DELIVERABLE 3.1chatziva.com/publications/W3SI_DV_8008_part_I.pdf¾ Power system security aspects are best served by VSC-HVDC reinforcements • 10 to 18 Gvar2 Economic competitiveness:

EC Contract n°: TREN/FP7EN/218903/˝IRENE-40˝

EUROPEAN COMMISSION DG-TREN

i

IRENE-40 REFERENCE W3 SI DV 8008 18/10/12 Internal partner reference: Filing N° Doc.Type Order N° Rev. N° Date ________________________________________________________________________________________________________

Alstom – TUD – ECN – ETH Zurich – Imperial – ICCS-NTUA – RWTH – ABB – Siemens

- DELIVERABLE 3.1 -

APPLICATION GUIDE FOR THE IMPROVEMENT OF

ECOLOGICAL SUSTAINABILITY, SECURITY AND COMPETITIVENESS BY INFRASTRUCTURAL

CHANGES

YES NO

Distribution List: Alstom 9

TUD 9

ECN 9

ETH Zurich 9

Imperial 9

ICCS-NTUA 9

RWTH 9

ABB 9

Siemens 9

EC DG - TREN 9

E

D

C

B

A

02/03/2012 Enrique Gaxiola (Siemens AG) Holger Müller (Siemens AG)

Frans Nieuwenhout (ECN)

Prof. Lou v.d. Sluis (TU Delft)

PU



ii



Liang Tao (Siemens AG) Hendrik Natemeyer (RWTH) Andreas Roehder (RWTH) Sebastian Winter (RWTH) George Orfanos (NTUA) Iraklis Skoteinos (NTUA) Özge Özdemir (ECN) Marit van Hout (ECN) Gerard Doorman (ECN) Jos Sijm (ECN) Frans Nieuwenhout (ECN) Spyros Chatzivasileiadis (ETH Zürich) Thilo Krause (ETH Zurich) George Pereira (Alstom Grid) Manuel Castro (ICL) Danny Pudjianto (ICL) Marko Aunedi (ICL) Goran Strbac (ICL) Christos Vasilakos Konstantinidis (Imperial) Rosy Wang (TUD) Zongyu Liu (TUD) George Papaefthymiou (TUD)

Rev. Date Drafted Checked Approved Status*

* I: Internal; PP: Restricted to Programme participant; RE: Restricted to specified group; CO: Confidential; PU: Public

REVISION

First issue 02.03.2012 Second issue 14.08.2012 Reduced to 409 pages Third issue 18.10.2012 Reduced to 344 pages Fourth issue 22.11.2012 Inclusion of full authors list



iii



Executive Summary Building on the Work Package 1 “Investigation” and Work Package 2 “Analyses” foundation ground we here present the modelling works for four domains of special interest. This report describes works entitled “Solutions” as undertaken in Work Package 3 of the IRENE-40 project, i.e. in the four dedicated Tasks 3.1, 3.3, 3.4 and 3.5. Input into Work Package 3 are the scenarios developed combined with the technology database information gathered and the modelling and definitions framework development and system analyses performed in the other eleven earlier tasks. Our scenarios were evaluated on three criteria: environmental sustainability, security and economic competitiveness. In addition we report on harmonization and coordination among national networks and its importance to the above pan-European approaches. Below we summarize the main findings w.r.t: Environmental sustainability:

• G&D scenarios (except BAU scenario) allow reaching emissions reduction target • Network development is highly necessary to reach CO2 reduction targets, especially in case

of a high penetration of renewables • Lack of network development would lead to up to 400 million additional tons of CO2 in

2050 (compared to extended grid reaching emission targets) • High resolution network model used to highlight regions with RES curtailment (‘ecological

weak network points’) • CO2 emission pricing can have a high influence on total generation costs and emissions,

especially in case of high penetration of CCS technology • Especially in the scenarios with a high share of RES (RES and DES), investments in

additional transmission infrastructure are highly profitable. Security: • Upgrade with HVDC is the preferable infrastructure scenario due to

¾ lowest cost for security (SC-OPF) AND operation (OPF) ¾ increased network utilization (hence more critical contingencies might arise) ¾ the possibility to achieve zero cost of security

• Overlay networks are preferable ¾ Less line-kilometers achieve the same effect as compared with internal

reinforcements Assumption: intra-country grid is more meshed than cross-border lines • Need for increased controllability is identified in all scenarios i.e. AC1

1 UHVAC (750kV) or HVAC++ (400kV): See this deliverable D3.1 & D3.2.

+FACTS or HVDC flows



iv



¾ Highly-loaded UHVAC & HVAC transmission lines congestion is resolved by an efficiently placed FACTS device

¾ Power system security aspects are best served by VSC-HVDC reinforcements • 10 to 18 Gvar2

Economic competitiveness:

reactive power series compensation is needed (i.e. FACTS) for the expansion benchmarks UHVAC or HVAC++

� Decarbonisation is technically and economically feasible. � Large scale renewable energy integration is achievable. � Different decarbonisation pathways drive dissimilar levels of investment in transmission

network infrastructure. � Expansion of cross-border grid contributes to efficient utilisation of low-carbon resources � Development of a European grid facilitates competitive cross-border trading and enhances

security of supply � Cost to consumers can potentially be reduced � Demand response can reduce need for transmission infrastructure reinforcements

Harmonization and coordination among national networks (This relates to European3

� Low levels of market unification might hamper the harmonised development of the Trans-European Grid as studied in IRENE-40.

Energy Markets Study/Analysis with four variants of capacity market unification level):

� Likely to end up between “3 island-continent” and “7 balancing cluster” markets in trade-off between large - to minimize curtailed RES energy (due to restricted peak generator capacity in certain high-RES regions: in favor of uniform European capacity market) - and small balancing circles - to minimize capacity market need (in favor of national separate markets)

� Without above capacity market: 1.) BAU & CCS scenarios seem best for peak generator; 2.) mid-cases (10% to 50% installation cost from capacity market) are RES and EFF; 3.) most challenging scenario appears to be DES

� Initially based on Financial Transmission Rights (FTRs) obligations: Necessary to establish efficient markets for long term transmission rights

� Essential for network codes to provide harmonized data exchange in a common format and contents w.r.t. grid model - on a regional-and EU-level for individual TSO´s - to perform efficient transmission capacity calculation and effective operational planning

� Locationally varying generation transmission tariffs can help minimize intra-zonal congestion given zonal definition of price areas. Transmission cost recovery should be on a “beneficiary pays” principle, pan-European wide based with a 2040 horizon view (Not to hamper competition Transmission Tariff Harmonization is a necessity).

2 Assumption: For ten 500 km long AC lines; each interconnection consisting of two similar voltage parallel paths inside the countries. 3 Note: Although Norway and Switzerland are not official EU member states, they are nonetheless integrated into the 27+2 IRENE scenario definition and thus included in the capacity market variant analysis.



v



� EU wholesale & balancing unification has a profound impact on the overall capacity market volume across Europe. The most efficient capacity market mechanism in the long run would be forward capacity requirement / reliability options system.

In Task 3.2 the resulting conclusions and works are further extended based on above work as summarized in Deliverable 3.2 Benchmark Network Scenario. The common bottlenecks and solutions found form the basis for establishing the exploitation roadmaps and strategies in Work Package 4.



vi



CONTENTS

1 INTRODUCTION ................................................................................................ 1

2 ANALYSIS ............................................................................................................ 1

3 RESULTS .............................................................................................................. 3 3.1 INFRASTRUCTURE SOLUTIONS FOR INCREASED ECOLOGICAL SUSTAINABILITY ........................................................................................................................ 3

3.1.1 Introduction ..................................................................................................................... 3 3.1.2 Model Framework ........................................................................................................... 3 3.1.2 Emission pricing to achieve environmental targets and environmental aspects of IRENE-40 scenarios (ECN) ....................................................................................................... 14 3.1.3 General results and comparison of low and high resolutions model (NTUA, RWTH) 27 3.1.4 Cross-border flows and impact of emission pricing (NTUA) ....................................... 28 3.1.5 Identification of ecological weak network points (RWTH) .......................................... 48 3.1.6 Assessment of grid development measures (RWTH) ................................................... 56

3.2 INFRASTRUCTURE MEASURES FOR INCREASED SECURITY .......................... 68 3.2.1 Introduction ................................................................................................................... 68 3.2.2 Comparison of technologies .......................................................................................... 87 3.2.3 Network modelling ...................................................................................................... 100 3.2.4 Generator modelling .................................................................................................... 100 3.2.5 Description of the AC-OPF and the security-constrained OPF algorithms ................ 100 3.2.6 Expansion scenarios – theoretical approach................................................................ 108 3.2.7 Expansion scenarios – simulations .............................................................................. 123

3.3 INFRASTRUCTURE SOLUTIONS FOR INCREASED COMPETITIVENESS .... 142 3.3.1 Introduction ................................................................................................................. 142 3.3.2 Generic analysis of different transmission infrastructure investment alternatives ..... 146 3.3.3 Future transmission infrastructure investment in Europe ........................................... 147 3.3.4 Technology options for the future transmission infrastructure investment in Europe 205

3.4 COORDINATION AND HARMONISATION AMONG NATIONAL NETWORKS223 3.4.1 European electrical network ........................................................................................ 223 3.4.2 EUROPEAN ENERGY MARKETS .......................................................................... 237

4 CONCLUSIONS ............................................................................................... 308 4.1 ENVIRONMENTAL SUSTAINABILITY .................................................................... 308 4.2 SECURITY-CONSTRAINED OPF AND COST OF SECURITY .............................. 310

4.2.1 Selection of critical contingencies............................................................................... 310 4.2.2 Cost of Security ........................................................................................................... 315

4.3 COMPETITIVENESS ..................................................................................................... 321 4.3.1 Introduction ................................................................................................................. 321 4.3.2 Response to the main questions .................................................................................. 322



vii



4.4 COORDINATION AND HARMONISATION AMONG NATIONAL NETWORKS327 4.4.1 Our recommendations for grid code and EC targets ................................................... 327 4.4.2 Recommendations for a Future Market Grid Code ..................................................... 335

5 ACKNOWLEDGEMENT ............................................................................... 345

6 ANNEX .............................................................................................................. 345 6.1 REFERENCES ................................................................................................................. 346

6.1.1 References §3.1 ........................................................................................................... 346 6.1.2 References §3.2 ........................................................................................................... 347 6.1.3 References §3.3 ........................................................................................................... 352 6.1.4 References §3.4 ........................................................................................................... 357



viii



FIGURES

Figure 1 – Scenario Synthesis input role into Work Package 3 Tasks ...................................... 1Figure 2 – Analysis in Work Package 3 ..................................................................................... 2Figure 3: Geographical allocation of countries in the low resolution model and

interconnections considered ............................................................................................... 6Figure 4: Detailed model of the European transmission network used in T3.1 ......................... 9Figure 5: Load distribution based on population density ......................................................... 10Figure 6: Power plant database ................................................................................................ 11Figure 7: Exemplary: Distribution of installed wind power capacity ...................................... 12Figure 8: Identified and implemented network development measures .................................. 14Figure 9: The physical representation of the electricity network of the EU27 + 2 and the

Balkan countries as modeled in COMPETES; the dotted lines represent external imports as included in the DESERTEC scenario from the African continent. ............................. 15

Figure 10: Average generation costs (in euro/MWh) in 2050 for all IRENE-40 scenario’s for both EU27 and EU27 + Norway and Switzerland. .......................................................... 19

Figure 11: CO2 emissions (ton CO2/MWh) per scenario in 2050, represented for both EU27 and EU27 + Norway and Switzerland. ............................................................................ 21

Figure 12: Contribution of generation per technology type (in %) to total generation per IRENE-40 scenario represented for both EU27 and EU27 + Norway and Switzerland. 22

Figure 13: General result discussion/model comparison: Average generation costs .............. 27Figure 14: General result discussion/model comparison: Average CO2-emissions ................ 28Figure 15: Average CO2 emissions (kg/KWh) for the five examined scenarios ..................... 30Figure 16: Total generation cost (million €) for the five examined scenarios ......................... 31Figure 17: Res curtailment (%) for the five examined scenarios ............................................. 32Figure 18: Graphical representation of each technology type contribution to total generation

for all scenarios with fixed NTC 2020 values ................................................................. 34Figure 19: Graphical representation of each technology type contribution to total generation

for all scenarios with the forecasted NTC values for years 2030, 2040 and 2050 .......... 35Figure 20: Congestion hours between the 71 interconnection for 2030, fixed 2020 NTC

values ............................................................................................................................... 38Figure 21: Congestion hours between the 71 interconnection for 2040, fixed 2020 NTC

values ............................................................................................................................... 40Figure 22: Congestion hours between the 71 interconnection for 2050, fixed 2020 NTC

values ............................................................................................................................... 42Figure 23: Average CO2 emissions (kg/KWh) for the additional lower CO2 prices scenarios 44Figure 24: Total generation cost (million €) for the additional lower CO2 prices scenarios ... 45



ix



Figure 25: RES curtailment (%) for the additional lower CO2 prices scenarios ..................... 46Figure 26: Ecological weak network point analysis 2020 ....................................................... 49Figure 27: Ecological weak network point analysis 2030 ....................................................... 50Figure 28: Ecological weak network point analysis 2040 ....................................................... 51Figure 29: Ecological weak network point analysis 2050, BAU ............................................. 52Figure 30: Ecological weak network point analysis 2050, CCS .............................................. 52Figure 31: Ecological weak network point analysis 2050, Efficiency (EFF) .......................... 53Figure 32: Ecological weak network point analysis 2050, Renewables (RES) ....................... 53Figure 33: Ecological weak network point analysis Desertec (DES) ...................................... 55Figure 34: Extension measures selected for analysis ............................................................... 57Figure 35: Additional infeed of RES caused by line “Diele-Niederrhein” .............................. 59Figure 36: Additional infeed of RES caused by line “Spain-France” ..................................... 60Figure 37: Additional infeed of RES caused by line “France-Germany” ................................ 61Figure 38: Changes in CO2-emissions caused by line “Diele-Niederrhein” ............................ 62Figure 39: Changes in CO2-emissions caused by line “Spain-France” ................................... 63Figure 40: Changes in CO2-emissions caused by line “France-Germany” .............................. 64Figure 41: Changes in costs caused by line “Diele-Niederrhein” ............................................ 65Figure 42: Changes in costs caused by line “Spain-France” ................................................... 66Figure 43: Changes in costs caused by line “France-Germany” .............................................. 66Figure 44: Present value ........................................................................................................... 67Figure 3.43 – Maximum active power loading as a function of the line length [EEH11] ....... 77Figure 3.44 – Illustrative Comparison of Line Reinforcements against Additional Lines

[VEN11] ........................................................................................................................... 77Figure 3.45 – AC-OPF: Annual Total Generation Costs Reduction (in %) due to the

expansion scenarios ......................................................................................................... 85Figure 3.46 – Cost of Security for the different generation and expansion scenarios. The

selected critical contingencies are the lines which during the OPF calculations had an average annual loading above 85%. Therefore, for each case a different number of contingencies is taken into account. ................................................................................. 85

Figure 3.47 – Dena Grid Study II Expansion and Annual Costs for different Transmission Options for consideration in Germany [DEN01] ............................................................ 89

Figure 3.48 – Isokeraunic Levels in Continental France (Annual mean number of Stormy days) ................................................................................................................................. 92

Figure 3.49 – St.Clair Curve illustrating the maximum active power loading as a function of the line length [EEH11]. 𝑃𝑆𝐼𝐿 U refers to the surge impedance loading of the line. ........... 94

Figure 3.50: Equivalent circuit of the VSC-HVDC line for OPF calculations ...................... 102Figure 3.51: Illustration of the Current-Injection Method ..................................................... 104



x



Figure 3.52 – Representation of the addition of a new AC line 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 Uin parallel with the existing interconnection 𝑍𝑒𝑥𝑡 U. If 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 > 𝑍𝑒𝑥𝑡 U and 𝐼𝑛𝑒𝑤,𝑙𝑖𝑚 ≥ 𝐼𝑒𝑥𝑡,𝑙𝑖𝑚 = 𝑁𝑇𝐶U, then the already existing interconnection will reach its capacity limit before the new line. . 109

Figure 3.55 – Loadings of Additional Lines: Comparison of theoretical results based on Eq. 3.39 (New Line: Theoretical Max Loading) and actual results from simulations (New Line: Actual Max. Loading). The total loading of the interconnection is the sum of the “Existing NTC Value” and the “New Line: Actual Max. Loading” [RES 2050 Scenario]. The “Existing NTC Value” is equal to 1.43*NTC(2010), where NTC(2010) is given in T2.3 Report. ..................................................................................................... 112

Figure 3.54 – Relationship between the impedance ratio and the net transfer capacity. The dashed line illustrates the trendline. ............................................................................... 113

Figure 3.55 – Adding a long parallel AC line to an existing interconnection. ...................... 114Figure 3.56 – Illustration of the additional internal impedance. ............................................ 114Figure 3.57 – Accounting for the network internal impedance with an equivalent 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡

𝑒𝑞. . 117 Figure 3.58 – Illustration of internal network reinforcements. .............................................. 119Figure 3.59 – BAU 2050 Scenario: Duration of Line Loadings in hours per year. The line

loadings are expressed in %. Lines that are loaded above 99% for more than 4000 hours per year are selected for reinforcement .......................................................................... 126

Figure 3.60 – CCS 2050 Scenario: Duration of Line Loadings in hours per year. The line loadings are expressed in %. Lines that are loaded above 99% for more than 4000 hours per year are selected for reinforcement .......................................................................... 127

Figure 3.61 – DES 2050 Scenario: Duration of Line Loadings in hours per year. The line loadings are expressed in %. Lines that are loaded above 99% for more than 4000 hours per year are selected for reinforcement .......................................................................... 128

Figure 3.62 – EFF 2050 Scenario: Duration of Line Loadings in hours per year. The line loadings are expressed in %. Lines that are loaded above 99% for more than 4000 hours per year are selected for reinforcement .......................................................................... 130

Figure 3.63 – RES 2050 Scenario: Duration of Line Loadings in hours per year. The line loadings are expressed in %. Lines that are loaded above 99% for more than 4000 hours per year are selected for reinforcement .......................................................................... 131

Figure 3.64 – AC-OPF: Annual Total Generation Costs for all scenarios ............................ 138Figure 3.65 – AC-OPF: Annual Total Generation Costs Reduction (in %) due to the

expansion scenarios ....................................................................................................... 139Figure 66: Technology options for efficient integration of low carbon generation systems . 148Figure 67: Overview of the modelling approach ................................................................... 149Figure 68: Schematic representation of the generation and transmission investment model 150Figure 69: Schematic representation of the system balancing model .................................... 152Figure 70: Europe – main interconnected transmission system ............................................. 153



xi



Figure 71: BAU scenario, Year 2030 – generation and demand background ....................... 154Figure 72: BAU scenario, Year 2050 – generation and demand background ....................... 155Figure 73: BAU scenario – Transmission network capacity requirements ........................... 159Figure 74: BAU scenario, Year 2030 – transmission network capacity reinforcements ....... 160Figure 75: BAU scenario, Year 2050 – transmission network capacity reinforcements ....... 161Figure 76: BAU scenario, Year 2050 – firm capacity margin ............................................... 162Figure 77: RES scenario, Year 2030 – generation and demand background ........................ 163Figure 78: RES scenario, Year 2050 – generation and demand background ........................ 164Figure 79: RES scenario – Transmission network capacity requirements ............................ 166Figure 80: RES scenario, Year 2030 – transmission network capacity reinforcements ........ 167Figure 81: RES scenario, Year 2050 – transmission network capacity reinforcements ........ 168Figure 82: Interconnector Spain – France: flow utilisation ................................................... 169Figure 83: Spain – normalised wind power output ................................................................ 169Figure 84: Spain – normalised solar power output ................................................................ 170Figure 85: Spain – correlation between renewable energy sources and maximum energy

exports ............................................................................................................................ 170Figure 86: Spain – correlation between renewable energy sources and maximum energy

imports ........................................................................................................................... 171Figure 87: Interconnector France – Belgium: flow utilisation ............................................... 172Figure 88: Belgium – normalised wind power output ........................................................... 172Figure 89: Interconnector Spain – France: flow utilisation ................................................... 172Figure 90: Interconnector UK – Belgium: flow utilisation .................................................... 174Figure 91: UK – normalised wind power output ................................................................... 174Figure 92: Interconnector France – Belgium: flow utilisation ............................................... 174Figure 93: Belgium – normalised wind power output ........................................................... 174Figure 94: Interconnector Germany – Denmark: flow utilisation ......................................... 175Figure 95: Denmark – normalised wind power output .......................................................... 175Figure 96: Denmark – normalised solar power output .......................................................... 176Figure 97: DES scenario, Year 2030 – generation and demand background ........................ 177Figure 98: DES scenario, Year 2050 – generation and demand background ........................ 178Figure 99: DES scenario – Transmission network capacity requirements ............................ 180Figure 100: DES scenario, Year 2030 – transmission network capacity reinforcements ...... 181Figure 101: DES scenario, Year 2050 – transmission network capacity reinforcements ...... 182Figure 102: Interconnector Spain – France: flow utilisation ................................................. 183Figure 103: Spain – normalised wind power output .............................................................. 183Figure 104: Spain – normalised solar power output .............................................................. 183Figure 105: Interconnector Italy – Austria: flow utilisation .................................................. 185Figure 106: Italy – normalised wind power output ................................................................ 185



xii



Figure 107: Italy – normalised solar power output ................................................................ 185Figure 108: Austria – normalised wind power output ........................................................... 185Figure 109: CCS scenario, Year 2030 – generation and demand background ...................... 186Figure 110: CCS scenario, Year 2050 – generation and demand background ...................... 187Figure 111: CCS scenario – Transmission network capacity requirements .......................... 189Figure 112: CCS scenario, Year 2030 – transmission network capacity reinforcements ...... 190Figure 113: CCS scenario, Year 2050 – transmission network capacity reinforcements ...... 191Figure 114: CCS scenario, Year 2050 – firm capacity margin .............................................. 192Figure 115: EFF scenario, Year 2030 – generation and demand background ....................... 193Figure 116: EFF scenario, Year 2050 – generation and demand background ....................... 193Figure 117: EFF scenario – Transmission network capacity requirements ........................... 196Figure 118: EFF scenario, Year 2030 – transmission network capacity reinforcements ...... 196Figure 119: EFF scenario, Year 2050 – transmission network capacity reinforcements ...... 197Figure 120: Transmission network capacity requirements .................................................... 200Figure 121: Transmission capacity contributed by demand side flexibility .......................... 201Figure 122: Transmission investment contributed by demand side flexibility ...................... 202Figure 123: Overview of the modelling approach ................................................................. 206Figure 124: Flow of information – technology database and scientific models .................... 206Figure 125: Power system infrastructure expansion technologies ......................................... 207Figure 126: The transmission network development within the framework of energy systems

........................................................................................................................................ 208Figure 127: BAU scenario, Year 2030 – technological transmission infrastructure solution 209Figure 128: RES scenario, Year 2030 – technological transmission infrastructure solution 210Figure 129: DES scenario, Year 2030 – technological transmission infrastructure solution 211Figure 130: CCS scenario, Year 2030 – technological transmission infrastructure solution 212Figure 131: EFF scenario, Year 2030 – technological transmission infrastructure solution . 213Figure 132: BAU scenario, Year 2050 – technological transmission infrastructure solution 215Figure 133: RES scenario, Year 2050 – technological transmission infrastructure solution 216Figure 134: DES scenario, Year 2050 – technological transmission infrastructure solution 217Figure 135: CCS scenario, Year 2050 – technological transmission infrastructure solution 218Figure 136: EFF scenario, Year 2050 – technological transmission infrastructure solution . 219Figure 2-137 EU Transmission Areas

[http://en.wikipedia.org/wiki/Wide_area_synchronous_grid] ....................................... 243Figure 2-138 EU Regional Initiative Segments (ERGEG, [ERGE06]) ................................. 244Figure 2-139: .......................................................................................................................... 251Figure 2-140: Payoff Curves for FTRs as Options or Obligations against Price Differential

........................................................................................................................................ 253



xiii



Figure 2-141: Payoff Curves for FTRs as Options or Obligations against Price Differential [WH11] .......................................................................................................................... 257

Figure 2-142 Annual per unit hydro profiles of all 27 EU states .......................................... 269Figure 2-143 Overview of balancing circle setting for variant 1 ........................................... 270Figure 2-144 Overview of balancing circle setting for variant 2 ........................................... 271Figure 2-145 Overview of balancing circle setting for variant 3 ........................................... 272Figure 2-146 Overview of balancing circle setting for variant 4 ........................................... 273Figure 2-147 Installation costs of different firm capacity technologies ................................ 274Figure 2-148 Marginal generation costs of different firm capacity technologies .................. 274Figure 2-149 Nuclear units' expected wholesale clearance price under varied dispatch level

........................................................................................................................................ 275Figure 2-150 Hydro units' expected wholesale clearance price under varied dispatch level . 275Figure 2-151 Coal units' expected wholesale clearance price under varied dispatch level ... 275Figure 2-152 Biomass units' expected wholesale clearance price under varied dispatch level

........................................................................................................................................ 276Figure 2-153 Gas units' expected wholesale clearance price under varied dispatch level ..... 276Figure 2-154 Oil units' expected wholesale clearance price under varied dispatch level ...... 276Figure 2-155 Sample annual energy balance dispatch result (hourly data) ........................... 278Figure 2-156 Peak gas unit average full load hours estimation for variant 1 ........................ 280Figure 2-157 Peak gas unit average full load hours estimation for variant 2 ........................ 280Figure 2-158 Peak gas unit average full load hours estimation for variant 3 ........................ 281Figure 2-159 Peak gas unit average full load hours estimation for variant 4 ........................ 281Figure 2-160 Peak oil unit average full load hours estimation for variant 1 ......................... 282Figure 2-161 Peak oil unit average full load hours estimation for variant 2 ......................... 283Figure 2-162 Peak oil unit average full load hours estimation for variant 3 ......................... 283Figure 2-163 Peak oil unit average full load hours estimation for variant 4 ......................... 284Figure 2-164 Gas units’ capacity market volume as percentage of installation cost for variant

1 ...................................................................................................................................... 285Figure 2-165 Gas units’ capacity market volume as percentage of installation cost for variant



4 ...................................................................................................................................... 286Figure 2-168 Oil units’ capacity market volume as percentage of installation cost for variant 1

........................................................................................................................................ 287Figure 2-169 Oil units’ capacity market volume as percentage of installation cost for variant 2

........................................................................................................................................ 288



xiv



Figure 2-170 Oil units’ capacity market volume as percentage of installation cost for variant 3 ........................................................................................................................................ 288

Figure 2-171 Oil units’ capacity market volume as percentage of installation cost for variant 4 ........................................................................................................................................ 289

Figure 2-172 Percentage of totally curtailed renewable generation per year for variant 1 .... 290Figure 2-173 Percentage of totally curtailed renewable generation per year for variant 2 .... 290Figure 2-174 Percentage of totally curtailed renewable generation per year for variant 3 .... 291Figure 2-175 Percentage of totally curtailed renewable generation per year for variant 4 .... 291Figure 2-176 Future energy supply mix under BAU scenario and variant 1 ......................... 294Figure 2-177 Future energy supply mix under BAU scenario and variant 2 ......................... 294Figure 2-178 Future energy supply mix under BAU scenario and variant 3 ......................... 295Figure 2-179 Future energy supply mix under BAU scenario and variant 4 ......................... 295Figure 2-180 Future energy supply mix under CCS scenario and variant 1 .......................... 296Figure 2-181 Future energy supply mix under CCS scenario and variant 2 .......................... 296Figure 2-182 Future energy supply mix under CCS scenario and variant 3 .......................... 297Figure 2-183 Future energy supply mix under CCS scenario and variant 4 .......................... 297Figure 2-184 Future energy supply mix under DES scenario and variant 1 .......................... 298Figure 2-185 Future energy supply mix under DES scenario and variant 2 .......................... 298Figure 2-186 Future energy supply mix under DES scenario and variant 3 .......................... 299Figure 2-187 Future energy supply mix under DES scenario and variant 4 .......................... 299Figure 2-188 Future energy supply mix under RES scenario and variant 1 .......................... 300Figure 2-189 Future energy supply mix under RES scenario and variant 2 .......................... 300Figure 2-190 Future energy supply mix under RES scenario and variant 3 .......................... 301Figure 2-191 Future energy supply mix under RES scenario and variant 4 .......................... 301Figure 2-192 Future energy supply mix under EFF scenario and variant 1 .......................... 302Figure 2-193 Future energy supply mix under EFF scenario and variant 2 .......................... 302Figure 2-194 Future energy supply mix under EFF scenario and variant 3 .......................... 303Figure 2-195 Future energy supply mix under EFF scenario and variant 4 .......................... 303Figure 4.1 – BAU 2050: Annual Average Line Loadings. High Loading: >85%. Medium

Loading: 50%-85%. Low Loading: <50%. The line width is proportional to the line capacity. ......................................................................................................................... 311

Figure 4.2 – CCS 2050: Annual Average Line Loadings. High Loading: >85%. Medium Loading: 50%-85%. Low Loading: <50%. The line width is proportional to the line capacity. ......................................................................................................................... 312

Figure 4.3 – DES 2050: Annual Average Line Loadings. High Loading: >85%. Medium Loading: 50%-85%. Low Loading: <50%. The line width is proportional to the line capacity. ......................................................................................................................... 313



xv



Figure 4.4 – EFF 2050: Annual Average Line Loadings. High Loading: >85%. Medium Loading: 50%-85%. Low Loading: <50%. The line width is proportional to the line capacity. ......................................................................................................................... 314

Figure 4.5 – RES 2050: Annual Average Line Loadings. High Loading: >85%. Medium Loading: 50%-85%. Low Loading: <50%. The line width is proportional to the line capacity. ......................................................................................................................... 315

Figure 4.6 – Comparison of the Cost of Security of each expansion scenario for RES 2050. The selected hour samples are the ones which were common in all three scenarios. .... 317

Figure 4.7 – Cost of Security for the different generation and expansion scenarios. The selected critical contingencies are the lines which during the OPF calculations had an average annual loading above 85%. Therefore, for each case a different number of contingencies is taken into account. ............................................................................... 319

Figure 4.8 – CCS 2050: Comparison of the Cost of Security results. Left set of bars: only the lines with an annual average loading above 85% are selected as critical contingencies in each expansion scenario (different number of contingencies in each case). Right set of bars: all expansion scenarios consider the same lines as critical contingencies. ........... 320

TABLES

Table 1: Countries considered within the low resolution network model ................................. 5Table 2: List of generation units per country ............................................................................. 7Table 3: BAU scenario fuel and CO2 emissions prices ............................................................. 8Table 4: Remaining scenarios fuel and CO2 emissions prices ................................................... 8Table 5: Overview of the average generation costs in €/MWh calculated by dividing total

renewable generation costs and total renewable generation etc. ..................................... 20Table 6: COMPETES results showing generation per technology, CO2 emissions, costs of

generation, curtailment of intermittent generation (% of total generation) and congestion rents, per scenario in 2050 using 2050 NTC values. ....................................................... 25

Table 7: Generation simulation results for the BAU scenario ................................................. 29Table 8: Generation simulation results for the high RES scenario .......................................... 29Table 9: Generation simulation results for the DESERTEC scenario ..................................... 29Table 10: Generation simulation results for the Efficiency scenario ....................................... 29Table 11: Generation simulation results for the CCS scenario ................................................ 30Table 12: Percentage of technology type contribution to total generation, BAU scenario ..... 32Table 13: Percentage of technology type contribution to total generation, high RES scenario

.......................................................................................................................................... 32Table 14: Percentage of technology type contribution to total generation, DESERTEC

scenario ............................................................................................................................ 33



xvi



Table 15: Percentage of technology type contribution to total generation, Efficiency scenario .......................................................................................................................................... 33

Table 16: Percentage of technology type contribution to total generation, CCS scenario ...... 33Table 17: Fuel and CO2 prices considered in the additional study cases ................................. 42Table 18: Generation simulation results for the high RES scenario with lower CO2 prices ... 42Table 19: Generation simulation results for the DESERTEC scenario with lower CO2 prices

.......................................................................................................................................... 43Table 20: Generation simulation results for the Efficiency scenario with lower CO2 prices .. 43Table 21: Generation simulation results for the CCS scenario with lower CO2 prices ........... 43Table 22: Increase of total CO2 emissions due to lower CO2 prices, in kg/KWh and % for the

four additional study cases ............................................................................................... 47Table 23: Decrease of total generation fuel cost due to lower CO2 prices, in M€ and % for the

four additional study cases ............................................................................................... 47Table 24: Cost parameters of selected network development measures .................................. 57Table 25: CO2-emissions resulting from construction of selected network development

measures ........................................................................................................................... 58Table 3.26 – Technology Comparison for Increased Security ................................................ 71Table 3.27 – Possibilities for combinations of the technology options ................................... 72Table 3.28 – RES 2050: Comparison table of generation costs for different expansion

scenarios ........................................................................................................................... 76Table 3.29 – RES 2050: OPF Results from different expansion scenarios ............................. 81Table 3.30 – Classification and ranking of all the selected Network Reinforcements according

to the number of generation scenarios they participate ................................................... 83Table 3.31 – Total amount of series compensation that is necessary for each generation

scenario ............................................................................................................................ 84Table 3.32 – Construction time for the transmission technologies [BEN01] .......................... 87Table 3.33 – Annual Generation Costs for the RES 2050 Scenario. The new AC and HVDC

lines have each a length of 500 km. ............................................................................... 108Table 3.34 – Parallel AC lines 400 kV/3000 MVA that were added to relieve overloading

during the five generation scenarios .............................................................................. 111Table 3.35 – Comparing the effect of the additional lines' length on the annual generation

costs ................................................................................................................................ 116Table 3.36 – Comparison table of generation costs for different expansion scenarios ......... 121Table 3.37 – RES 2050: OPF Results from different expansion scenarios ........................... 122Table 3.38 – Line Reinforcements that are necessary in order to ensure that load is served

100% of the time in 2050. .............................................................................................. 123Table 3.39 – BAU 2050: Location and transmission capacity of the line reinforcements .... 126Table 3.40 – CCS 2050: Location and transmission capacity of the line reinforcements .... 127



xvii



Table 3.41 – DES 2050: Location and transmission capacity of the line reinforcements ..... 129Table 3.42 – EFF 2050: Location and transmission capacity of the line reinforcements ..... 130Table 3.43 – RES 2050: Location and transmission capacity of the line reinforcements ..... 131Table 3.44 – Classification and ranking of all the selected Network Reinforcements according

to the number of generation scenarios they participate ................................................. 132Table 3.45 – Series compensation for the AC-400 kV and AC-750 kV expansion scenarios,

expressed in degree of compensation and amount of reactive power. (Assumptions: lines are 500 km long; for each interconnection there are two parallel paths of similar voltage inside the countries) ....................................................................................................... 136

Table 3.46 – Total amount of series compensation that is necessary for each generation scenario .......................................................................................................................... 136

Table 3.47 – Valuation Results for BAU 2050 ...................................................................... 140Table 3.48 – Valuation Results for CCS 2050 ....................................................................... 140Table 3.49 – Valuation Results for DES 2050 ....................................................................... 141Table 3.50 – Valuation Results for EFF 2050 ....................................................................... 141Table 3.51 – Valuation Results for RES 2050 ....................................................................... 141Table 52: BAU scenario – summary of the main characteristics of the system .................... 156Table 53: BAU scenario – summary of the main costs components of the system ............... 157Table 54: RES scenario – summary of the main characteristics of the system ..................... 164Table 55: RES scenario – summary of the main costs components of the system ................ 165Table 56: DES scenario – summary of the main characteristics of the system ..................... 178Table 57: DES scenario – summary of the main costs components of the system ................ 179Table 58: CCS scenario – summary of the main characteristics of the system ..................... 187Table 59: CCS scenario – summary of the main costs components of the system ................ 188Table 60: EFF scenario – summary of the main characteristics of the system ...................... 194Table 61: EFF scenario – summary of the main costs components of the system ................ 194Table 62: Summary of the main characteristics of the system in the year 2050 ................... 197Table 63: Summary of the energy production of the various generation technologies in the

year 2050 ........................................................................................................................ 198Table 64: Summary of the main costs components of the system in the year 2050 .............. 199Table 3-65 Main characteristics of national electricity transmission tariffs ........................ 256Table 4.1 – BAU 2050: Critical Contingencies for the three expansion scenarios ............... 311Table 4.2 – CCS 2050: Critical Contingencies for the three expansion scenarios ................ 312Table 4.3 – DES 2050: Critical Contingencies for the three expansion scenarios ................ 313Table 4.4 – EFF 2050: Critical Contingencies for the three expansion scenarios ................. 314Table 4.5 – RES 2050: Critical Contingencies for the three expansion scenarios ................ 315Table 4.6 – Number of snapshots (in % of common samples size) which do not follow the

general trend in Figure 4.6 ............................................................................................. 318



xviii



Table 4.7 – Number of Snapshots with zero Cost of Security ............................................... 318



1



1 INTRODUCTION

The document has been issued by Siemens AG and represents IRENE-40 project works as undertaken in Work Package 3 Task 3.1 lead by RWTH Aachen, Task 3.3 lead by ETH Zürich, Task 3.4 lead by Imperial and Task 3.5 lead by TU Delft.

2 ANALYSIS

Based on the Workpackage 1 and Work Package 2 works we here present the modelling works for three domains of special interest as shown in Figure 1, complemented with harmonisation insights. Input into Work Package 3: The scenarios developed by ECN and validated by Siemens AG combine with the technology information gathered and the system analyses performed in the other tasks of Work Package 2 to form the input into Work Package 3 (Figure 1). In a following step the four scenarios were evaluated on three criteria: environmental sustainability, security and economic competitiveness. In Task 3.2 the resulting conclusions and works are further extended based on above work as illustrated in Figure 2. The common bottlenecks and solutions found form the basis for establishing the exploitation roadmaps and strategies in WP4 (Figure 2).

Figure 1 – Scenario Synthesis input role into Work Package 3 Tasks



2



Figure 2 – Analysis in Work Package 3

T3.1 EnvironmentalAssessment

EnvironmentalBottlenecks

PreventativeMeasures &

Technologies

Recommendations

T3.3 SecurityAssessment

Security Bottlenecks


Technologies

Recommendations

T3.4 EconomicAssessment

Economic Bottlenecks


Technologies

Recommendations

Network ScenariosBusiness-as-usual Renewables Renewable + Desertec Low-carbon

WP2 – Generation & Consumption ScenariosBusines-as-usual Renewables Renewables + Desertec Low-carbon

Common Bottlenecks

Preventative Measures & Technologies(balancing three objectives)

Recommendations

WP4 – Roadmap



3



3 RESULTS

3.1 INFRASTRUCTURE SOLUTIONS FOR INCREASED ECOLOGICAL SUSTAINABILITY

3.1.1 Introduction

The transformation of the energy systems that is initiated to reach the goal of environmentally friendly future and in the face of decreasing fossil resources lead to new challenges on the European electrical transmission network. The goal of IRENE-40 is to create a roadmap for these energy networks to adapt and fit to the changing requirements. The objective of Task 3.1 in this context is to indicate and evaluate infrastructural measures to increase ecological sustainability. The reduction of the CO2 emissions is today the main environmental objective of the European energy policy. The members of the European union agreed to reduce the CO2 emissions by 20% until 2020 (part of the 202020 goals). However, there is no consensus yet on further goals after 2020. It is however assumed that a reduction of CO2-emissions by 95% until 2050 is necessary to limit the limit the global temperature increase to 2°C what is an objective with a widespread agreement. Four of five IRENE-40 generation and demand scenarios meet this objective. The change of the demand and generation pattern however leads to widespread development needs for the transmissions network in Europe that could be a main bottleneck for the transformation process and therefore for the achievement of the environmental goals. Within Task 3.1 this will be pointed out and the impact of network developments on the total CO2 emissions will be highlighted.

3.1.2 Model Framework

For the purpose of Task 3.1 there has been proposed the utilization of two network models in order to investigate the solutions for increased ecological sustainability in the expanded European energy market until the Year 2050. The first model proposed is country-node model in which each European country is represented by a node which is connected to adjacent countries by a single transportation line. The purpose of such a low resolution model is to identify a general tendency of future congestions between European countries and to draw the needed focus on specific interconnections which will require major reinforcement’s in the future European grid. The second model is a more detailed network model in which several basic nodes are considered within each European country examined along with



4



multiple interconnections to other countries. The purpose of such a high resolution model is to examine even further the possible future congestions not only between adjacent European countries but also inside specific countries giving the advantage of more detailed possible network expansion solutions.

3.1.2.1 Reduced network model (NTUA)

Each model approach has its own benefits. A country based model is suitable to minimize the cost objective function while it can fulfill pre-defined technical and operational constraints. The model offers the following advantages.

• It covers a large geographic scope of European member countries. • It enables the analysis of interconnection strategies to create a single pan-European

grid and development of new markets for cross-border trade. • It requires less computational effort, suitable for hourly simulation, and offers

reasonable accuracy of the results. The initial low resolution model proposed for the purpose of the IRENE-40 Work package 3 was the country-node model developed by TU DELFT within Task 1.5. This model is a country-node AC OPF model based on the UCTE 2008 winter model developed in MATPOWER. The model consists of 26 buses with each bus representing one country, 46 branches (aggregated cross-border lines) and 103 generators. In this model 36 countries are in the database, and only UCTE member countries are considered. For the purpose of Task 3.1 it was decided not to use directly the already developed network model but instead develop a new low resolution model, different than TU-DELFT’s in order to cover a larger geographical area since in the initial low resolution UCTE model there were no interconnections with the Scandinavian or the Baltic countries, Great Britain and Ireland. The low resolution model developed by NTUA for the purpose of investigations for solutions for increased ecological sustainability in the expanded European energy market within the project IRENE-40 will be presented in this section of the Deliverable while simulation results for the five proposed generation demand scenarios proposed within Task 2.3 will be presented in section 3.

3.1.1 The 35+2 node low resolution network model

As already mentioned the purpose of the low resolution network model is to cover a big enough geographical area in order to include all the countries not only inside the EU but also in the general European peninsula with the addition of non-European specific countries of interest necessary for the examination of the future of the pan-European energy network. Therefore, the expanded model developed by NTUA for the purpose of Task 3.1 consists of 37 countries, each of which is represented by a node. Each node-country is connected to its adjacent node-countries by a single transmission line enabling the investigation of country



5



per country transfer energy needs for each scenario examined. The countries included in the low resolution network model along with their respective 2 digit ISO country code are presented in Table 1 while a geographical allocation is demonstrated in Figure 3 along with the considered country to country interconnections.

Table 1: Countries considered within the low resolution network model

No Country 2 Digit ISO Country Code 1 Albania AL 2 Austria AT 3 Bosnia and Herzegovina BA 4 Belgium BE 5 Bulgaria BG 6 Switzerland CH 7 Cyprus CY 8 Czech Republic CZ 9 Germany DE 10 Denmark DK 11 Estonia EE 12 Spain ES 13 Finland FI 14 France FR 15 Greece GR 16 Croatia HR 17 Hungary HU 18 Ireland IE 19 Italia IT 20 Latvia LV 21 Lithuania LT 22 Luxemburg LU 23 Morocco MA 24 Montenegro ME 25 FYROM MK 26 Malta MT 27 Netherlands NL 28 Norway NO 29 Poland PL 30 Portugal PT 31 Romania RO 32 Serbia RS 33 Sweden SE 34 Slovakia SK 35 Slovenia SV 36 Tunisia TN 37 United Kingdom UK



6



Figure 3: Geographical allocation of countries in the low resolution model and

interconnections considered

From the 37 countries included in the model, two countries, Cyprus and Malta, are considered autonomous, therefore not being connected to any other adjacent countries due to their geographical position, while the interconnections between Spain and Morocco and Tunisia and Italy (both shown with yellow lines in Figure 3) are only considered in the case of the DESERTEC scenario. As a total, between the remaining 33 countries there is a total of 71 interconnections for all scenarios plus two necessary extra interconnections for the DESERTEC scenario. The reduced network model used for the simulation of the four scenarios is a transportation model, based on Net Transfer Capacity (NTC) between adjacent interconnected countries. The NTC values used in this model derive from the NTC tables given as an input by ECN as



7



a part of Task 2.3 with a modification for Denmark, for which the initial NTC divided in Denmark-East and Denmark-West is not used; instead Denmark is treated as a whole with a single NTC value for each decade. The NTC values used for 2010 are averages of summer 2010 and winter 2009/2010 values published by ENTSO-E and include interconnection reinforcements planned in the ENTSO-E 10YNDP for 2010. The NTC values for 2020 are the sum of NTC values for 2010 and network reinforcements planned in the ENTSO-E 10YNDP for the period 2011-2019 and they are an average of the derived NTC values for summer 2020 as well as winter 2020. The data for 2030/40/50 have been derived by starting to assume that the sum of the NTC capacities of each country in 2050 will be related to 20% of total installed generation capacity plus 10% of the intermittent generation capacity. In between 2020 and 2050 a constant growth rate is taken to arrive at the values in 2030 and 2040. All the NTC data used for the four examined scenarios can be found in Excel files as an input of Task 2.3.

Table 2: List of generation units per country

No Generation Type Generation Technology 1

Ren

ewab

le e

nerg

y

Hydro (pump storage excluded) 2 Hydro (pumping systems) 3 Wind on-shore 4 Wind off-shore 5 Solar PV 6 Solar CSP 7 Marine (tidal, waves) 8 Geothermal 9 Biomass and Waste fired 10

Ther

mal

ene

rgy

Hard coal 11 Lignite 12 IGCC 13 IGCC CCS 14 CCGT (natural gas) 15 CCGT CCS 16 CHP (gas) 17 CHP CCS 18 GT and Gas boilers (natural gas) 19 IC engines (derived gases) 20 Oil fired 21 Nuclear energy Nuclear power plants

Regarding generation, for each country there are considered 21 generating units representing one generation technology each. Therefore, on each country node the generation input hourly is the sum of the existing generation outputs of each of those technologies. The list of the 21 considered generation units/technologies per country is shown in Table 2.

Special attention is hereby given to:



8



• Hydro and pump storage

• Renewable energy sources

• Installed capacities per technology per country, efficiency

3.1.1.1.1 Fuel prices, CO2 emissions prices and fuel costs The average fuel prices per type of conventional generating unit and CO2 emissions prices for each decade step examined derive from the Task 2.3 input by ECN. It should be noted that those prices are different for the baseline Business As Usual (BAU) scenario and the remaining scenarios. Table 3 and Table 4 demonstrate the prices for each type of fuel for decades 2020 to 2050 along with CO2 emissions price for the BAU and the rest of the scenarios examined.

Table 3: BAU scenario fuel and CO2 emissions prices

Type 2020 2030 2040 2050 Gas (€/GJ) 8,4 10,8 13,8 16,6 Coal (€/GJ) 3,5 5,8 7,0 8,3 Lignite (€/GJ) 2,6 4,9 6,1 7,4 Biomass (€/GJ) 3,5 5,8 7,0 8,3 Nuclear (€/GJ) 0,06 0,06 0,06 0,06 Oil (€/GJ) 12,0 18,5 21,7 24,8 CO2 (€/t) 15,8 21,6 26,6 30,2

Table 4: Remaining scenarios fuel and CO2 emissions prices

Type 2020 2030 2040 2050 Gas (€/GJ) 8,4 17,0 17,5 18,5 Coal (€/GJ) 3,5 8,2 8,5 8,9 Lignite (€/GJ) 2,6 7,3 7,6 8,0 Biomass (€/GJ) 3,5 8,2 8,5 8,9 Nuclear (€/GJ) 0,06 0,06 0,06 0,06 Oil (€/GJ) 12,0 26,1 26,8 27,5 CO2 (€/t) 15,8 32,4 75,5 86,3

Having determined the efficiency of each type of generating unit and by knowing the fuel price, CO2 amount of emissions and emissions price, the fuel cost per type of unit/technology per country can be calculated by the following equation:

𝑓𝑢𝑒𝑙 𝑐𝑜𝑠𝑡 = 𝑓𝑢𝑒𝑙 𝑝𝑟𝑖𝑐𝑒𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 + 𝐶𝑂2𝑒𝑚𝑖𝑠𝑠𝑖𝑜𝑛𝑠𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑐𝑦 ∙ 𝐶𝑂2𝑝𝑟𝑖𝑐𝑒



9



3.1.1.2 high resolution network model (RWTH)

Some of the analysis that are performed in Task 3.1 require a detailed network model of the European transmission network including the major European load and generation sites with element-sharp data. That means that the most important substations, transmission lines, transformers, generation units and so on need to be included in the model as separate, disaggregated elements. For that purpose a model of the European power transmission grid with several thousand nodes and more than 2500 dispatchable generators was developed. On the one hand the model consists of the basic grid layer containing substations, transmission lines and transformers; on the other hand layers for generation and demand are developed and used. For the purpose of Task 3.1 the model is limited to the former UCTE region. The model will be extended to include the United Kingdom, Ireland and Scandinavia in Task 3.2.

Figure 4: Detailed model of the European transmission network used in T3.1

3.1.1.2.1 Grid model layer The grid layer is based on the UCTE study model that has been geographically referenced in a first step. With given geographical coordinates for all nodes and further information of the edges, the network model can be illustrated as in Figure 4. For transmission lines and transformers, the necessary information for load flow calculations is given by the study

380kV +

220kV

©IFHT



10



model. The information about generation and demand however, turned out to be insufficient for the analysis of future bottlenecks in case of transformation processes as only static values are given. Furthermore no clear differentiation between load and different generation types is available as only one aggregated power figure for each node of the model is provided. Due to this reason the already mentioned layers for generation and demand are integrated in the model what is explained in the following.

3.1.1.2.2 Demand layer Besides the generation, the network load is the other main driver for power flows within the transmission grid. Equal to the generation layer, the spatial distribution and the temporal behaviour of the load are required for an adequate modelling. The temporal behaviour in absolute values is given in the scenario data and has been discussed in previous chapters. Therefore, the spatial distribution needs to be determined, i.e. the assignment of the load to the nodes of the grid model. As frequently discussed in literature a strong correlation between population density distribution and load distribution exists. Figure 5 illustrates how this correlation looks like in the case of the Netherlands. Therefore the population density data is used as load density and allocated to the nodes of the grid model by certain heuristics. Afterwards some manual correction is necessary to avoid overloaded nodes, i.e. nodes that gain a share of the load that is too high to be supplied by the grid. The nodal values are scaled up to the total national load of the country as no further spatial-temporal information is given.

Figure 5: Load distribution based on population density

3.1.1.2.3 Generation layer The generation layer is based on a power plant database as illustrated in Figure 6. Besides existing power plants also planned generation units are included, as well as for example wind farms. In a first step this database is adapted according to the scenario data to create a coherent power plant portfolio for each scenario and decade. For this, based on the age structures of the national power plants and the scenarios, the shut down and extension needs between the decades are determined. In a second step, the extensions are projected

population density real load distribution

TenneT



11



geographical data by certain heuristics. That way a new geographically referenced power plant database is created for each decade and scenario. The power plant entries in the database are afterwards allocated to the nodes of the grid model, which is made possible due to the geographical information contained in the models.

Figure 6: Power plant database

For renewable energy sources like wind or photovoltaic that are usually installed in small but multitudinous units a mapping of each single unit is not reasonable and also not covered by the data basis. Therefore node-related values are identified providing an assumption for the installed capacity of, for instance, wind power or photovoltaic units connected to the nodes. For wind power this information can be derived from the power plant database that includes enough entries to create a reasonable distribution of the installed capacity. These nodal values are scaled to the total national installed capacity given by the scenario data. Only offshore wind power is added manually and treated separately as for this only little information is provided by the database. Figure 7 provides an exemplary overview of the resulting distribution of installed wind power capacities for a given scenario. For photovoltaic and other solar power generation units in Europe the entries in the power plant database are not sufficient to get a coherent picture of the geographical distribution of this source of power. Therefore assumptions based on, for instance, the distribution of the population and of the total solar irradiation are taken. For Germany, currently by far the

Nuclear

Coal

Gas

Wind

Hydro

Other



12



country with the highest installed capacity of photovoltaic, more precise data for the distribution of the installations is available and used for a more precise mapping. The solar power import from northern Africa in the scenario “Desertec” is manually allocated to nodes near Gibraltar (generation “Morocco”) and Sicily (generation “Tunisia”) as agreed in the consortium.

Figure 7: Exemplary: Distribution of installed wind power capacity

3.1.1.2.4 Simulation with the Optimal Power Flow At this stage the main layers grid, generation and demand have been created. For the simulations however, especially for the generation, additional data is needed to determine the operation of power plants. The model is simulated with the AC Optimal Power Flow (OPF). Whereas the standard Power Flow calculation (PF) needs a pre-defined generation, the OPF sets up the operating points of all conventional generators itself. Within the limits of the generators and the boundary conditions of the grid, e.g. voltage limits or the maximal power flow on transformers and lines, the OPF calculates a cost-minimal power generation. Each power plant is modelled with a linear cost function representing its marginal generation costs. The generation costs are hereby defined as the sum of fuel costs and CO2-emission costs. The marginal costs may also refer to other costs of operation, which is not apportioned in the following however. For the modelling of the fuel costs, the efficiency, the CO2 content of the fuel and the CO2-price, figures given in the scenarios are taken as well. Thus, for every power plant the marginal generation costs are determined including the costs for the emission certificates. In addition the absolute CO2 emission can be calculated for each power plant as soon as the power output is identified by the Optimal Power Flow.



13



The OPF-formalism used for this assessment does not take region-based dispatching strategies of the generation units into consideration but assumes a single European market. On the technical side it has to be considered that the calculated operating state does not necessarily be (n-1) secure. Furthermore, no temporal constraints like ramps and minimum operating times of power plants are taken into account, leading to every simulation being regarded as an isolated point in time. However the advantage of making use of the Optimal Power Flow is the integration of grid constraints and it is assumed that in the face of scenario calculations these inaccuracies are passable to get a first estimation of the future utilization of the transmission network. Through various optimization measures for the OPF-problem it was possible to reduce the computing times for a single simulation of the described model to less than one minute. By means of a high performance cluster with parallel computing it is possible to calculate complete years in hourly resolution for each scenario based on the given data.

3.1.1.2.5 Integration of wind and solar power time curves / unsupplied loads

3.1.1.2.6 Grid development All over Europe lines are constructed and planned in the face of current and possibly even future bottlenecks, resulting in a constant development of the European transmission grid. As these lines have a significant impact on the future utilization and development needs of the network, it is inevitable to keep the network model as up to date as possible and to integrate the development measures into the model. A survey of grid development measures has been done to identify planned initiatives. Most of the information can be found in publications of ENTSO-E, for instance in the Ten Year Network Development Plan, or in those of the individual TSOs. These measures are integrated into the network model by identifying the two nodes that are connected by the new line. The technical parameters are taken from the IRENE technology database with knowledge about voltage, number of circuits and the length of the line that can be derived from the position of the nodes and the technology (e.g. AC or DC, cable or overhead line). All measures are marked with a timestamp representing the probable finishing date. The measures are illustrated in Figure 8.



14



Figure 8: Identified and implemented network development measures

3.1.2 Emission pricing to achieve environmental targets and environmental aspects of IRENE-40 scenarios (ECN)

Expected costs of CO2 emissions due to transmission grid losses:

The cost of CO2 emissions due to grid line losses is expected to be relatively low.

3.1.2.1 CO2 emission reduction in the IRENE-40 scenario runs with the COMPETES model

3.1.2.1.1 Short introduction to the COMPETES model Each node is connected to neighboring nodes by a single interconnection per border, where exogenously determined NTC values represent the transmission capacity on a certain interconnection (see Figure 9). The Balkan countries (Montenegro, Macedonia, Croatia, Serbia, Bosnia-Herzegovina and Albania), are included in order to better link Greece to the rest of the 27 European Union countries.



15



Figure 9: The physical representation of the electricity network of the EU27 + 2 and the

Balkan countries as modeled in COMPETES; the dotted lines represent external imports as included in the DESERTEC scenario from the African continent.

3.1.2.1.2 Underlying assumptions and procedures The following dataset was distributed to IRENE-40 partners:

• Fuel prices in €/GJ & CO2 prices in €/ton CO2 determined for BAU scenario and low-carbon scenario’s

• Installed power capacities for EU27 (+ Norway and Switzerland) • Efficiencies of future power plants • Fuel emissions (kg/GJ) • Hourly demand levels for EU27 (+ Norway and Switzerland) • Hourly generation of wind onshore, wind offshore, solar CSP and solar PV per

country



16



• External imports from Africa in the DESERTEC scenario Procedures and inputs that (might) differ between COMPETES and other low resolution models:

• Impact of including Balkan countries on model results • Efficiencies of existing power plants • Pre-stage Hydro Pump Storage and Conventional Hydro Procedure • NTC values; were distributed by ECN to IRENE-40 partners as transmission capacity

starting values, but they are not part of the scenario database • Network topology (see Figure 9)

Special attention is hereby given to:

• Efficiencies

• Hydro Pump Storage & Hydro Conventional Procedure

3.1.2.2 Results of scenario runs with COMPETES

3.1.2.2.1 Average generation costs Figure 10 shows the average generation costs in euro/MWh, calculated by dividing the total generation costs by total MWh of electricity generated in 2050 in the five different IRENE-40 demand and generation scenarios. The total generation costs for the EU27 and the EU27 plus Norway and Switzerland are calculated as the sum of the marginal generation costs per MWh per fuel type times the amount of generated electricity per technology commissioned in a certain year using a certain fueltype. Marginal generation costs are based on the power plant efficiencies, the fuel prices, maintenance costs, the CO2 emissions per fuel and the CO2 prices. Intuitively, different fuel- and CO2 prices as well as the share of renewable and thermal installed capacities in total installed capacity, should result in different levels of average generation costs. Figure 10 can be explained using Table 5, Table 6 and Figure 12. When only shortly scanning Figure 12, one would expect the generation costs, regardless of the magnitude of generation, to be lowest in the scenarios with the highest share of zero and/or low cost generation4

4 Zero cost generation is generation from Solar PV and Solar CSP. Low cost generation is hydro power, wind power, geothermal power and marine (tidal & waves) since the marginal costs of these technologies are only determined by relatively low maintenance costs. Also nuclear power has low generation costs due to the relatively small contribution of fuel costs. Of total renewable generation costs, biomass & waste fired technologies make up the highest share, since waste fired generation has a cost for CO2, and biomass co-firing

in



17



total generation, i.e. the DES and the RES scenario, followed by the EFF scenario having a lower but still significant share of low cost generation in total generation. Intuitively, the CCS and BAU scenario are expected to have the highest average generation costs in euro/MWh since non-zero cost generation is most prominent in these scenarios. It is however difficult to say from Figure 12 whether the CCS or the BAU scenario would result in the highest average generation costs, since CO2 costs play a small role in the CCS scenario, while in the BAU scenario fuel and CO2 costs are significantly lower than in the CCS scenario. In addition, ECN takes into account the age of plants w.r.t. its efficiency. How much is generated by which plant might therefore also explain a part of the relative differences between the scenarios in Figure 10. To be more specific, the share of generation from conventional plants (no CCS technologies) operational before 2020 in total thermal generation in 2050, is highest in the BAU scenario. In this scenario around 30 percent of total thermal generation is generated by conventional plants that were in commission before 2020. For the other four scenarios this percentage is significantly lower5

, leading irrespective of the fuel-, and CO2 costs, and share in low carbon generation, to higher generation costs in the BAU scenario. And so, on the one hand one can expect the average generation costs to be higher in the BAU scenario due to higher share of generation from less efficient thermal plants, and CO2 costs involved, while on the other hand, fuel prices and CO2 prices are significantly higher in the CCS scenario. This could thus result in a balance in the average generation costs in euro/MWh.

In Table 5 the average generation costs per category (renewable, thermal; seperated into CCS and conventional, and nuclear) are shown, calculated by dividing the total sum of inter alia renewable generation costs by total renewable generation. Table 5 will give more insight into the differences in total average generation costs as shown in Table 6. From Table 5 it can be seen that the CCS, RES and DES scenario have significant higher average thermal conventional generation costs compared to the other two scenarios, and especially the BAU scenario. The reason why the CCS scenario has a somewhat higher average thermal conventional generation cost is because conventional generation is almost exclusively generated by expensive Gas Turbines, while in RES and DES less expensive Coal PC, IGCC, CCGT and CHP plants also generate power. In the DES and RES scenario, average thermal conventional generation costs in euro/MWh are relatively high, but due to the high share of low cost generation, the impact of high fuel and CO2 prices on the average generation costs is to some degree lowered. Figure 12 shows that the share of non-zero cost

and biomass standalone have a fuel price. For simplicity, renewables will be referred to in total as low cost generation. 5 RES: 4,1% ; DES: 4,4% ; CCS: 0,36% ; EFF: 3,8%



18



generation in total generation is higher in the EFF scenario than in the DES and RES scenario. Therefore, especially fuel prices (since the share of CCS generation is significant) will play a more prominent role in the EFF scenario than in the RES and DES scenario, resulting into somewhat higher average generation costs in Figure 10. Even though the share in low cost generation is the highest in the RES and DES scenario, back-up power in the form of the third most expensive technology of Gas Turbines needed when low cost generation is not sufficient to meet demand, is significant (see Table 6 and the relative high average thermal conventional generation costs as shown by Table 5). Therefore, average generation costs in euro/MWh in the RES and DES scenario are not significantly lower from the other three scenarios. Since the demand for electricity in the EFF scenario is significantly lower (10 percent increase between 2010 and 2050) than for CCS, DES and RES scenario (60 percent increase between 2010 and 2050), total generation costs as shown in Table 6 are the lowest in the EFF scenario for the simple reason that less has to be generated for supply to meet demand. Furthermore, from Figure 10 a clear difference in the level of average generation costs in euro/MWh can be seen between the RES and DES scenario. This difference can be explained from the assumption that the external imports of solar generation from Africa do not have an intermittent characteristic since these are assumed to be constant hourly imports. Therefore, on an annual basis, more cheap renewable generation is available in the DES scenario compared to the RES scenario (see Figure 12), and the dispatch of cheap generation is more stable than in the RES scenario. Consequently, this will result into fewer extreme hours when significant amounts of expensive peak load power is required because generation from renewables is not sufficient to meet demand. The combined impact in the RES scenario of the higher frequency of zero or close to zero marginal costs, with a higher frequency of high prices due to the dispatch of inter alia expensive gas turbines, results in relatively high annual average prices compared to the DES scenario. The intermittent characteristic of wind and solar generation is especially a problematic issue since these intermittent units are more or less bound to a specific location, e.g. wind installed capacity is especially concentrated in the western part of Europe, while solar installed capacity is more concentrated around the Mediterranean sea. Therefore, when the share of intermittent installed capacity in total installed capacity is significant, in case the wind does not blow or the sun does not shine, the issue for supply to meet demand is more concentrated in certain areas and therefore relying more heavily on the capacity to transmit (relative cheap) generation from one country to another. In case transmission capacity constrains the flow of cheap generation from one node



19



to another, more expensive generation has to subsitute for cheap generation resulting into relatively high average generation costs in euro/MWh. Also in the EFF scenario intermittency has an impact on the average generation costs since conventional generation from gas turbines, installed to operate as back-up power, does have a relatively high share in total generation in the EFF scenario compared to the CCS and especially the BAU scenario. This implies that in some cases in the EFF scenario locally available and imported cheap generation is not sufficient to meet demand, increasing the demand for conventional back-up power. The reason why the average generation costs in euro/MWh are lower for the EU27 +2 countries is because Norway and Switzerland have significant amounts of cheap hydro power, therefore the average generation costs will be lower since the total generation will increase at a higher rate than the total generation costs. In summary, in the CCS scenario where the share of low cost generation in total generation is lowest compared to the other four scenario, significantly high fuel prices will result into the highest average generation costs, followed by the BAU scenario where as opposite to the CCS scenario, also CO2 costs will strongly determine the level of the average generation costs. Since in the RES, DES and EFF scenario the share of low cost generation in total generation is significant, average generation costs are lower compared to the BAU and CCS scenario.

Figure 10: Average generation costs (in euro/MWh) in 2050 for all IRENE-40 scenario’s for

both EU27 and EU27 + Norway and Switzerland.

0

10

20

30

40

50

60

BAU CCS RES DES EFF

EU27

EU27+2



20



Avg. generation cost in €/MWh: BAU RES DES EFF CSS EU27 EU27

+2 EU27

EU27+2

EU27

EU27+2

EU27

EU27+2

EU27

EU27+2

Renewable 22 19 12 11 11 10 11 9 20 18 Thermal conventional 100 100 162 162 162 162 156 156 164 164 Thermal CCS 91 91 106 106 106 106 96 96 106 106 Nuclear 1 1 1 1 1 1 1 1 1 1

Table 5: Overview of the average generation costs in €/MWh calculated by dividing total renewable generation costs and total renewable generation etc.

3.1.2.2.2 CO2 emission comparison per MWh Figure 11 shows somewhat other results concerning the relative differences between the scenarios than Figure 10, since only CO2 emissions determine the relative differences, and the fuels used may not play a part as a consequence of installed CCS technologies. Figure 11 shows that the generation in the BAU scenario is mostly produced by conventional units which use fossil fuels resulting into relative high CO2 emissions per MWh because CCS technologies are not introduced at a high level. Since in the EFF scenario higher shares in total generation are generated by CCS technologies, and in the RES and DES scenario peak load generation is mostly generated by Gas Turbines, the latter two scenarios show significantly higher levels of CO2 emissions. The reason why the RES, DES and EFF scenario are much more in line in Figure 10 is due to the fuel prices playing relatively more significant role in the EFF scenario compared to the RES and DES scenario. As already explained above, in the RES and DES scenario there is a relatively high demand for transmission capacity due to intermittency, resulting in extreme cases where cheap power generation is not able to reach the locations where it is needed; thus CO2 emitting conventional power plants are dispatched instead. In addition, the intermittent character will not only demand for conventional back up power in high wind situations when there is not sufficient transmission capacity to distribute cheap generation across Europe, but might also result in demand for back up power in case there is relatively low generation from intermittent sources in a certain location, increasing the demand for conventional back up generation. The observation of lower CO2 emissions in the DES scenario compared to the RES scenario can be explained because the latter discussed impact of intermittency on the demand for back up power is of less significance in the DES scenario, because part of the total solar generation in Europe does not have an intermittent characteristic (i.e. imports from Morocco and Tunisia).



21



Figure 11: CO2 emissions (ton CO2/MWh) per scenario in 2050, represented for both EU27

and EU27 + Norway and Switzerland.

3.1.2.2.3 Scenario generation mix and overview table Figure 12 shows the share of the generation of certain technologies in 2050 for the five scenarios both for EU27 and EU27 + 2. Gas CCS consists of CCGT CCS and CHP CCS; Gas consists of CCGT (natural gas), CHP gas, GT and Gas Boilers (natural gas) and IC engines (derived gas); Coal CCS consists of IGCC CCS; Coal consists of lignite, coal PC and IGCC; and RES consists of Hydro conventional, Hydro Pump Storage, wind onshore, wind offshore, solar PV, solar CSP, Marine (tidal & waves), geothermal, and biomass and waste fired. Note that oil fired generation which is the most expensive fuel to generate electricity is not used as peak load power in any of the five scenarios. Figure 12 more or less shows the differences between the assumptions in the five scenarios: RES and DES have the highest share in RES technologies (roughly around 80 percent) and are in need of back-up generation (especially gas); CCS scenario has the highest share in CCS technologies (mainly gas CCS). And in the EFF scenario, renewable technologies have the highest share in total generation in 2050 and CCS technologies make up about 15 percent in total generation.

0

0,02

0,04

0,06

0,08

0,1

0,12

0,14

0,16

BAU EFF CCS RES DES

EU27

EU27+2



22



Figure 12: Contribution of generation per technology type (in %) to total generation per IRENE-40 scenario represented for both EU27 and EU27 + Norway and Switzerland.

Table 6 shows an overview of the main results from the 2050 scenario runs with COMPETES. In the runs as represented in this table, especially in the RES and DES scenario, there is quite some curtailment of intermittent renewable generation (mainly wind offshore). In these scenarios, matching renewable supply and demand is highly dependent on the transmission capacity available since renewable generation nodes are located relatively far away from most of the consumption nodes. This issue is even more explicit in the DES scenario, where for example imports from Morocco have to pass Spain and France to reach other countries in the centre of Europe. Consequently, the calculated total theoretical congestion rents6

6 Total congestion rentscenario,interconn.=∑�∑ABS�Nodal Price Ahour- Nodal Price Bhour� *�lowhour)�

are also highest in this scenario run mainly due to higher levels of congestion on the Spain to France interconnection (curtailment of wind generation is about a factor 2.5 higher in Spain in DES scenario compared to RES scenario). The EFF scenario also has a high share in renewable installed capacity. However, congestion rents are lower in the EFF scenario due to the fact that lower demand for electricity will demand less transmission capacity than in a high demand case were cheap supply is highly bounded to

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

EU27 EU27 + 2 EU27 EU27 + 2 EU27 EU27 + 2 EU27 EU27 + 2 EU27 EU27 + 2

RES DES CCS EFF BAU

Nuclear

Gas CCS

Gas

Coal CCS

Coal

RES



23



certain locations. Lowest congestion rents are observed in the BAU scenario since in this scenario the potential of certain countries to produce zero and/or low cost generation which highly depends on the location (e.g. Norway has a high potential to exploit hydro power), is not exploited to a high extent. Instead, supply of electricity is more organized at a local level, where local conventional plants are in most cases able to generate sufficient supply at relatively low cost levels in comparison to the low-carbon scenarios. Since fuel prices and CO2 prices are not that high in the BAU scenario, the need for low or zero cost supply of renewables is less significant in order to minimize generation costs, therefore resulting into low demand for cross-border transmission capacity and consequently in lower levels of congestion. When looking at total generation costs in the EU, as already shortly discussed in section 3.5.1, highest generation costs can be seen in the CCS scenario due to the relatively high share of non-zero costs generation in total generation and high fuel prices. Even though CO2

prices and fuel prices are lower in the BAU compared to the low carbon scenarios, since CCS technologies are not introduced to a relative high extent and the share of generation from less efficient thermal conventional power plants is rather high, total generation costs in the BAU scenario are higher than the EFF, DES and RES scenario. In addition, due to the greater impact of the intermittent character of wind and solar generation in the RES scenario compared to the DES and EFF scenario, there is a need for conventional back up power of the relatively expensive gas turbines resulting into higher total generation costs than in the DES and EFF scenario. To be more specific, the generation cost of gas turbines is roughly €5.7 billion lower in the DES than in the RES scenario. Furthermore, as a consequence of lower demand for generation in the EFF scenario due to a relative efficient use of generation, total generation costs are also low. Intuitively, it is not surprising for the RES and DES scenario to have the highest percentages of curtailed intermittent generation7

7 Curtailed generation is not included in the generation totals as calculated by the COMPETES model.

, since transmission capacity to minimize total generation costs in especially these two scenarios plays a significant role. In the DES scenario the percentage of curtailed intermittent generation is higher than in the RES scenario. This is due to a simple rule of thumb which was used when formulating the scenarios to estimate the NTC values in the different scenarios. In 2050 for each country, the sum of the NTC values on all borders was assumed to be equal to 20% of the total installed generation capacity plus 10% of domestically installed intermittent generation. In case of the DES scenario, there is less domestic intermittent generation in Italy and Spain, therefore resulting in lower NTC values for these countries. In this simple rule of thumb no account has been made of imports



24



from Africa in the DES scenario. These scenario results therefore illustrate the limitations of a too simple rule of thumb. Even though NTC values are increased significantly in 2050 by using the rule of thumb depending on the 2050 situation in each scenario, congestion seems to play a significant role in constraining the cheap renewable generation to be transmitted efficiently. In order to say something on (optimal) transmission investment decisions, data on transmission investment costs are also needed since a copperplate situation, where there are no transmission constraints lowering the cost to the consumer, might also be far from optimal due to significant costly transmission investments.

An option is: Adding CCS to gas turbines in RES and DES scenarios to meet EU CO2 target for 2050.

EC

Contract n°:

TR

EN

/FP7EN

/218903/˝IRE

NE

-40˝ E

UR

OPE

AN

CO

MM

ISSION

D

G-T

RE

N

25

________________________________________________________________________________________________________ A

REV

A – TU

D – EC

N – ET

H Zurich – Im

perial – ICC

S-NTU

A – R

WT

H – A

BB

– Siemens

Table 6: CO

MPETES results show

ing generation per technology, CO

2 emissions, costs of generation, curtailm

ent of intermittent

generation (% of total generation) and congestion rents, per scenario in 2050 using 2050 N

TC values.

2050 BA

U

RE

S D

ES

EFF

CC

S

Generation (TW

h) EU

27 EU

27 + 2

EU27

EU27 +

2 EU

27 EU

27 + 2 EU

27 EU

27 + 2 EU

27 EU

27 + 2

Hydro C

onventional 527.80

736.03 499.67

719.79 499.24

719.35 379.97

600.06 499.51

719.60 H

ydro Pump Storage

89.60 91.89

108.82 111.29

108.82 111.29

55.72 58.18

108.82 111.29

Wind onshore

462.50 488.04

732.36 766.32

718.90 753.24

385.80 419.33

304.99 339.33

Wind offshore

314.14 319.29

631.88 647.37

596.73 612.62

323.85 337.48

203.21 220.60

Solar PV

116.33 116.44

963.31 963.93

460.51 461.13

435.38 436.00

250.35 250.97

Solar CSP

45.1 45.1

125.8 125.8

811.2 811.2

89.6 89.6

76.5 76.5

Rest R

ES 17.9

17.9 56.7

56.7 57.6

57.6 58.1

58.1 59.2

59.2 G

eothermal

15.3 15.3

110.5 110.5

109.3 109.3

81.3 81.3

118.9 118.9

Biom

ass & W

aste Fired 595.7

595.7 489.2

494.0 442.3

446.7 230.4

230.8 464.2

471.7 C

oal PC

78.5 78.5

15.3 15.9

14.8 15.2

12.8 13.2

7.4 7.7

Coal C

CS

64.9 64.9

17.0 17.6

15.3 15.7

236.5 245.9

526.8 538.7

Lignite 151.9

151.9 0.0

0.0 0.0

0.0 0.0

0.0 0.0

0.0 IG

CC

207.6

207.6 24.2

29.2 19.9

23.3 0.0

0.0 0.0

0.0 C

CG

T (natural gas) 227.5

231.1 35.8

36.0 28.6

28.8 41.0

41.0 0.0

0.0 C

CG

T CC

S 174.4

174.4 59.3

59.3 45.6

45.6 152.1

152.2 1111.7

1111.9 C

HP gas

282.7 282.7

9.6 10.2

6.0 6.4

6.0 6.0

0.0 0.0

CH

P CC

S 89.4

89.4 46.4

46.4 39.5

39.5 55.0

55.8 391.0

392.2 G

T and Gas B

oilers 0.1

0.1 219.4

220.0 187.2

187.5 56.8

56.9 23.3

23.4 IC

engines (derived gas) 0.0

0.0 0.0

0.0 0.0

0.0 0.0

0.0 0.0

0.0 O

il fired 0.0

0.0 0.0

0.0 0.0

0.0 0.0

0.0 0.0

0.0 N

uclear energy 1005.5

1005.5 794.5

794.5 799.5

799.5 797.6

797.6 799.5

799.5 Total generation (TW

h/year) 4466.8

4711.7 4939.6

5224.8 4961.0

5244.0 3397.9

3679.5 4945.5

5241.5 C

O2 em

issions (gram/kW

h) 136.4

129.6 31.1

30.3 25.9

25.1 18.0

16.7 6.5

6.1 Total generation costs (€bn/year)

172.9 173.3

108.2 109.7

94.0 95.1

83.6 84.6

264.3 266.2

Total Congestion R

ent EU27+2 &

Balkan (€bn/year)

5.3 20.4

44.0 17.9

6.8 C

urtailment Interm

ittent Generation (%

) 0.02%

0.02%

3.29%

3.15%

4.23%

4.04%

1.28%

1.30%

0.00%

0.00%


EUROPEAN COMMISSION

DG-TREN

Page : 27 of 383

IRENE-40 Task 3.3 Final Report W3ET TN 3004 11/07/2012 Internal partner reference: Filing N° Doc.Type Order N° Rev. N° Date _________________________________________________________________________________________________________

Alstom Grid – TUD – ECN – ETH Zurich – Imperial – ICCS-NTUA – RWTH – ABB - Siemens

3.1.3 General results and comparison of low and high resolutions model (NTUA, RWTH)

This chapter is to discuss and compare general outcomes of the two network model, i.e. the reduced network model, set up by NTUA for the purpose of Task 3.1, and the full resolution network model used by RWTH. Due to the high difference between the model, a comparison is only reasonable at a very high level, which is why only the generation costs and the total CO2-emission related to the total energy consumption are compared.

Figure 13: General result discussion/model comparison: Average generation costs

Figure 13 shows the average generation costs resulting from the simulation of a complete year within a certain scenario and decade in the NTUA and RWTH models. Between 2020 and 2030 the average costs increase in all scenarios what is mainly because of increasing fuel and CO2-trading prices.

2020 low 2020 high 2030 low 2030 high 2040 low 2040 high 2050 low 2050 high0

10

20

30

40

50

60

70

80

90

year, model (low resolution NTUA vs. high resolution RWTH)

Avg

. gen

erat

ion

cost

s [€

/MW

h]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 28 of 383



Figure 14: General result discussion/model comparison: Average CO2-emissions

Figure 13 shows the total CO2-emissions produced by operating power plants related to the total consumption. For the reduced network model it can be observed, that the average emissions reduce strongly until 2050 – except for the mentioned BAU scenario with a significant higher values. Main conclusion: The comparison of the NTUA low resolution network model RWTH high resolution detailed network model is a first estimation what happens if the network is not developed. The result is that a non-developed network will lead to higher generation costs and the fact that the targets regarding emission reduction will not be met. Task 3.2 addresses into more detail as to explaining differences between the model results as reported upon in Deliverable 3.2.

3.1.4 Cross-border flows and impact of emission pricing (NTUA)

For the purpose of Task 3.1, and based on the model described in section 3.1.1.2, two sets of annual simulations for the target years 2020, 2030, 2040 and 2050 for all the 5 scenarios were carried out. In the first simulation set the 2020 NTC values were considered for all the examined scenarios until 2050. In the second simulation set, the forecasted NTC values for years 2030, 2040 and 2050 provided by ECN were taken into account for the corresponding target years. This way the

2020 low 2020 high 2030 low 2030 high 2040 low 2040 high 2050 low 2050 high0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

year, model (low resolution NTUA vs. high resolution RWTH)

Tota

l CO

2-em

issi

ons

[t CO

2/MW

h]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 29 of 383



comparison of relative results between the two sets of simulations, can provide a first estimate of the interconnection reinforcement impact on cross-border flows, annual generation mix production and CO2 emissions.

3.1.4.1 GENERATION MIX

Table 7 to Table 11 display the total energy generation in TWh, the CO2 emissions in kg/KWh, the total generation cost in million € (which includes the operation cost and the cost of CO2 emissions), the percentage of RES curtailment for each examined target year (2020, 2030, 2040, 2050) and the average generation cost in €/MWh for both cases.

Table 7: Generation simulation results for the BAU scenario

Year 2020 2030 2040 2050 NTC value NTC20 NTC20 NTC30 NTC20 NTC40 NTC20 NTC50

Generation Sum (TWh) 3604.49 3977.00 3966.28 4315.77 4288.47 4693.50 4636.36 CO2 (kg/KWh) 0.2708 0.1777 0.1755 0.1751 0.1716 0.1522 0.1411

Total Generation Cost (M€) 72438 111768 110626 149335 145760 190401 182025 RES curtailment 0.0614% 0.1099% 0.0418% 0.2778% 0.0410% 0.7010% 0.0319%

Avg. generation cost (€/MWh) 20.10 28.10 27.89 34.60 33.99 40.57 39.26

Table 8: Generation simulation results for the high RES scenario





Table 9: Generation simulation results for the DESERTEC scenario





Table 10: Generation simulation results for the Efficiency scenario





EUROPEAN COMMISSION

DG-TREN

Page : 30 of 383




Table 11: Generation simulation results for the CCS scenario




Avg. generation cost (€/MWh) 20.10 48.34 48.24 51.18 50.87 52.81 51.96 Figure 15 to Figure 17 summarize the results presented in the above tables regarding the average CO2 emissions in kg/KWh, the total cost in million € and the percentage of RES curtailment for each examined case for the five considered scenarios.

Figure 15: Average CO2 emissions (kg/KWh) for the five examined scenarios

0,0000

0,0500

0,1000

0,1500

0,2000

0,2500

0,3000

2020 2030 NTC20

2030 NTC 30

2040 NTC20

2040 NTC40

2050 NTC20

2050 NTC50

BAU

RES

DES

EFF

CCS


EUROPEAN COMMISSION

DG-TREN

Page : 31 of 383



Figure 16: Total generation cost (million €) for the five examined scenarios

0

20000

40000

60000

80000

100000

120000

140000

160000

180000

200000

220000

240000

260000

280000

2020 2030 NTC20

2030 NTC30

2040 NTC20

2040 NTC40

2050 NTC20

2050 NTC50

BAU

RES

DES

EFF

CCS


EUROPEAN COMMISSION

DG-TREN

Page : 32 of 383



Figure 17: Res curtailment (%) for the five examined scenarios

In Table 12 to Table 16 the percentage share on total generation for each general type of production technology is. presented.

Table 12: Percentage of technology type contribution to total generation, BAU scenario


RES 40.59 46.32 46.48 47.48 47.91 48.53 49.44 Coal 24.43 13.79 13.63 15.62 15.64 13.72 12.95

Coal CCS 0.00 0.21 0.22 0.70 0.77 1.26 1.38 Gas 10.90 14.35 14.23 9.26 8.38 8.27 7.41

Gas CCS 0.00 1.59 1.63 4.49 4.71 6.80 7.13 Nuclear 24.03 23.74 23.81 22.46 22.60 21.42 21.69

Table 13: Percentage of technology type contribution to total generation, high RES scenario


RES 40.59 47.58 47.68 59.29 60.40 72.29 76.53 Coal 24.43 17.60 18.40 6.79 5.35 1.94 1.63

Coal CCS 0.00 0.13 0.14 0.25 0.34 0.27 0.31 Gas 10.90 14.00 13.02 12.21 12.05 8.28 4.50

Gas CCS 0.00 1.61 1.65 4.40 4.76 2.28 2.11

0%

1%

2%

3%

4%

5%

6%

7%

2020 2030 NTC20

2030 NTC30

2040 NTC20

2040 NTC40

2050 NTC20

2050 NTC50

BAU

RES

DES

EFF

CCS


EUROPEAN COMMISSION

DG-TREN

Page : 33 of 383



Nuclear 24.03 19.08 19.12 17.06 17.11 14.95 14.92

Table 14: Percentage of technology type contribution to total generation, DESERTEC scenario


RES 40.59 55.21 55.38 66.41 67.85 73.57 78.23 Coal 24.43 14.71 14.84 5.76 3.91 1.80 1.54

Coal CCS 0.00 0.13 0.13 0.22 0.30 0.23 0.28 Gas 10.90 9.68 9.28 7.81 7.39 8.06 3.55

Gas CCS 0.00 1.30 1.35 3.09 3.79 1.63 1.70 Nuclear 24.03 18.97 19.01 16.70 16.76 14.72 14.69

Table 15: Percentage of technology type contribution to total generation, Efficiency scenario


RES 40.59 43.83 43.92 51.19 51.86 60.66 62.69 Coal 24.43 20.07 20.59 6.48 5.38 0.03 0.03

Coal CCS 0.00 0.51 0.53 3.02 3.17 6.19 6.29 Gas 10.90 11.82 11.14 7.88 7.72 5.51 3.43

Gas CCS 0.00 3.03 3.05 10.36 10.79 6.40 6.03 Nuclear 24.03 20.70 20.73 21.02 21.04 21.21 21.53

Table 16: Percentage of technology type contribution to total generation, CCS scenario


RES 40.59 40.04 40.00 41.13 40.78 42.96 42.78 Coal 24.43 8.09 7.96 2.46 1.44 1.08 0.46

Coal CCS 0.00 5.54 5.63 8.14 8.44 10.18 10.65 Gas 10.90 0.18 0.07 0.23 0.02 1.40 0.21

Gas CCS 0.00 27.02 27.18 30.88 32.26 28.75 30.36 Nuclear 24.03 19.10 19.11 17.13 17.05 15.62 15.53


EUROPEAN COMMISSION

DG-TREN

Page : 34 of 383



Figure 18: Graphical representation of each technology type contribution to total generation for all

scenarios with fixed NTC 2020 values

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2020 2030 BAU

2030 RES

2030 DES

2030 EFF

2030 CCS

2040 BAU

2040 RES

2040 DES

2040 EFF

2040 CCS

2050 BAU

2050 RES

2050 DES

2050 EFF

2050 CCS

Nuclear

Gas CCS

Gas

Coal CCS

Coal

RES


EUROPEAN COMMISSION

DG-TREN

Page : 35 of 383



Figure 19: Graphical representation of each technology type contribution to total generation for all

scenarios with the forecasted NTC values for years 2030, 2040 and 2050

3.1.4.2 CROSS-BORDER FLOWS

From the results it is obvious for all five scenarios that interconnections reinforcements are necessary in order to reduce overall system generation cost, RES curtailment and CO2 emissions. It should be noted that the overall generation cost is expected to be higher if reserves requirements and technical minimum of the generators are considered. The extra capacity that should be installed regarding the interconnections is very much influenced by the targets and the data inputs of each scenario. For the RES scenario, if the NTC 2020 values are extrapolated to 2050, there is 31.5

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2020 2030 BAU

2030 RES

2030 DES

2030 EFF

2030 CCS

2040 BAU

2040 RES

2040 DES

2040 EFF

2040 CCS

2050 BAU

2050 RES

2050 DES

2050 EFF

2050 CCS

Nuclear

Gas CCS

Gas

Coal CCS

Coal

RES


EUROPEAN COMMISSION

DG-TREN

Page : 36 of 383



billion Euros reduction in system cost (without considering any generation cost for hydro units) and 5.13% reduction in RES curtailment. Of course, in real power system simulation this reduction should be higher since in our approach no congestions inside each country’s network were taken into account. On the other hand, in CCS scenario the cost reduction for 2050 is only 8.7 billion Euros since generation production is less influenced by network congestions and overall CO2 emissions are the smallest among all the scenarios. Figure 20 to Figure 22 display the number of total congestion hours (counted both ways) for the 71 interconnections between the examined European countries as described in paragraph 3.1.1 for the years 2030, 2040 and 2050 in the case where the 2020 fixed NTC values are considered. Some general observations are presented below:

• The interconnections connecting the Scandinavian countries with the continental Europe are congested many hours of the years for all the scenarios due to their high renewables installed capacity. The installed pumping hydro in these countries can alleviate RES curtailment in the rest of the Europe if these interconnections are strengthened.

• Germany plays a significant role in all the scenarios since it is very much dependent on the capability of its interconnections to export its high renewables production or cover its high demand when renewables production is relative low. This also results to high congestions at the interconnections of its neighbouring countries (Poland, Netherlands, etc).

• UK is a net exporter in all the scenarios and its interconnections are highly used. The same applies for Portugal, Spain and France.

• In the DES scenarios, the CSP production is transferred to Central Europe through Italy’s and Spain’s interconnections, which causes congestions problems throughout all continental Europe .

• Another obvious hotspot is the Poland-Lithuania interconnection which’s capacity is not sufficient for all the scenarios for almost all the hours of the years.

• In RES scenario many hours of congestion occur in the Balkan region due to the fact that Greece becomes a net exporter and its renewables production affects the flows throughout that entire region.

Overall in the RES and DES scenarios, there is a need to strengthen the North ↔ South backbone of Europe’s interconnections. In BAU scenario, the occurred congestions in 2050 are very much alike with 2020 congestions while in EFF scenario there is a tendency for less North ↔ South flows. In CCS scenario, the congested hours are much less from RES, DES and EFF scenario although there is increased use of the Scandinavian and Baltic countries interconnections.


EUROPEAN COMMISSION

DG-TREN

Page : 37 of 383



0 1000 2000 3000 4000 5000 6000 7000 8000 9000

AT-CZ AT-DE AT-HU

AT-IT AT-SI

AT-CH BE-FR BE-DE BE-NL BE-UK BG-GR BG-RO BG-MK BG-RS CZ-DE CZ-PL CZ-SK

DK-DE DK-NL DK-SE

DK-NO EE-FI

EE - LV EE-LT FI-SE

FI-NO FR-DE FR-IT FR-ES FR-UK FR-CH DE-NL DE-PL DE-SE DE-CH DE-NO

GR-IT GR-AL

GR-MK HU-RO HU-SK HU-SI

HU-HR HU-RS IE-UK

IT-SI IT-CH IT-AL IT-HR IT-ME LV-LT LT-PL LT-SE

NL-UK NL-NO

PL-SK PL-SE PT-ES

RO-RS SI-HR

SE-NO UK-NO AL-ME AL-MK AL-RS

BA-HR BA-ME BA-RS HR-RS ME-RS MK-RS

BAU RES DES EFF CCS


EUROPEAN COMMISSION

DG-TREN

Page : 38 of 383



Figure 20: Congestion hours between the 71 interconnection for 2030, fixed 2020 NTC values


EUROPEAN COMMISSION

DG-TREN

Page : 39 of 383



0 1000 2000 3000 4000 5000 6000 7000 8000 9000

AT-CZ AT-DE AT-HU

AT-IT AT-SI


DK-DE DK-NL DK-SE

DK-NO EE-FI

EE - LV EE-LT FI-SE


GR-IT GR-AL


HU-HR HU-RS IE-UK


NL-UK NL-NO

PL-SK PL-SE PT-ES

RO-RS SI-HR



BAU RES DES EFF CCS


EUROPEAN COMMISSION

DG-TREN

Page : 40 of 383





EUROPEAN COMMISSION

DG-TREN

Page : 41 of 383



0 1000 2000 3000 4000 5000 6000 7000 8000 9000

AT-CZ AT-DE AT-HU

AT-IT AT-SI


DK-DE DK-NL DK-SE

DK-NO EE-FI

EE - LV EE-LT FI-SE


GR-IT GR-AL


HU-HR HU-RS IE-UK


NL-UK NL-NO

PL-SK PL-SE PT-ES

RO-RS SI-HR



BAU RES DES EFF CCS


EUROPEAN COMMISSION

DG-TREN

Page : 42 of 383




3.1.4.3 impact of emission pricing

In order to investigate the impact of CO2 emission pricing in the future European grid within the purpose of IRENE-40 we decided to carry out further simulations on the low resolution model by modifying the CO2 emission price. Therefore four further cases were considered for the high-RES, DESERTEC, Efficiency and CCS scenarios where the future prices of CO2 over the years 2030 to 2050 were lower than the price specified in Table 17, while fuel prices were considered the same, representing a less motivated to heavily charge greenhouse emissions EU. Investigating the BAU scenario with lower CO2 prices was not considered necessary since it is obvious that the simulation results would deviate even more from the initial EU emissions targets which were initially the goal of the formulated scenarios. Table 17 sums up the fuel costs per type of technology along with the initial CO2 prices proposed for the IRENE-40 high-RES, DESERTEC, Efficiency and CCS scenarios and the modified CO2 prices used for the scope of this study case. It should be noted that the generation scenario installed capacity was based on the CO2 prices presented in the previous section; thus it is reasonable to expect changes in the generation mix output of the scenarios as well as the congestion hours for the 71 examined interconnections.

Table 17: Fuel and CO2 prices considered in the additional study cases

Type 2020 2030 2040 2050 Gas (€/GJ) 8,4 17,0 17,5 18,5 Coal (€/GJ) 3,5 8,2 8,5 8,9

Lignite (€/GJ) 2,6 7,3 7,6 8,0 Biomass (€/GJ) 3,5 8,2 8,5 8,9 Nuclear (€/GJ) 0,06 0,06 0,06 0,06

Oil (€/GJ) 12,0 26,1 26,8 27,5 Initial CO2 (€/t) 15,8 32,4 75,5 86,3 New CO2 (€/t) 15,8 26 35 42

Table 18 to Table 21 display the total energy generation in TWh, the CO2 emissions in kg/KWh, the total generation cost in million € (which includes energy generation cost and CO2 cost), the percentage of RES curtailment for the examined target years (2030, 2040, 2050) and the average generation cost in €/MWh for both cases of fixed and forecasted NTC values for the additional lower CO2 prices test cases.

Table 18: Generation simulation results for the high RES scenario with lower CO2 prices

Year 2030 2040 2050 NTC value NTC20 NTC30 NTC20 NTC40 NTC20 NTC50

Generation Sum (TWh) 4124.57 4117.09 4600.81 4573.92 5331.97 5356.90 CO2 (kg/KWh) 0.2184 0.2225 0.1267 0.1147 0.0540 0.0342


EUROPEAN COMMISSION

DG-TREN

Page : 43 of 383



Total Generation Cost (M€) 173997 171864 159015 151957 139711 111868 RES curtailment 0.0755% 0.0178% 0.8401% 0.0827% 6.8220% 1.6848%

Avg. Generation Cost (€/MWh) 42.19 41.74 34.56 33.22 26.20 20.88

Table 19: Generation simulation results for the DESERTEC scenario with lower CO2 prices





Table 20: Generation simulation results for the Efficiency scenario with lower CO2 prices





Table 21: Generation simulation results for the CCS scenario with lower CO2 prices





Figure 23 to Figure 25 summarize the results presented in the above tables regarding the average CO2 emissions in kg/KWh, the total cost in million € and the percentage of RES curtailment for each additional case for the four examined scenarios.


EUROPEAN COMMISSION

DG-TREN

Page : 44 of 383



Figure 23: Average CO2 emissions (kg/KWh) for the additional lower CO2 prices scenarios

0,0000

0,0500

0,1000

0,1500

0,2000

0,2500

0,3000

2030 NTC20

2030 NTC30

2040 NTC20

2040 NTC40

2050 NTC20

2050 NTC50

RES

DES

EFF

CCS


EUROPEAN COMMISSION

DG-TREN

Page : 45 of 383



Figure 24: Total generation cost (million €) for the additional lower CO2 prices scenarios

0

20000

40000

60000

80000

100000

120000

140000

160000

180000

200000

220000

240000

260000

280000

2030 NTC20

2030 NTC30

2040 NTC20

2040 NTC40

2050 NTC20

2050 NTC50

RES

DES

EFF

CCS


EUROPEAN COMMISSION

DG-TREN

Page : 46 of 383



Figure 25: RES curtailment (%) for the additional lower CO2 prices scenarios

By comparing Table 7 to Table 11 with Table 18 to Table 21, it is shown that the generation mix is affected by the change in CO2 emission prices since a part of CCS technologies output is replaced by conventional thermal generation output, especially until 2040 when the non-CCS thermal generation cost with the new lower CO2 prices is comparable with the CCS generation cost. The same applies for the actual emissions in kg/KWh and the total generation cost in million €. It should be noted that CO2 prices reduction is not a simple cost saving measure. With introduction of CO2 pricing, generators will charge more for their electricity even if they do not have to pay for the emission permits. A lower CO2 price means a stronger reliance on alternative means to achieve CO2 reduction targets by 2050. These constitute a cost to society which is likely to be of the same magnitude as the cost reduction to consumers due to lower CO2 trading prices. Therefore a direct comparison between the total cost of generation between the high and the reduced CO2 prices scenarios is off the scope of the initial scenarios. Regarding CO2 emissions and fuel costs, Table 22 and Table 23 demonstrate the increase of total CO2 emissions and the decrease of fuel costs due to the reduced CO2 prices and the change in the final fuel mix. The results are provided both in absolute and as a percentage compared to the initial high price CO2 emissions results for the four additional study cases. The increase in CO2 emissions and fuel costs is major for year 2030 for the CCS scenario since until then conventional thermal units are preferred instead of CCS units, resulting in higher emissions, but it significantly lowers

0%

1%

2%

3%

4%

5%

6%

7%

2030 NTC20

2030 NTC30

2040 NTC20

2040 NTC40

2050 NTC20

2050 NTC50

RES

DES

EFF

CCS


EUROPEAN COMMISSION

DG-TREN

Page : 47 of 383



over the next decades where CCS units get to replace the old thermal plants. For the rest of the scenarios there is a raise on average CO2 emissions for the year 2040 which again seems to become much smaller for the year 2050. The 30% emission increase in the efficiency scenario between the NTC 2030 and the NTC 2040 cases when comparing the high and low emissions price results is caused by the fact that roughly 130TWh of lignite and hard coal plants participated in the final generation mix instead of the less pollutant CCGT and CCGT CCS plants causing as well an increase of the total emissions between 2030 and 2040. Those results underline even more the necessity of proper evaluation of CO2 prices in the future since important deviations can occur when future scenarios are considered under economical and environmental approaches.

Table 22: Increase of total CO2 emissions due to lower CO2 prices, in kg/KWh and % for the four additional study cases

Scenario 2030 2040 2050 NTC20 NTC30 NTC20 NTC40 NTC20 NTC50

RES 0.0033 0.0041 0.0214 0.0245 0.00040 0.0007 1.5186% 1.8699% 20.3038% 27.8633% 0.7409% 2.2117%

DES 0.0042 0.0052 0.0162 0.0168 0.00027 0.0006 2.4352% 3.0545% 20.1310% 27.8642% 0.5232% 2.1757%

EFF 0.0613 0.0076 0.0231 0.0246 0.0000 0.0000 26.9062% 3.3082% 26.8286% 33.0204% 0.1211% 0.1251%

CCS 0.0823 0.0889 0.0023 0.0020 0.0000 0.0000 98.4002% 108.4045% 9.1527% 14.2841% 0.2068% 0.6242%

Table 23: Decrease of total generation fuel cost due to lower CO2 prices, in M€ and % for the four additional study cases

Scenario 2030 2040 2050 NTC20 NTC30 NTC20 NTC40 NTC20 NTC50

RES 460 561 6694 9130 50 119 0.3043% 0.3776% 4.6066% 6.3968% 0.0389% 0.1139%

DES 1000 1105 4783 6731 53 134 0.8236% 0.9244% 4.1862% 6.1077% 0.0440% 0.1416%

EFF 835 946 6848 8399 151 274 0.5903% 0.6771% 5.4374% 6.7187% 0.1577% 0.3187%

CCS 10842 11716 6603 9832 429 887 5.7656% 6.2395% 2.9163% 4.2769% 0.1633% 0.3346%


EUROPEAN COMMISSION

DG-TREN

Page : 48 of 383



3.1.5 Identification of ecological weak network points (RWTH)

3.1.5.1 Definition and methology

As discussed in the previous chapters, one main goal for the future of the European energy supply is the reduction of the CO2-emissions. Therefore, a strong development of renewable energy sources takes place substituting fossil generation units. Usually these units are volatile in power infeed, remote to the load centers and regionally highly concentrated like especially offshore wind farms in the North Sea or solar power plants in southern regions. Due to this reason the transmission system is facing the challenge to integrate the renewable generation and transmit it to the more distant load centres as compared to average past conventional generation placement. Without extension measures in the network the transmission is not possible, the power has to be curtailed and replaced by other, emission-burdened generation. In the worst case, loads cannot be supplied due to bottlenecks in the network. Because of the motivation in the following, ecological weak network points are defined as point or regions in the grid where a substantial amount of energy from renewable sources remains unused as the network is unable to completely transmit it to the load centers. As described previously, for the simulations the high resolution model is used and already known network extensions up to 2020 are incorporated in the model. No further changes in the network are assumed and therefore the identification of the ecological weak network points is meant to be a first step to identify necessary grid development measures for later investigation like for example performed in Task 3.2. As described in the previous chapter, infeed from renewables is modeled as a dispatchable generator providing power at zero costs. Curtailment therefore takes place if the Optimal Power Flow does not completely utilize the power of such a generator. Besides curtailed power from RES, also unsupplied loads are identified. Therefore all loads are modeled as dispatchable but belayed with a high penalty factor. So only in case of technical necessity to dispatch the load this is simulated. The simulations are then performed as described, for every scenario and decade a complete year in hourly resolution is simulated with the optimal power flow. For the identification of the ecological weak network points, for every node the unused energy from renewable energy sources and the unsupplied loads are summed up and illustrated on a map. The results within the different scenarios are presented and discussed in the following.

3.1.5.2 Illustration and discussion of REsults

2020 The curtailment of renewable energy is rather low in 2010 and 2020: it can be stated that no load is unsupplied in these years. The visualisation of these results therefore only shows some minor


EUROPEAN COMMISSION

DG-TREN

Page : 49 of 383



ecological weak network points. The positive peaks with colours green-yellow-red indicate curtailed energy from renewable energy sources. Unsupplied load will be illustrated as negative peaks in blue. It has to be noted that the height of the peaks is proportional to the yearly average values of the curtailed or not supplied energies. Therefore, no information about the temporal correlation of the different weak network points is given in the figures. Furthermore it is possible that local compensation exists, i.e. that at one note renewables are curtailed in some hours and the load cannot be supplied in other hours. In this case, only the difference of the values is illustrated in the figures.

Figure 26: Ecological weak network point analysis 2020


EUROPEAN COMMISSION

DG-TREN

Page : 50 of 383



2030


In 2030 the different scenarios begin to differ. The results of the ecological weak network point analysis in the scenarios BAU, CCS, Efficiency and Renewables are illustrated in Figure 27. The Desertec scenario is discussed separately later in this chapter. In the scenarios BAU, CCS and Efficiency only small changes compared to 2020 are visible due to the fact that in these scenarios the installed capacities of the renewables increase with a lower rate than in the Renewables scenario. In that scenario the total amount of curtailed energy from renewable energy sources is higher compared to the other scenarios as already observed previously. The spatial distribution however equals those of the other scenarios. Generally it shows up that the North Sea shore is an ecological weak network point. The grid is not able to integrate all the power from the offshore power plants. In Spain and Portugal ecological weak network points with less average curtailed energy but uniform distribution appear. Furthermore spots with loads that cannot be supplied exist

BAU

RES

CCS

EFF


EUROPEAN COMMISSION

DG-TREN

Page : 51 of 383



in Spain. Only in the Efficiency scenario, the scenario with a considerable lower demand, all loads can be supplied. 2040


For 2040 the number of nodes that cannot be sufficiently supplied increases in all scenarios, only in the Efficiency scenario still all loads can be supplied. Especially the CCS scenario with its high demand with no DSM leads to unsupplied loads. The curtailed energy from renewable energy sources especially increases in the Renewable scenario. The ecological weak network points in Spain and at the North Sea shore evolve further like already described and some minor sports appear. 2050

BAU

RES

CCS

EFF


EUROPEAN COMMISSION

DG-TREN

Page : 52 of 383



For the year 2050 the results are analysed in detail in the following as then the ecological weak network points and the development needs emerge most clearly.

Figure 29: Ecological weak network point analysis 2050, BAU

Figure 30: Ecological weak network point analysis 2050, CCS

Compared to the other scenarios, the ecological weak network points in the BAU and CCS scenarios are least significant. This is due to the fact that the generation in those scenarios is more centralized


EUROPEAN COMMISSION

DG-TREN

Page : 53 of 383



and close to the load centers as in the other scenarios with a high share of renewables. As this is similar to the current situation, the existing grid is still good suited for this generation pattern. On the other hand a high amount of unsupplied loads can be observed due to the reason that both scenarios, BAU and CCS, include an increase in demand without DSM. In fact, CCS is the scenario with the highest peak demand as it combines the highest energy demand – like Renewables and Desertec – with the absence of DSM. The high demand causes load that are too high to be supplied.

Figure 31: Ecological weak network point analysis 2050, Efficiency (EFF)

Figure 32: Ecological weak network point analysis 2050, Renewables (RES)


EUROPEAN COMMISSION

DG-TREN

Page : 54 of 383



The Efficiency scenario shows the already discussed ecological weak network points at the North Sea shore and in Spain. Whereas in the simulation of 2040 no unsupplied loads are visible, in 2050 unsupplied loads especially exist in Poland. The most conspicuous results emerge in the scenario Renewables 2050. At the North Sea shore, especially in the Netherlands, Germany and Belgium the peaks of the unused energy from renewables are clearly visible. The highly concentrated installed capacity of about 95GW offshore wind power in this scenario and in those three countries cannot be integrated and transmitted to the load centres by the grid. Therefore the North Sea shore is an ecological weak network point facing the development of the offshore wind farms. Furthermore it can be stated that the 2020 grid (2010 grid with network extensions) can take up about 25GW at the North Sea shore. This however is just estimation as the figure depends on load, the spatial distribution of the wind power and furthermore it has to be considered that the links to Great Britain and Scandinavia are missing. Nevertheless it gives an indication for the transport capacity needed for this hotspot. The second weak network point that is mentioned here is Southern Europe and especially the Iberian Peninsula. For this region the scenario data purports a large share of solar power, photovoltaic (PV) and concentrated solar power (CSP). In addition, Spain has 20GW installed capacity in wind power onshore already in 2010 what increases to about 76GW in 2050. In comparison to the offshore wind power, this power is highly decentralized and distributed. Therefore, the ecological weak network point in southern Europe does not appear as striking peak put as flat and distributed. For the analysis of unsupplied load it turns out that especially in southern Poland grid development is necessary to supply all consumers. Obviously the uneven spatial distribution of generation and demand causes this problem. It has to be considered that particularly the distribution of the generation depends on assumptions and projections, but nevertheless grid a development need can be stated and is also confirmed by current discussions about network extension needs, for example in ENTSO-E’s Ten Year Network Development Plan (cf. grid development in Poland illustrated in Figure 10). Desertec scenario The Desertec scenario is discussed separately as this scenario can only take place with strong network extension measures, most probably a HVDC Overlay grid as also proposed by the Desertec consortium. The ecological weak network point analysis however focusses on a slightly developed network that is not able to integrate the power imported from Northern Africa. Thus, in Figure 33 striking peaks are visible at the connection points. Aside these peaks, the results in the Desertec scenario are similar to those of the Renewables scenario in terms of the ecological weak network point at the North Sea shore and the unsupplied loads in Poland. This is caused by the fact that the Desertec and the Renewables scenario do only differ in the installed capacities of photovoltaic and CSP (concentrated solar power) whereas wind offshore and the load remains equal.


EUROPEAN COMMISSION

DG-TREN

Page : 55 of 383



Figure 33: Ecological weak network point analysis Desertec (DES)

2030 2040


EUROPEAN COMMISSION

DG-TREN

Page : 56 of 383



3.1.6 Assessment of grid development measures (RWTH)

3.1.6.1 Approach

As discussed previously, the identification of the ecological weak network points can be used as indicator for future development needs. Different technological solutions addressing these development needs have to be identified and assessed with respect to their ecologic and also economic benefits. In the following a methodology will be presented and applied to three different extension measures allowing such a multi-criteria assessment. As positive ecologic impact of a new transmission line a reduction of the total CO2-emission is considered. This reduction can be obtained if the transmission capacity allows an increased infeed from generation units with zero or low CO2-emissions, especially renewable energy sources, and a decreased infeed from high emission power plants. To determine the change in generation pattern, all simulations are performed both with and without the line included in the network model. That way, the reduction in emissions, unused energy from renewable energy sources and generation costs caused by the network development can be derived. In the following three different development measures are assessed. Each measures contains a 380kV AC double circuit line. The limitation on three single lines is due to the high computational effort that is needed for the simulations as for each line the simulations have to be performed again. Furthermore it allows getting first results for the impact of single line and a validation for the methodology. In the following tasks the results and the methodology are going to be used for the assessment of different network scenarios. The three lines address urgent development needs in the network. As shown in the previous chapter, a high need for network extension at the German North Sea shore exists due to the infeed from the offshore wind farms. This is addressed with a line connecting the 380kV substation ‘Diele’, where a many wind farms will be connected, with the load centres of western Germany at the substation “Niederrhein”. This line has been mentioned in the first German “dena”-study and is planned to be built until 2015. The second and third line is addressing the need for interconnector development. One line is crossing the border between Spain and France, the other the border between France and Germany. Both lines are motivated by the fact that Germany and Spain show – already in these days – a high share of renewable energy sources whereas the France energy supply is based on nuclear power. In times of a high infeed from renewables the lines are therefore used to export the power from Spain and Germany whereas in times of low infeed from renewables the lines are utilized for in import – especially from France – into those countries. The need for development of the interconnectors between Spain and France is also frequently discussed. The interconnectors between France and Germany furthermore showed a high utilisation in preliminary investigations.


EUROPEAN COMMISSION

DG-TREN

Page : 57 of 383



To avoid cannibalization effects, the simulations with and without the line under consideration are always carried out under the implementation of the other two lines. In fact, all three lines are already included in the extension measures illustrated in Figure 34. Besides the simulation of the grid with all extension measures, three additional simulations are performed where in each the respectively line is removed. The three extension measures are illustrated in Figure 34.

Figure 34: Extension measures selected for analysis

For the further assessment on the footprint of each line, technology-specific parameters are taken from the IRENE-40 Technology Database. Herewith it is possible to estimate the investment costs as well as the maintenance costs incurring annually. Table 24 presents estimates for these economic characteristics for each considered measure. The discounted present values of the given cash flows have been calculated for the base year 2010 by taking an interest rate of 5% into account.

Table 24: Cost parameters of selected network development measures

ES

FR

FR

DEBE


EUROPEAN COMMISSION

DG-TREN

Page : 58 of 383



Line Length [km] Investment [€] Annual maintenance [€] Present value [€] Diele-Niederrhein 238 440 300 000 47 650 441 117 630

Spain-France 236 436 600 000 47 650 437 417 630 France-Germany 342 632 700 000 47 650 633 517 630

The total economic footprint can now be determined by relating the present values to the savings in power generation to be realised throughout the time scope until 2050. Furthermore the CO2-emissions that are needed for the production of the material and the construction of an overhead line are given in Table 25. The figures are based on the Cigré report “Life cycle assessment for overhead lines” [9] and are included in the IRENE-40 technology database.

Table 25: CO2-emissions resulting from construction of selected network development measures

Line Length [km] Emissions while constr. [tCO2] Diele-Niederrhein 238 10 23

Spain-France 236 1 015 France-Germany 342 1 471

3.1.6.2 Discussion of results

In the following section the impact of the three selected extension measured are analysed in detail. The indicators that are discussed are the additional infeed of renewable energy sources enabled by the measures and the changes in CO2-emissions and generation costs caused by these measures. Additional infeed of renewable energy sources As discussed, future bottlenecks in the grid can prevent the infeed of renewable energy sources. Therefore additional transmission capacity created by network extension can increase the capability of the grid to integrate the renewable energy. The differences between the integrated energy from renewables in simulation with and without each of the analysed extension measures are therefore discussed.


EUROPEAN COMMISSION

DG-TREN

Page : 59 of 383



Figure 35: Additional infeed of RES caused by line “Diele-Niederrhein”

The network extension measure “Diele-Niederrhein” in Germany directly addresses the ecological weak network point at the North Sea shore. As a starting point, Figure 35 presents the additional infeed from renewable energy sources, which can be anticipated based on the scenario-related development for the different decades. First of all it can be observed that in all scenarios a steady increase of the additional renewable energy takes place. In 2020 hardly any difference shown for the inclusion of the additional line although in this year ecological weak network points exists in the North Sea region (cf. Figure 26). Obviously this is not the case at the specific node, so the extension does not allow additional infeed. Here it has to be considered that no (n-1)-security as well as no special stability constraints besides thermal line rating and voltage band are applied. These constraints would reduce the capacity of the existing grid and thereby make the extension necessary. In all 2030 scenarios more renewable energy can feed into the grid due to the extension measure. The CCS scenario shows the least increase as the installed capacity of offshore wind farms – being the main driver of the development needs in this region – is lower than in all other scenarios. On the other hand, the scenarios Renewables and Desertec lead to the largest increase as here the installed offshore wind power is the highest. As mentioned, those scenarios do only differ in the installed capacities of photovoltaic and CSP – not in wind power. In the following decades the additionally usable energy from renewable sources increases constantly with the also increasing peaks in the ecological weak network point analysis. In the 2050 scenarios Desertec and Renewables the network extension allows to use approximately 9 TWh/a electrical energy additionally. This is equal to a constant power infeed of about 1 GW.

2020 2030 2040 20500

1

2

3

4

5

6

7

8

9

10

year

∆ E

RE

S [TW

h/a]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 60 of 383



Figure 36: Additional infeed of RES caused by line “Spain-France”

The additionally used renewable energy caused by the extension between Spain and France is lower compared to the previously discussed line as here the ecological weak network point is more distributed and can therefore not be addressed directly. Nevertheless, the new connection line does have a significant impact on the infeed of renewables, as shown in Figure 36. Especially in the high renewables scenarios Desertec, Efficiecy and Renewables, a substantial amount of carbon-free electrical energy in enabled to feed into the grid. About 1.2 TWh of annually generated energy can be substituted by means of renewables in 2050 which is much less compared with “Diele – Niederrhein”, but nevertheless remarkable as no concentrated ecological weak network point is adressed. In the CCS scenario on the other hand, the additionally used energy from renewables is low and even decreasing over the decades. It was already discussed in the ecological weak network point analysis that in this scenario no specific ecological weak network point exists in southern Europe.

2020 2030 2040 20500

0.2

0.4

0.6

0.8

1

1.2

1.4

year

∆ E

RE

S [T

Wh/

a]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 61 of 383



Figure 37: Additional infeed of RES caused by line “France-Germany”

In the region of the French-German border no specific ecological weak network point was identified, nevertheless a high utilisation of this interconnector takes place. Thus, assessing the differences in the renewable energy it turns out that the potential of the extension between France and Germany in this concern is much lower compared to the two other network extensions. In some scenarios it is even negative. Only in the Desertec scenario, which leads to flows from southern to northern Europe, a discussable amount of renewable energy can be used. Change of CO2-emissions caused by grid development measures The main aim of the European energy policy is the reduction of CO2-emissions. The previously discussed integration of renewable energy sources is an important step to this goal but also other solutions exist like for example CCS. Therefore a network extension measure may have a positive ecological impact although it does not allow an increase of the infeed of renewable energy sources if a different low emission technology unit is enabled to operate. On the other hand an extension measure might lead to a high additional infeed from RES but only to a low CO2-emission reduction. Thus, in the following the changes in CO2-emission caused by the different extension measured are analysed. For this purpose the total CO2-emission produced by all generation units are determined in the different OPF-simulations with and without the specific extension measures. As the extension measure usually allows a different – in most cases cheaper – operation of the power plant portfolio also different total CO2-emission arise leading to the differences that are discussed in this section.

2020 2030 2040 2050-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

year

∆ E

RE

S [T

Wh/

a]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 62 of 383



Figure 38: Changes in CO2-emissions caused by line “Diele-Niederrhein”

Figure 38 showing the changes in generation’s total CO2-emissions caused by the network extension measure “Diele-Niederrhein” is similar to Figure 35 illustrating the additional infeed from renewable sources caused by this line. This fact indicates again the high importance of this line for the infeed of renewables that directly leads to a reduction of CO2-emissions. Only the bars for the BAU scenario are in relation to the other bars a little bit higher that in Figure 35 what is due to the fact that the average CO2-emissions are the highest in this scenario so that the additional infeed lead to comparably higher CO2-savings.

2020 2030 2040 20500

1

2

3

4

5

6

year

∆ C

O2

[Mt C

O2/

a]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 63 of 383



Figure 39: Changes in CO2-emissions caused by line “Spain-France”

The changes of the CO2-emissions caused by “Spain-France” are highest in 2020 and decrease until 2030. Whereas the in the BAU scenario a constant decrease until 2050 is visible, the other scenarios increase after between 2030 and 2050. The reason for this trend is an inversion of the power flow direction. In 2020 main direction of the power flow is from France to Spain as – except of the hours with very high wind infeed in Spain – the cheap nuclear energy in France is exported to Spain with higher average prices. As also the average emissions in France are lower than in Spain the line allows a reduction of CO2-emissions. In the following time steps of the scenario data however, Spain faces a strong increase in renewables sources what leads to decreasing prices and average emissions in Spain. In 2030 and especially in the “high-RES”-scenarios (Renewables, Desertec and Efficiency) the average emissions of France and Spain are almost balanced so that only small changes in CO2-emissions occur. With the further rising share of renewables in Spain the power flow direction turns towards France more and more often and the extension measure lead to increasing CO2-savings.

2020 2030 2040 20500

0.5

1

1.5

2

2.5

year

∆ C

O2

[Mt C

O2/

a]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 64 of 383



Figure 40: Changes in CO2-emissions caused by line “France-Germany”

The interconnection extension “France-Germany” also causes the highest CO2-savings in 2020. The reasons are similar to the previously discussed measure: In 2020 the average emission level is lower in France than in southern Germany where many nuclear power plants where shut down. In the later decades, the increasing share of renewables in Germany – here especially photovoltaic – lowers the gap between the average emissions of the countries so that the CO2-savings decrease. To finalize the analysis of the impact of a grid development measure on the CO2-emissions it can also be mentioned that the CO2-emissions produced by the construction of a line can be totally neglected in comparison with the emissions arising from the power plants and even in comparison with the changes of the generation’s emissions resulting from network extension. Even in only one year of operation the emissions that can be saved due to the network development are much more than the construction-emissions that are about 1000 tCO2 in all three cases. Other direct ecological effects resulting from an overhead line like land use and visibility will not be discussed in the following as the emphasis of these factors do only depend on the public perception. Change of generation costs caused by grid development measures The objective function of the optimal power flow algorithm represents the generation costs including the fuel costs and the CO2-emisson costs. Next to the investment and annual maintenance costs, the economic impact of a single line therefore includes the difference of generation costs, which can be determined by comparative simulations with and without the line. If for example an extension allows to increase the infeed of cheap – in term of variable costs – generation units and to decrease the infeed of expensive sources, the line enables a positive economic impact due to fuel

2020 2030 2040 20500

0.5

1

1.5

2

2.5

year

∆ C

O2

[Mt C

O2/

a]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 65 of 383



cost savings. For a complete economic evaluation however, these cost savings have to be considered in comparison with the investment and maintenance costs. Therefore, the present values of the reductions of generation costs that are considered as economic gain are calculated. In order to provide a coherent view of the economic impact of the different grid development measures, the resulting generation cost changes are presented, followed by a summary of the overall present values.

Figure 41: Changes in costs caused by line “Diele-Niederrhein”

The reduction of generation costs enabled by the line “Diele-Niederrhein” bears a resemblance to the increase of used renewable energy illustrated in Figure 35. The generation costs decrease as the wind power, which is simulated with marginal costs of zero, is enabled to feed into the grind and replace more expensive thermal generation units.

2020 2030 2040 20500

200

400

600

800

1000

1200

1400

year

∆ C

[M€/

a]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 66 of 383



Figure 42: Changes in costs caused by line “Spain-France”

The change of generation costs caused by the interconnector extension between Spain and France is quite constant in 2020 and 2030. The generation costs can be reduced by roughly 200 Million Euro per year in this period. Except for the BAU scenario where a light decrease of the savings occurs, the differences in generation costs increase in the following decades. The increasing share of renewable energy sources in Spain leads to this effect.

Figure 43: Changes in costs caused by line “France-Germany”

2020 2030 2040 20500

100

200

300

400

500

600

700

year

∆ C

[M€/

a]

BAUCCSDESEFFRES

2020 2030 2040 20500

50

100

150

200

250

300

year

∆ C

[M€/

a]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 67 of 383



The costs differences for the extension France-Germany decrease after 2020. It is not before 2050 that the values increase significantly again. In 2050, the Efficiency scenario leads to values twice as high as all other scenarios. The reason for this is that in all scenarios – except for BAU – the installed capacity of nuclear power in France is equal, although the load and the installed power of renewable energy sources changes. For the Efficiency scenario it implies that a high amount of cheap energy is available in France leading to a good utilisation of the considered line.

Figure 44: Present value

The line “Diele-Niederrhein” result in very high present values in the scenarios with a high share of renewables as expected. The high importance of this line for the integration of the offshore wind power has already been discussed. The interconnector extension between Spain and France shows good results in all scenarios with small deviations. However, it has to be considered that the investment cost for this extension should be higher than given by the database as the necessary crossing of the Pyrenees is expensive. Negative net present values are the result of the extension France-Germany. The relative small savings of generation costs lead to the conclusion that this specific extension measure has a negative economic impact. Only in the BAU scenario small revenues are possible. §4.1 sumarizes the overall environmental sustainability conclusions drawn from above §3.1.

DI - NR ES - FR FR - DE-500

0

500

1000

1500

2000

2500

3000

3500

4000

Net

Pre

sent

Val

ue [M

€]

BAUCCSDESEFFRES


EUROPEAN COMMISSION

DG-TREN

Page : 68 of 383



3.2 INFRASTRUCTURE MEASURES FOR INCREASED SECURITY

3.2.1 Introduction

3.2.1.1 Technology Survey

A survey on relevant technologies that facilitate transmission expansion was established [W3ET TN 3004 B], which is used to obtain the here presented results for security. The focus is primarily on the effect of each of these technologies on power system security. This, to a wide extent, addresses both security of supply as well as infrastructure security. Both ac and dc systems - including power-electronics-based solutions – are reviewed. Construction of ac lines is probably one of the first options considered in transmission expansion planning. AC transmission relies on a mature and reliable technology, widely used, and with many standardised equipment, e.g. protection components. It allows simple connectivity between the nodes, and easy injection/withdrawal of power along an ac transmission path. The construction of additional ac links has a significant positive effect on most aspects of security, as it assists in congestion relief, alleviates oscillatory instability, and at the same time, it enhances both transient and voltage stability. AC transmission is probably the option usually preferred for shorter transmission lengths, while, along with compensation (or FACTS devices), it provides an attractive solution for distances up to 300-400 km. An increased number of new projects involves higher voltage levels for transmission, e.g. above 700 kV, especially in systems that are now developing their interconnections, such as India. Overhead lines demonstrate a better performance than cables or Gas-Insulated Lines, in many aspects. Both cables and GIL have a limited transmission length, due to the need of reactive compensation. Cables also do not allow transmission capacities as high as overhead lines, while widespread construction of GIL still lacks maturity (as a state-of–the-art technology) and acceptance at some TSOs (countries) due to the use of CFK-gas SF6 as insulating medium. Nevertheless, for short transmission lengths cables and GIL (e.g. up to 30-40 km for cables and 60-80 km for GIL) have location specific advantages (e.g.less visual impact). In conjunction with the lack of public acceptance for new overhead lines, which is often a major problem, especially in Europe, cables or GIL might be preferred in some occasions. Line reinforcements, such as replacing the conductors of existing lines with high temperature low sag (HTLS) conductors, dynamic line rating (DLR), or upgrading to higher voltage levels, have also been considered. From the three options, voltage upgrading is probably the most effective with


EUROPEAN COMMISSION

DG-TREN

Page : 69 of 383



respect to power system security. DLR and HTLS help relieve congestions, by adding additional capacity to the network, but in terms of dynamic security they probably have an insignificant, or even negative, effect. In comparison with AC systems, HVDC technology still is the only technology that allows submarine cable connections longer than 100 km, while it proves a more economical solution for distances longer than about 600 km. With its active power flow control capability, HVDC can prevent overloadings on parallel AC lines and can assist towards the damping of inter-area oscillations. There are two main types of HVDC technology: the Line Commutated Converter (LCC) and the Voltage Source Converter (VSC). LCC-HVDC is a more mature technology, with several applications around the world for over 50 years. The currently largest project is in China, having an installed power of 6,400 MW, and a distance of 2,000 km. In order for LCC-HVDC to operate, it needs a strong AC network as well as extensive filters at the AC converter stations side (i.e. reactive power) equipment. With the more recent introduction of high power rating IGBTs, the alternative VSC-HVDC was developed which exhibits several advantages over the conventional LCC-HVDC, as it can operate in weaker networks, hardly needs compensation equipment or harmonic filters, has black-start capabilities, and allows independent control of the active and reactive power. The latter allows the VSC-HVDC to assist in voltage stability problems. Nevertheless, the LCC technology exhibits less losses (0.7% for LCC vs. ~1% for VSC expected till 2015-2020 [DEN10]), and it allows for larger transmission capacity; VSC is currently limited to 1,000 MW and is projected to reach 2,400 MW by 2020. With a view in the future though, the most important advantage of the VSC technology is that it allows the operation of DC multi-terminal systems, and, thus, it is able to form HVDC grids. Nevertheless, short-circuit protection, i.e. the unavailability of a dc circuit breaker at the moment, poses a barrier in the practical realisation of dc grids. Flexible AC Transmission Systems (FACTS) are designed in order to provide compensation, injecting or withdrawing reactive power at weak network points in the AC grid. Within the survey, standard compensation equipment (i.e. a mechanical switched reactor or capacitor connected in series or in parallel), Phase-Shifting Transformers, Synchronous Condensers, as well as power-electronics-based equipment are incorporated in the extended FACTS family. FACTS equipped with power electronics can actively control voltage and active power flow. FACTS can be divided in two categories: shunt compensation and series compensation. FACTS connected in parallel are fixed or switchable shunt reactors/capacitors, Static Var Compensators (SVCs) and STATCOMs. FACTS connected in series consist mainly of Fixed Series Compensation


EUROPEAN COMMISSION

DG-TREN

Page : 70 of 383



(FSC – in effect, a series connected capacitor), Thyristor Controlled Series Compensation (TCSC) and Phase-Shifting Transformers (PSTs). FACTS connected in parallel provide voltage support. Especially SVCs and STATCOMs, equipped with power electronics, can actively control the voltage, assisting in voltage stability and, additionally, damp inter-area oscillations. They can further provide help during transient instability in reducing the reactive loading of the generators. STATCOMs, based on the voltage-source technology, demonstrate a better performance than standard (LCC technology based) SVCs, as they are effective also during low network voltage. FACTS connected in series can increase the line flows and assist in voltage and transient stability issues. TCSC and PST can actively control the power flow, preventing the overloading of parallel lines. The TCSC assists also in power oscillation damping, essentially, in a more effective way than SVC or STATCOM. HVDC systems, and specifically VSC-HVDC, combine many of the merits of both parallel and series FACTS devices, while they allow fast controllability. At the same time, they add additional capacity and assist in the increase of N-1 security. Although security problems are usually location-dependent and each demonstrates its own specificities, in general, VSC-HVDC systems can be probably considered as the most effective measure for both steady-state and dynamic security. Table 3.26 summarizes the main findings of the technology survey. A (+) is assigned to a technology if it can contribute positively towards the alleviation of the specific security problem. If a technology is assigned two (++), it means that it is a quite effective means against the respective problem. A zero (0) implies that the technology has neither a positive nor a negative effect, while a minus (-) denots that the use of this technology rather aggravates than alleviates the problem in question. The evaluation of the technologies in Table 3.26 has been carried out with respect to their effect on each security problem purely from a technological point of view, without incorporating any cost considerations in the comparisons.


EUROPEAN COMMISSION

DG-TREN

Page : 71 of 383



Table 3.26 – Technology Comparison for Increased Security

In Table 3.27, the possibilities for combination of the different technologies are presented. In the simulations that we carried out, we focussed on three expansion scenarios: new AC 400 kV overhead lines (“AC OHL, same voltage” in the tables), new AC 750 kV overhead lines (OHL) (“AC OHL, extra high voltage” in the tables) and VSC-HVDC OHL and cables. In both AC scenarios, however, we assumed implicitly reactive compensation mechanisms in the form of Fixed Series Compensation.

EC

Contract n°:

TR

EN

/FP7EN

/218903/˝IRE

NE

-40˝

EU

RO

PEA

N C

OM

MISSIO

N

DG

-TR

EN

Page : 72 of 383

IRE

NE-40 T

ask 3.3 Final Report

W3E

T

TN

3004

11/07/2012

Internal partner reference: Filing N

° D

oc.Type

Order N

° R

ev. N°

Date

_________________________________________________________________________________________________________ A

lstom G

rid – TUD

– ECN

– ETH Zurich – Im

perial – ICC

S-NTU

A – R

WT

H – A

BB

- Siemens

Table 3.27 – Possibilities for combinations of the technology options

EC

Contract n°:

TR

EN

/FP7EN

/218903/˝IRE

NE

-40˝

EU

RO

PEA

N C

OM

MISSIO

N

DG

-TR

EN

Page : 73 of 383

IRE

NE-40 T

ask 3.3 Final Report

W3E

T

TN

3004

11/07/2012

Internal partner reference: Filing N

° D

oc.Type

Order N

° R

ev. N°

Date

_________________________________________________________________________________________________________ A

lstom G

rid – TUD

– ECN

– ETH Zurich – Im

perial – ICC

S-NTU

A – R

WT

H – A

BB

- Siemens

EC Contract n°: TREN/FP7EN/218903/IRENE-40

EUROPEAN COMMISSION DG – TREN

IRENE-40

Page: 74 of 383

IRENE-40 REFERENCE W3IM TN 4005 C 22/03/12

Internal partner reference: Filing n° Doc. type Order n° Rev. n° Date

Alstom – TUD – ECN – ETH Zurich – Imperial College London – ICCS-NTUA – RWTH – ABB – Siemens

3.2.1.2 Network Modelling

A single-node per country model is used for the simulations in Task 3.3. The model represents the power exchanges between the EU-27 states, i.e. including UK, Ireland, as well as Norway, Switzerland, and the Balkan States. The power network inside each country is assumed copperplate. The lines connecting the nodes represent the actual physical interconnecting lines between the countries. For most of the network, the line data are single-line aggregated representations of the UCTE (now ENTSO-E) data. The line capacity of the equivalent lines is set equal to the respective Net Transfer Capacity NTC of 2010, increased by 40%, i.e. 1.4*NTC(2010). This increase is taken into account for three reasons: (a) reactive power flows, (b) ±10% variation of voltage, and (c) line limits before the N-1 criterion is applied. The modeling approach, in a way represents the flow-based allocation mechanism. Flow-based allocation mechanisms are under a test phase in different regions of Europe [KUR10] and are expected to be the major market operation scheme of the European interconnected system in the coming years. In this mechanism, the capacity allocation for the cross-border flows takes place after the computation of the physical flows. For this reason, the TSOs determine the power transfer distribution factors as well as the border capacities of their interconnections [ERL07] and the allocation office carries out a DC-OPF. In our case, the border capacities are the 1.4*NTC(2010) values, while instead of the PTDFs we have the R,X, and B values of the lines. Finally, instead of a DC-OPF we run an AC-OPF and a Security-Constrained OPF.

Modifications due to the Security-Constrained Optimal Power Flow (SC-OPF): The Security-Constrained OPF (SC-OPF) determines the dispatch ensuring that the N-1 security criterion is satisfied. In the case of the aggregated lines – as described above – a single line outage in our model would imply the outage of all the interconnecting lines between two countries. Since most European countries are interconnected by more than one parallel line, we modify our assumption, by representing each interconnection with two identical parallel lines. The impedance of the new lines is calculated accordingly to maintain the equivalent electric characteristics (R and X are doubled, B is reduced by half). Each of the two parallel lines shares half of the initial line capacity.

Special attention is hereby given to: Generator Modelling

3.2.1.3 Expansion Scenarios – Theoretical Approach

Our theoretical analysis for the AC expansion is divided in three steps. First, we assume new parallel AC lines with lengths similar to the existing interconnections. In the second step, we assume longer AC lines, so that weak network points inside each country can be bypassed. In the



IRENE-40

Page: 75 of 383




third step, we examine internal network reinforcements and compare this with the case of long AC lines. Main Message: From an economic point of view, due to the meshed structure of the European power system, an AC overlay network will probably not be the optimal solution. If we go for long AC lines, then we must certainly add series compensation. With fixed series compensation (i.e. a capacitor) the effective length of the line can decrease significantly and the impedance ratios for about 400-500 km long lines will be equivalent to the ones in Section 3.2.6.2 (see also Table 3.34, third column, as well as Figure 3.55 for the line loadings; all the lines in both the Table and the Figure were considered to have a length similar to the interconnection length.). A TCSC (i.e. controllable capacitor) is expected to result in at least similar or probably better results, as long as it can achieve the same degree of compensation. And, based on our simulation results, an HVDC grid will have even better results (in Table 3.28 it is shown that HVDC lines decrease the generation costs 1.6 times more compared to the cost decrease of the "short" AC lines expansion scenario). These conclusions are based on our considerations of a single-node per country European model, which, however, we tried to extend through theoretical calculations in order to account, to a certain extent, also for the internal network of the countries. From a security point of view, it is clear that controllable flows [AC with FACTS and/or HVDC] will usually have a better performance than AC lines, as they offer multiple opportunities, for power flow control (effect also on N-1), voltage stability, and power oscillations damping. Quantitative indices for the power flow control, such as cost of security, will be presented in the simulations’ part. Main Conclusions:

- Adding one additional long HVAC-400 kV/3000 MVA transmission line, we cannot achieve a power transfer higher than 2*NTC in total [for the whole interconnection path], although the capacity limit of 3000 MVA is usually much larger than the NTC value; this serves as an upper bound. Actually, the power transfer is often significantly lower (see also

- -

- Table 3.34 and Figure 3.55). - For UHVAC lines (750 kV AC overlay supergrid) the upper bound lies around 2.6*NTC. - If we also take into account the internal impedances of the existing grid, then this upper

bound decreases. For a 500 km line, it may fall e.g. to 1.6*NTC. Nevertheless, by adding series compensation capacity can increase and reach again the levels of about 2*NTC.

- We compare the case of local network reinforcements vs. long interconnecting AC lines. We show that for building the same kilometers of new lines, and as long as the internal network



IRENE-40

Page: 76 of 383




is more meshed than the interconnections, a long AC line is more preferable than carrying out internal reinforcements.

- Probably controllable flows, in the form of HVDC lines, or AC lines with FACTS, are a more preferable way of expansion (this conclusion is not derived theoretically but is based on our simulation studies, as in the following table):

Table 3.28 – RES 2050: Comparison table of generation costs for different expansion scenarios

Base Case 10 HVAC lines + 9 HVDC

(500 km long)

10 HVAC lines + 9 HVDC

(HVAC equal to interconnection length)

19 VSC-HVDC lines

(500 km length)

106.31 billion € 98.53 billion € 95.29 billion € 88.74 billion € Additional conclusions:

- Upgrading the lines, instead of adding new lines, can lead to better line utilization, but still the addition of a new line allows for more transfer capacity.

- Compensation equipment can improve significantly the performance of long AC lines. Long AC lines are not that suitable in a meshed grid with several interconnections, but could probably be useful in connecting remote generating units (although in such cases HVDC lines could be a more cost-effective option).

Additional Remarks:

- An AC expansion, which assumes that only the congested interconnecting lines should be reinforced, and the full capacity of the new parallel lines can be used, is not valid.

- The AC expansion should also account for reinforcement within a country’s network, or ways to bypass the weak network points.

- When selecting the lines to be reinforced, except for the average loading one should also look at the impedance values.

- Long AC lines: for long AC lines the stability limit is the limiting factor. This is lower than the thermal limit (see following figure). Still, the limits we extract are more constraining than the stability limit.

- Building additional lines increases not only the thermal limit but also the stability limit. In comparison, line reinforcements, e.g. voltage upgrade or HTLS increase only the thermal limit of the line.



IRENE-40

Page: 77 of 383




Figure 3.45 – Maximum active power loading as a function of the line length [EEH11]

Figure 3.46 – Illustrative Comparison of Line Reinforcements against Additional Lines [VEN11]

3.2.1.3.1 Long AC Interconnections Through our studies, we should identify the interconnections for which their transmission capacity should increase. The transmission capacity can increase either by reinforcing the internal network, in order to increase the NTC value, or by adding a longer direct line which will connect two non-

St.Clair Curve 1: until 80 km: thermal limit 2: Between 80 and 320 km: voltage drop 3: above 320 km: stability limit (note: 𝑃𝑆𝐼𝐿 probably lies around 1300 MW for 400 kV double circuit and around 2100 MW for 750 kV)

Line Upgrades Additional Lines



IRENE-40

Page: 78 of 383




congested network points between the two networks, e.g. in the form of a Supergrid. In this way, weak network points can be bypassed. In this section we will study the case of long AC interconnecting lines, while in following sections we address the internal network reinforcements and compare them with the case of long AC lines. Through our analysis, we show that, in any case, a long AC line cannot carry more power than a shorter AC line which is installed only along the interconnection. The results are based on the assumption that the internal network in each country is more meshed than the external network (interconnections). Furthermore, we extract a relationship that calculates the equivalent impedance of the long AC line for our simplified single-node-per-country model. The value of this impedance, and, thus, the maximum power that can flow through the new line, depends on two factors: (a) how much more meshed is the internal network with respect to the external, and (b) how long is the segment of the line inside the country with respect to the external (i.e. the external equals the interconnection length). As an example, we assume that we have a 250 km long AC line, of which 50 km are the interconnection length (external) and 200 km are within the regional networks (internal). We further assume that for each interconnecting line, there exist at least two parallel paths of similar voltage level inside the country (for the European network this would probably be a reasonable assumption; for example, Germany has two interconnecting lines with Czech Republic. But from e.g. Magdeburg there are certainly more than four different transmission paths through which power can be transported to the German borders.). Then the maximum total loading that this interconnection can achieve will decrease from 2*NTC to 1.6*NTC in the best case. Here, series compensation can provide significant help. Assuming 70% compensation on the 250 km line, the effective reactance becomes equivalent to a 70 km line. As a result, if we combine long AC lines with reactive compensation, we can achieve power flows equal to the case of “short” interconnections, i.e. lines which are as long as current interconnections. Concerning the annual generation costs, the results can be considered equal to the third column of Table 3.28. In the next section, we will extract some properties for the “shorter” interconnection lines, having in mind that for our studies they represent long lines with significant compensation equipment installed.



IRENE-40

Page: 79 of 383




3.2.1.3.2 “Short” lines – Reinforcing only the interconnection We find that in case we add one parallel line8

between countries, with length equal to the average existing interconnection length, the maximum power that can flow over the new line is equal to the NTC value of that interconnection. As a result, for the total interconnection capacity, after the reinforcement, it holds:

𝐼𝑡𝑜𝑡 ≤ 2 ∙ 𝑁𝑇𝐶

This upper limit holds true for the vast majority of the interconnections. In fact, the average limit over 18 interconnections, along which we add a new parallel line due to their high loading, is 𝐼𝑡𝑜𝑡 ≈ 1.5 ∙ 𝑁𝑇𝐶. The reason of the significantly low utilization of the new lines lies on the impedance values. The current prefers to flow along the paths with minimum impedance. As a result, interconnections with low impedance values are usually the ones that reach first their capacity limit. A low impedance interconnection implies that there are several 220 kV and 400 kV lines in parallel, which connect the two countries. The addition of a new AC line – which will naturally have higher impedance in comparison to the sum of all existing lines of that interconnection – allows only a fraction of the transmitted power to flow over the new line. We further show that there is a correlation between the impedance values of the existing interconnections and their respective NTC values. The higher the NTC, the lower is the impedance value. This confirms the logical assumption that in higher NTC values, there are usually more parallel lines connecting the two countries. Still, we point out that the low utilization of the new lines should not lead to the installation of lines with lower MVA limits. Lines operating on lower nominal currents and voltages imply also higher impedance values. In such cases, the additional power flow over the new line will be even smaller.

3.2.1.3.3 Internal Network Reinforcements vs. Long AC Interconnections In this section, we compare the bypass of weak network points by long AC interconnections against their elimination by adding shorter parallel lines inside the country, and then reinforcing the interconnection. Again, we take the reasonable assumption for the European system that the networks inside the countries are more meshed than the external network. For proper comparison, we assume no compensation in the long AC line. We show that, below a certain level of internal reinforcements, a long AC interconnection can carry more power. More specifically, we show that if we had a certain number of line kilometers we could build, it would be better to build a long AC interconnection from point A to point C (A and C in different countries), than build a line from

8 With one parallel line we imply a 400 kV 3000 MVA double-circuit line (NB: the only line that can outperform the double-circuit 400 kV is a 750 kV line – analysis on that will follow. All other lines will perform worse).



IRENE-40

Page: 80 of 383




point A to point B inside the same country (point B not an isolated node), and then an interconnection from B to C. As an example, assume that we want to reinforce the interconnection Germany-Switzerland with a 400 kV/3000 MVA transmission line. The length of the interconnection is 20 km. We also assume that the non-congested substation has a distance of 100 km from the interconnection point (e.g. somewhere inside Germany), while the internal network in Germany is more meshed that the cross-border lines (again the assumption that for each interconnecting line there exist at least two parallel paths of similar voltage level inside the country). Then if we go for the option of internal network reinforcements, we should build at least 2 parallel 400 kV lines from that substation to the German border (i.e. 2x100 km+20 km), in order to achieve the same result as a 120 km long AC line.

3.2.1.3.4 Line Upgrades Based on our considerations, upgrading an existing line, e.g. from 220 kV to 400 kV, is less effective than adding a parallel 400 kV line in terms of total power capacity. Assume an interconnection consists of one 220 kV line and one 400 kV line. By upgrading we get two 400 kV lines. By adding a line, we get one 220 kV and two 400 kV lines. Obviously, the second case will allow more current to flow through. The benefit of upgrading is that it allows a better utilization of the lines. Assuming a given amount of power, below the thermal limits in both cases, then adding a new line results in relatively lower line loadings, especially for the 220 kV line. According to draft calculations the fraction of the power transmitted over the 220 kV line will be about 11%, while each of the two 400 kV lines will carry 44.5% of the exchanged power. In the case of upgrading, the two 400 kV lines will each share a loading of 50%. In terms of total power capacity though, adding a new 400 kV line will result in a capacity 1.13 times higher that upgrading. If this extra 13% of capacity is not significant, then a line upgrade would make more sense. Nevertheless, it should be here noted that for security considerations, adding a new line will be always preferred against a line upgrade, as this better assists in the fulfillment of the N-1 criterion. More specifically, assuming that there are no more parallel lines, this N-1 calculation should be based on the worst N-1 case which is the loss of one 400kV line. Thus, in the case of two 400kV lines, the transfer capability of the interconnection will be that of one 400kV (with one line outaged). In the case of two 400kV and one 220kV line the transfer capability will be that of one 220kV and one 400kV line (with one 400kV line outaged).

3.2.1.3.5 750 kV AC Lines If instead of 400 kV Lines we used 750 kV lines, the results will be obviously better, but still the improvement will not be incredibly large. Table 3.29 presents the results from an AC-400 kV expansion, an AC-750 kV expansion and an HVDC expansion. The AC-750 kV expansion results in about 1.3 billion Euros additional generation cost savings per year. The HVDC scenario, on the other hand, achieves a cost decrease about five times as much as the AC-750 kV scenario (6.5 billion Euros/year).



IRENE-40

Page: 81 of 383




Table 3.29 – RES 2050: OPF Results from different expansion scenarios

Base Case 10 EHVAC lines

400 kV 3000 MVA + 9 HVDC

10 UHVAC lines 750 kV 3900 MVA

+ 9 HVDC 19 HVDC lines

106.31 billion € 95.29 billion € 94.01 billion € 88.74 billion € According to draft calculations, based on the impedance values of a double-circuit 400 kV/3000 MVA line, and a single-circuit 750 kV/3900 MVA line, all the limits we have calculated for the additional 400 kV lines can be increased by a factor of 1.61. This means that the average limit of the 18 interconnections will increase from 𝐼𝑡𝑜𝑡 ≈ (1 + 0.5) ∙ 𝑁𝑇𝐶 to 𝐼𝑡𝑜𝑡 ≈ (1 + 0.8) ∙ 𝑁𝑇𝐶 if we add a parallel 765 kV line. Minimum and maximum limits for the 18 interconnections we studied will change from 1.14*NTCÆ1.23*NTC and from 2.16*NTCÆ2.87*NTC. Here, it should be also taken into account that the power which will flow through the additional lines, depends also on the impedance ratio with respect to the parallel paths.

3.2.1.4 Expansion Scenarios – Simulations

3.2.1.4.1 Cost of Security and SC-OPF The goal of the simulations is to identify the appropriate expansion measures in order to address steady-state security. As all expansion scenarios should result in a system which is N-1 secure – at least for the considered critical contingencies – we decided to compare the expansion options with respect to the additional cost they incur to the system in order to satisfy the N-1criterion. Therefore, we introduce the “Cost of Security” index, which is calculated as the difference in the generation dispatch costs between a “standard” OPF and a Security-Constrained OPF (SC-OPF) (see Eq. 3.1).

𝐶𝑜𝑆 = 𝐶𝑆𝐶𝑂𝑃𝐹 − 𝐶𝑂𝑃𝐹 Eq. 3.1

The “Cost of Security” reflects the additional generation costs, which result from the dispatch of more expensive generators, in order to maintain the system N-1 secure. As it becomes evident, except for an AC-OPF we also need an SC-OPF algorithm. In Section 3.2.5, the SC-OPF algorithm we developed is described. It should be noted that the novelty of this algorithm is that it takes into account post-contingency control actions of the VSC-HVDC lines. Because of this additional flexibility, the SC-OPF can achieve a lower cost dispatch, as it does not need to act proactively – by dispatching more expensive generators – in order to reduce line loadings in case a contingency occurs. Instead, it can rely to a certain extent to the fast response of VSC-HVDC lines, so that as soon as a contingency occurs, they will undertake the appropriate corrective actions.



IRENE-40

Page: 82 of 383




3.2.1.4.2 Expansion Scenarios Based on Table 3.26, it seems that the most effective means to increase steady state security is the addition of AC or HVDC lines. We will not study the addition of AC cables or GIL, due to their length limitations. As regards the options of DLR, HTLS and Voltage Upgrade, their effect is not as positive as the addition of a new line (see also Section 3.2.1.3.4 for voltage upgrades). The only FACTS devices that have an effect on congestion relief and steady-state security are the PST and TCSC. Although we did not study explicitly the addition of such devices in the network, we did incorporate fixed compensation mechanisms in our studies. During our simulation studies, we would expect to identify the interconnections, thee capacity of which should increase in order to increase the level of security. The interconnection capacity can increase either by internal network reinforcements, which should have an effect on the NTC values, or by long AC or DC lines which will connect two non-congested network points in neighbouring countries. Based on the conclusions from our theoretical considerations in Section 3.2.1.3, we decided to study long AC lines, since, as we have seen, for the same amount of line-kilometers, i.e. equal to the distance between two non-congested network points in neighbouring countries, a long AC line can carry more power than internal AC network reinforcements. Furthermore, we decided to add compensation to the AC lines, in order to reduce their effective length. As we have seen in Section 3.2.1.3.1, taking into account our modelling approach, the effective length of the additional AC lines is in the same order of magnitude as the existing interconnection lengths (e.g. 70 km). For the sake of simplicity, we assume that every AC line we add in the system has an effective length equal to the respective parallel interconnection. Thus, we achieve two goals. First, the results of the AC expansion scenarios are towards the optimistic side, as the necessary line length in order to bypass congested points will be longer than the interconnection length, which implies a higher impedance. Second, assuming implicitly the compensation mechanisms, gives us the freedom to account for almost any possible AC line length, by just varying the amount of compensation. Last but not least, it should be noted that in our studies we could not account for the effect of additional line-km, when used only for internal line reinforcements. We have shown, indeed, that the necessary line-km will be more than in the case of a long AC line in order to achieve the same result. However, we cannot predict what is going to be the exact effect of e.g double length of line-km for internal reinforcements, compared to long AC lines. It might lead to results similar to the long AC lines, but it could also increase significantly the NTC value, and as a result have a more significant positive contribution. According to the above consideration, we studied three expansion options:

- Addition of AC-400 kV /3000 MVA double-circuit lines with fixed series compensation - Addition of AC-400 kV /3000 MVA double-circuit lines with fixed series compensation - Addition of VSC-HVDC lines (500 kV/3000 MVA, for comparison reasons)



IRENE-40

Page: 83 of 383




3.2.1.4.3 Simulation Results As base case we assumed the existing interconnections in 2010. Running the generation scenarios for 2050, it became evident that in order the load to be served 100% of the time, the capacity of the lines DK-NO and NL-NO should be increased by 1700-4800 MVA, according to the generation scenario. For certain generation scenarios, additional lines should also be reinforced. The system with these additional capacities served as our base case for our calculations. In a first step, we ran an AC-OPF for each generation scenario, in order to identify the lines that were most often highly loaded. We determined that the lines to be reinforced in each scenario should be the lines that are loaded above 99% of their capacity for more than 4000 hours per year. Table 3.30 presents the lines, whose transmission capacity should increase for each generation scenario. In order to increase the transmission capacity we investigated the three different expansion options.

Table 3.30 – Classification and ranking of all the selected Network Reinforcements according to the number of generation scenarios they participate

Expansion on Land Submarine Cables AT-DE All EE-FI All FR-ES All FR-UK All IT-SI All IE-UK All RO-RS All GR-IT All LT-PL All (planned) NL-NO All FR-CH All – except CCS PL-SE All DE-CH All – except CCS NL-UK All (planned) AT-HU BAU+RES DE-NO All (planned) DE-PL DES+RES LT-SE All (planned) GR-MK BAU+EFF DE-SE BAU+EFF PT-ES CCS+DES DK-NO CCS AT-SI DES BE-FR EFF DE-NL RES Table 3.31 presents the estimates for the total amount of series compensation that will be necessary to install for the AC-400 kV and AC-750 kV expansion options.



IRENE-40

Page: 84 of 383




Table 3.31 – Total amount of series compensation that is necessary for each generation scenario

AC-400 kV Total Compensation (Gvar)


BAU 2050 13.9 15.5 CCS 2050 9.4 10.5 DES 2050 16.0 17.8 EFF 2050 14.2 15.8 RES 2050 15.8 17.6 Figure 3.45 presents the percentage reduction in the annual generation costs for each generation scenario and expansion option (results from AC-OPF calculations). As it can be observed, the HVDC expansion option results in the higher cost savings for the European power system for all future generation scenarios. Nevertheless, based on the valuation studies we carried out in Section 3.2.7.5, for all expansion options and generation scenarios, the payback period is between 1-5 years.

0%

5%

10%

15%

20%

25%

BAU 2050 CCS 2050 DES 2050 EFF 2050 RES 2050

Cost Reduction of Expansion Scenarios

AC 400 kV

AC 750 kV

HVDC



IRENE-40

Page: 85 of 383




Figure 3.47 – AC-OPF: Annual Total Generation Costs Reduction (in %) due to the expansion scenarios

Following the AC-OPF calculations, we ran the SC-OPF algorithm for representative samples. In Section 4.2.2, we carry out an analysis which shows that the selected samples can be representative for the whole year, allowing us to gain valuable insights with respect to the performance of each expansion option, but also estimate an order of magnitude for the Cost of Security (CoS). As critical contingencies for each case, we assume the lines that have an annual average loading above 85% during the AC-OPF calculations. The results of our calculations are presented in Figure 3.46.

Figure 3.48 – Cost of Security for the different generation and expansion scenarios. The selected

critical contingencies are the lines which during the OPF calculations had an average annual loading above 85%. Therefore, for each case a different number of contingencies is taken into

account.

As it can be observed from Figure 3.46, in four out of five generation scenarios, the HVDC option achieves lower Cost of Security and, thus, a better performance. The HVDC expansion results in additional CoS savings of about 1to 2 billion Euros in comparison to the AC scenarios. And this happens, despite the fact that all HVDC scenarios consider at least the same, and often a greater, number of critical contingencies than the two AC scenarios. The reason why in the CCS 2050

0,00

1,00

2,00

3,00

4,00

5,00

6,00

7,00

8,00


Billi

ons o

f Eur

os

Cost of Security

AC 400 kV

AC 750 kV

HVDC



IRENE-40

Page: 86 of 383




scenario, the HVDC expansion results in a higher CoS than the AC counterparts is exactly because of this fact. In the AC scenarios the outage of the line PL-SE was not considered as a critical contingency as it had an average annual loading below 85% during the AC-OPF calculations. In Figure 4.8, we present the results of the CCS 2050 scenario, when for all expansion options we have selected exactly the same critical contingencies. From there, it is obvious that the HVDC option has a better performance. Concluding this section, according to our simulation results, it seems that an expansion based on VSC-HVDC lines can have the most positive effect both to generation cost savings but also to the Cost of Security – at least for a set of critical contingencies similar to all expansion options.



IRENE-40

Page: 87 of 383




3.2.2 Comparison of technologies

A survey on relevant technologies for increased security in power systems including real-world examples was established comprising the following:

1. New overhead AC transmission lines 2. AC cables 3. Gas insulated lines (GIL) 4. Dynamic line rating 5. High temperature low sag (HTLS) lines 6. Voltage upgrading 7. Classical high voltage direct current transmission (Classical HVDC) 8. Voltage source converter HVDC transmission technology (VSC-HVDC) 9. Flexible AC transmission systems (FACTS) 10. Phase shifting transformer 11. Synchronous condenser

For each of these the following points where described in full detail: • Main Characteristics • Contribution to Steady State Security • Contribution to Dynamic Security • Costs • Time to deploy • Shortcomings • Projections for Future Development

The above was used to derive the hereafter reported results.

3.2.2.1 Time to Deploy

The time to deploy and implement the various technologies, as with cost, is very difficult to pinpoint in general terms. Each individual project needs to be assessed on a case by case scenario as local geographical, environmental and political conditions play a critical part in determining this parameter of time deployment. However, this does not mean that attempts to quantify this parameter have not been made. Table 3.32 does give a general relative comparison between the primary transmission media and a general indication of deployment time.

Table 3.32 – Construction time for the transmission technologies [BEN01]

Please note that the value referring to HVDC lines corresponds to both overhead lines and cables. Obviously, overhead lines would have a construction time towards the lower limit, while typical



IRENE-40

Page: 88 of 383




values in case of cables would move towards the upper limit. It should further be mentioned, that the minimum time for building AC overhead lines is less than for HVDC, as no AC/DC converters need to be constructed. For the modifications of AC overhead lines, i.e. solutions provided by dynamic line rating, high temperature lines and voltage upgrading, the construction time is very likely to be well under the minimum specified for overhead lines (i.e. 12 days/km). For the dynamic line rating option and dependent on the extent employed, the construction time is not likely to exceed a single week for the whole line concerned. For the high temperature lines option, the conductor replacement on the Belgian example, cited under the real world example section, was approximately six months for the two of 19 km lines – working out to 4.7 days/km [GOF01]. As regards voltage upgrading, again this can vary considerably dependent on particular circumstances and pre-preparation. For the voltage upgrade example, cited under the real world example section, the upgrade was achieved within three months as compared to a speculated minimum two year build – a factor of one eighth and less – implying a 1.5 to 4 days per km construction period.

3.2.2.2 Costs

The cost to deploy and implement the various technologies, as with time, is very difficult to pinpoint in general terms. Each individual project needs to be assessed on a case by case scenario as local geographical, environmental and political conditions play a critical part in determining this parameter of cost. However, this does not mean that attempts to quantify this parameter have not been attempted. Figure 3.47 does give a relative comparison between the primary transmission media and a general indication of costs as it pertains to Germany.



IRENE-40

Page: 89 of 383




Figure 3.49 – Dena Grid Study II Expansion and Annual Costs for different Transmission Options for consideration in Germany [DEN01]

Basic – 380kV overhead line option FLM – Using Dynamic Line Rating option TAL – Modifying existing lines with High temperature conductors Hybrid – 800km 4400MW HVDC from Schleswig-Holstein (north) to Baden-Württemberg (south-west) with additional 3100km overhead lines GIL – Underground Gas Insulated Lines HVDC – note that the values for HVDC corresponds to underground cables, while for AC to overhead lines. Therefore, a direct comparison is not completely possible.

The dena grid study concluded that for short transmission lengths and small capacity (100 km / 1000 MW lines) the 380 kV AC overhead lines provided the best solution. On the other hand, for long transmission lengths and high capacity (400 km / 4000 MW) HVDC lines’ performance proved superior. For transmission lengths and capacities between the two extremes a combination of the two technologies provided the best results. Nevertherless, [DEN01] points out that such sample evaluations are not suitable for generalisation, as each individual project has its own specificities.



IRENE-40

Page: 90 of 383




Further cost comparisons related to the AC line reinforcements, or FACTS devices, are relatively difficult due to the uniqueness of each project. Nevertheless, new AC, GIL or HVDC lines are expected to cost more than any of these reinforcements or elements.

3.2.2.3 Steady-State Security

An IEEE/CIGRE Joint Task Force on Stability Terms and Definitions in 2004 defined the Security of a power system as referring to the ‘degree of risk in its ability to survive imminent disturbances (contingencies) without interruption of customer services’ [IEE01]. Further to this is that it ‘relates to robustness of the system to imminent disturbances and, hence, depends on the system operating condition as well as the contingent probability of disturbances’. Referral was also made to the NERC (North American Electric Reliability Council), which defines: Security – the ability of the power system to withstand sudden disturbances such as electric short circuits or nonanticipated loss of system components. [IEE01] It must be stressed, again, that the following comparisons is just an attempt to highlight the strengths of each technology with respect to different criteria. A generalization is not always possible due to the individual characteristics of each project. Clearly, the most robust system is one which is over-designed (some may say is not ‘cost-effective’) and has many transmission corridors such that the loss of a significant amount of transmission links still enables uninterrupted supply to consumers (customers). So, essentially, the robustness and by implication security, is limited by the investment made in network infrastructure. In the following, the effect of each technology on the steady-state security is examined based on two criteria: the creation of new transmission links and the addition of power transfer capacity.

3.2.2.3.1 Creation of additional transmission links Whatever format the transmission media takes, if there is not enough to deal with credible contingencies, then the network lacks ‘security’. So, irrespective of form, quantity of transmission path is a key component of network security. As a result, the option of building an additional overhead line or cable, either AC, HVDC or GIL is the optimal solution for increased steady-state security, as it increases the security margin with respect to the N-1 criterion. Considering the type of technology that should be used for the additional lines, dc systems may demonstrate superior performance in several indices due to the full controllability they offer. Nevertheless, ac lines present a handful of advantages that need not be neglected. First, ac transmission is a mature technology, widely used, with many standardised equipment developed for



IRENE-40

Page: 91 of 383




it; most importantly, the protection equipment such as circuit breakers. For dc transmission, dc circuit breakers are not available at the moment, thus not allowing the construction of a fully dc grid. Projections show, though, that a dc circuit breaker will exist in the near future [HAF11] (probably beyond 2020, i.e. in the 2030-2050 timeframe). Second, it is easy to tap power in an ac grid, allowing for simple connectivity between nodes or the simple creation of a new node for injection/withdrawal of power. For dc connections, this is not so straightforward, and rather costly, as a new multi-terminal converter substation should be built at the point of connection. Third, the inherent characteristics of ac connections and Kirchhoff laws allow for a relatively simpler control of the whole system. For example, when frequency reduces, generator units provide additional power through their stored kinetic energy (inertia) without any control action necessary. Control areas, interconnected through ac lines, can offer, with almost no control involved, instant support to neighbouring areas, keeping the frequency within limits. Such features are also some of the reasons that allowed ac systems to become so large during the past century. For each of these actions, dc systems need to be actively controlled. The advancements in power electronics have made such control actions possible and reliable, resulting in a performance often superior to the ac counterpart. However, complexity increases as well. A failure to one of these elements, and not the line itself, can lead to the outage of the dc link. It should be noted, though, that current HVDC links’ availability due to unscheduled outages amounts to about 99% [ABB05]. An additional consideration relates to the single-phase operation of poly-phase systems. Especially AC, but also DC, systems could offer such capabilities, which can be considered important for the maintenance of the power system security. However, the practice of enabling single pole operation is not a current inherent feature of many network overhead ac lines. The overwhelming vast majority of overhead ac lines are operated with only three pole operation enabled. Many utilities do not need nor desire single pole operation due to mainly the inability of existing network circuit breaker facilities and the desire not to introduce further phase imbalances – no matter how short lived, into their networks9

9 Three pole operation means that the circuit breaker only trips three phase, i.e. even if the fault is only on one phase, say the A phase, then the circuit breaker trips all three phases A, B and C. The clearance of all three phases is performed within one cycle of each other (at current zeros) after a typical (400kV) three cycle opening time. Three phase tripping mechanisms are standard for all voltages up to 132kV at least. Historically, only circuit breakers operating three phase were available up to say the late 1970s, early 1980s. Even today, many utilities do not employ single phase tripping at all, even at 400kV – for example, National Grid in the UK. Single pole operation of circuit breakers at transmission voltages of 400kV and 275kV is not common standard practice on developed networks – primarily because of ‘tradition’ and the lack of need. In developing networks, however, this scenario is different and the pressure is constantly towards the evolution to lower voltage levels of more and more single pole operation facilities being made

. On the other hand, LCC-HVDC bipole systems are designed so that they



IRENE-40

Page: 92 of 383




can operate with only a single-pole, in case the parallel pole fails. Maintaining 50% of the transmission capacity can be considered quite beneficial, given the fact that LCC-HVDC lines are used for bulk power transmission. This is not case for the VSC-HVDC, where in case of a disturbance the whole link must be outaged. Comparing further cables to overhead line transmission, overhead lines demonstrate in many ways the best performance. They have higher capacity and voltage limits, less limitations concerning the transmission lengths, and can be repaired more easily in case of failure. Nevertheless, for short lengths, cables have less losses, while faults on cables tend to be rarer. In this context, physical environment is an important feature which can affect the continued maintenance of network security. To paraphrase ‘not all ac overhead lines are equal’. For example, in areas of high lightning occurrence levels, ac overhead lines present a higher risk to maintaining stability than those with lower or negligible levels of lighting occurrence (Figure 3.48). In such cases, the options of underground cables (either ac or dc) or GIL demonstrate an advantage.

Figure 3.50 – Isokeraunic Levels in Continental France (Annual mean number of Stormy days) Original Source: Métérologie Nationale [SCH01]

Considering cable transmission specifically, GIL overcome many of the limitations that AC cables pose with respect to transmission length and capacity. Except for the increased costs, however, there are concerns about the environmental footprint of GIL due to their use of SF6. Research is ongoing on how SF6 can be eliminated from the GIL gas mixture. On the other hand, HVDC cables are not limited by either length or environmental concerns. They do, however, have a lower

available – mainly for maintaining stability. A circuit breaker with single pole operation capability is more complex and more costly than one with only a three pole operation capability.



IRENE-40

Page: 93 of 383




capacity limit than GIL, and share the same limitations, as described above, about multi-terminal substations and control issues. It should be noted here that regarding submarine interconnections, for lengths exceeding 60 km, HVDC cables is the only technology that has been used up to now.

3.2.2.3.2 Additional Power Transfer Capacity Except for the quantity of transmission paths, a further comparison can be made with reference to the sustained loading of each solution. Again, the addition of a new line in such a case will have a large effect on the total power that can flow along a corridor. And more specifically, the higher the voltage level, the higher the added capacity, e.g. an additional 750 kV line will show a better performance than an additional 400 kV line. Nevertheless, when it comes to loadability issues, line reinforcements can often compete equally well as an alternative option. Among line reinforcements, voltage upgrading seems to be one of the most effective solutions here, especially if the voltage level difference between the old and the new line is relatively large. As seen in Section Error! Reference source not found., an upgrade from 66 kV to 220 kV increased the capacity limit by 233%. Depending on the project, an upgrade to High Temperature Low Sag (HTLS) lines can be equally effective. For example, in Section Error! Reference source not found., a high-temperature low sag line in 150 kV has increased the thermal capacity by as much as 120%. For higher voltages, e.g. 400 kV, a line upgrade to HTLS may not result in such impressive performance improvements, but, still, it can increase the ampacity of the line somewhere between 50% and 100%. Similar is the case when considering voltage upgrade, if the voltage difference is not that large. Dynamic line rating (DLR) is probably the least cost and fastest to deploy solution, from the three possible AC upgrades. However, it also has a lesser effect on the capacity limit increase of the line. Nevertheless, if the need for capacity increase is not too big, or in combination with wind power plants, DLR can be quite effective, as increased wind results to increased power flow and, at the same time, increased line capacity limits. Considering the limitations of the AC upgrade options, and referring first to the time duration of the increased capacity limit, dynamic line rating seems to have an inferior performance, for two reasons. First, the capacity limit varies with temperature, while for the other options the capacity increase is constant. Second, dynamic line rating seems to be significantly effective during summer, but during winter the bottleneck is usually on the transformer rating, where DLR has no effect. An additional limitation which both HTLS upgrade and DLR options share is their resulting in increased thermal losses in the power system. On the other hand, voltages upgrade effects in the opposite direction, contributing to the reduction of thermal losses in the system. As already mentioned, comparing AC upgrades with the construction of additional lines and considering the capacity added, a new line will usually be more effective than an upgrade.



IRENE-40

Page: 94 of 383




However, when it comes to a corridor where parallel AC lines of different voltage levels and thermal capacities exist, the utilization of the parallel AC lines (both old and new) will be constrained from the capacity limit of the “weakest” link. If the utilization factor is low, and the effective capacity increase10

is not significantly more, then the AC upgrade of the “weakest” link may be preferred over the building of an AC line; and this, irrespective of the additional permission procedures and time that a new line build might need.

A limitation that AC lines demonstrate is related to their length. As shown in Figure 3.49, for line lengths above 80 km, limiting factors such as voltage drop along the line and stability limits play a more important role than the thermal line capacity. As a result, for longer distances, HVDC lines, having the similar or even smaller nominal capacity than AC lines, demonstrate a better performance regarding the effective current carrying capacity. The limitations of the long AC lines can be overcome to a certain extent either with fixed compensation devices or with FACTS. For example, a series compensated 300 km line may be able to allow currents up to its thermal limit (assumed: 70% compensation). Except for static compensation, FACTS devices can also contribute in increasing the loading of long AC lines. Although shunt devices such as SVCs and STATCOMs may assist in voltage drop, the most important positive effect for increased power transfer is provided by a TCSC (thyristor controller series compensation).

Figure 3.51 – St.Clair Curve illustrating the maximum active power loading as a function of the line

length [EEH11]. 𝑃𝑆𝐼𝐿 refers to the surge impedance loading of the line.

1. until 80 km: thermal limit; 2. Between 80 and 320 km: voltage drop;

10 The effective capacity increase is dependent on the impedances of the parallel lines forming the transmission path.



IRENE-40

Page: 95 of 383




3. above 320 km: stability limit Note: 𝑃𝑆𝐼𝐿 should lie around 1300 MW for 400 kV double circuit and around 2100 MW for 750 kV [OSW05]

Except for compensation purposes, controllable devices, such as TCSCs and PSTs (phase-shifting transformers) can actively control the power flow. As a result, they assist in congestion management by routing the power through less loaded lines, and, thus, contribute to the increase of the security margin with respect to the N-1 criterion. If we neglect the effect of line lengths on AC lines, TCSCs (and PSTs) do not actually add additional capacity to the network. But they help redistribute the power flows over the existing lines, allowing for a better utilization of the whole system. Here, it should be also noted, that comparing TCSCs with PSTs for steady-state security, PSTs will probably lead to an increase of the system losses, while TCSCs will effect towards the opposite direction. Going a step further, one could say that for a high amount of congested lines, or when limited transmission paths exist, a new line or an AC upgrade can prove more effective with respect to congestion relief and steady-state security. However, when already several transmission paths exist, and there is a limited amount of weak transmission links, a controllable device can be equally effective in congestion relief, while it can further enhance steady-state security by routing the power through other corridors, when an outage occurs.

3.2.2.3.3 Conclusions on Steady-State Security Concluding this section, all technologies considered in this survey can contribute to the steady-state security. The addition of new lines, either ac or dc, contributes the most to the enhancement of steady-state security, as it creates a new transmission link and, at the same time, increases the power transfer capacity of the system. Both of these factors contribute significantly to the N-1 security. For long transmission, the HVDC option is probably the most appropriate. For shorter transmission lengths (e.g. less than 300-400 km), AC lines, with series compensation if necessary, can perform equally well. In case of cables, HVDC is probably the best solution above 60 km. If a new line cannot be built, an upgrade should be considered. Most effective upgrade would be the voltage upgrade, especially if the line will be upgraded from a low to a significantly higher voltage rating. If the voltage level should be kept the same, an upgrade to HTLS can be considered. If a small, or temporary, capacity increase is necessary DLR might be applicable. TCSCs and PSTs add controllability to the power flow. Optimally, such devices should be installed on under-utilized transmission lines – when the objective is to increase the power transfer capacity. Controllable devices can overcome the constraints in power flow imposed by weak parallel links, while they can actively contribute towards the maintenance of the N-1 security, even after an outage. Therefore, their performance is not easy to be compared with the other “passive” AC options. Attempting a comparison, although this may be true only in limited cases, adding a FACTS device at a high-voltage underutilized line will probably have an effect somewhere between voltage upgrading and



IRENE-40

Page: 96 of 383




HTLS upgrading; but instead of gaining additional transfer capacity, performance increases by better utilizing the existing line, as well as in terms of controllability.

3.2.2.4 Small-Signal Stability

The IEEE/CIGRE Joint Task Force on Stability Terms and Definitions in 2004 proposed the following definition of Power system stability: ‘Power system stability is the ability of an electric power system, for a given initial operating condition, to regain a state of operating equilibrium after being subjected to a physical disturbance, with most variables bounded so that practically the entire system remains intact’ [IEE01].

From this over-riding definition, the task force introduced subcategories in which one is the ‘Small-disturbance (or small-signal) rotor angle stability’ which is ‘concerned with the ability of the power system to maintain synchronism under small disturbances. The disturbances are considered to be sufficiently small that linearization of system equations is permissible for purposes of analysis’ [IEE01].

With respect to small-signal stability there are different types of oscillation modes existing in a typical power system [KUN94]: ▪ Local modes or machine-system modes ▪ Inter-area modes ▪ Control modes ▪ Torsional modes In large power systems, the main problems are inter-area oscillations between groups of generators at a very low frequency. In essence, ‘passive’ transmission media, such as AC lines and AC upgrades, play virtually no ‘dynamic’ influence during the time frame of a small disturbance. Nevertheless, the addition of a new AC line (either overhead line, cable or GIL) reduces the effective reactance of the whole transmission path and, as a result, decreases the electrical distance. The option of voltage upgrading has also a similar effect. A shorter electrical distance means a stronger electrical connection between the two nodes, which has a positive effect towards the elimination of interarea oscillations, which occur between the two areas connected through this transmission path. The option of HTLS upgrading or DLR can have a negative effect on small-signal stability as it allows a higher line loading.

On the other hand, HVDC transmission systems embedded in or interconnecting AC networks are used, already today, for adding positive damping to possibly existing oscillations modes. Therefore,



IRENE-40

Page: 97 of 383




the active power transmitted is modulated in order to mitigate the oscillations. The controller is usually called Power Oscillation Damping Control (POD). For classical HVDC this POD control is already applied in few projects [TER10]. With this technology only the active power can be used for damping control. Nevertheless, the positive impact on oscillations is proven. The VSC-HVDC on the other hand can control active and reactive power independently from each other. Thus, both controls can be used on order to increase power oscillation damping. By modulating the reactive power output, the voltage at the HVDC or a nearby bus bar can be controlled. Thus, the VSC-HVDC can perform similar power oscillation damping (POD) tasks as STATCOMs and SVCs, which are applied since several years [LGS09]. The TCSC can – similar to the HVDC systems – control the active power flow across a line or a corridor. Thus, an interarea oscillation damping control can be added to work against oscillation in the power flow and thus increase the damping of critical or selected modes. STATCOMs and SVCs change the voltage at a nearby bus bar in the transmission system and influence the active power flow across the network in this way.

3.2.2.4.1 Conclusions on Small-Signal Stability The most positive effect towards the increase of small-signal stability is provided by the VSC-HVDC lines, due to the increased controllability they offer, i.e. independent control of active and reactive power. The LCC-HVDC technology can also perform successfully power oscillation damping tasks, followed by the TCSCs, which probably demonstrate a performance almost similar to the HVDC-LCC technology. SVCs and STATCOMs can contribute to damping of interarea modes, but, probably, at a lesser extent than the rest of the “active” control options. As interarea oscillations mainly occur due to weak electrical connections, the construction of a parallel AC link (either OHL, cable or GIL) – or voltage upgrading – can contribute to the elimination of such modes. Nevertheless, the power electronic based options (i.e. HVDC lines and FACTS) are able to target specific frequencies of oscillations and damp them effectively. Such possibilities cannot be provided by the “passive” AC elements.

3.2.2.5 Transient Stability

The IEEE/CIGRE Joint Task Force on Stability Terms and Definitions in 2004 discussed transient stability as being ‘concerned with the ability of the power system to maintain synchronism when subjected to a severe disturbance, such as a short circuit on a transmission line’ [IEE01]. The time frame of interest is indicated as ‘usually 3 to 5 seconds following the disturbance’ and ‘it may extend to 10-20 seconds for very large systems with dominant inter-area swings’.



IRENE-40

Page: 98 of 383




Transient stability (or first swing stability) is mainly affected by the strength of the network at the generator bus bars and operating point of the synchronous generators. Automatic Voltage Regulators (AVRs), which are installed on the generators, have only very little effect, during the transient phenomena. Concerning the network, strong electrical connections between the synchronous generators help significantly towards the maintenance of synchronism. Additional AC transmission links, or voltage upgrading, help towards this direction. It should be also noted that such “passive” solutions, on the one hand, enable faults to ‘travel’ but, on the other hand, do enable further “active” devices, primarily generators, to help deal with the incident. The larger and more interconnected the network, the more likely it is to deal with a severe incident. AC connections, however, come with the inherent feature of spreading the risk. “Active” solutions such as HVDC lines, on the other hand, keep the different areas isolated, thus not allowing the disturbance to spread. Classical HVDC as well as VSC-HVDC systems do not contribute to the fault level, so both have only very limited impact on transient stability. Although HVDC systems have a fast controllability of active power flows, their effect on transient stability is negligible. Nevertheless, as VSC-HVSC systems can provide reactive power and voltage control to the network, the loading of surrounding synchronous generators might be reduced. This change in the operating point of the generators is improving the stability margin. Also FACTS devices can influence transient stability margins by changing the steady-state operating point and by a reduction of loading of synchronous generator.

3.2.2.5.1 Conclusions on Transient Stability Transient stability is mainly dependent on the strength of the network and the operating point of the (synchronous) generators. In order to keep the system secure during transient phenomena, control actions on the generator level are usually taken. Strong electrical connections between the synchronous generators, either by adding ac transmission links or by voltage upgrading, may possibly have the largest effect towards the increase of transient stability margin, when considering network solutions. HVDC systems and FACTS devices, in comparison with the construction of new ac links, have probably a lesser contribution to transient stability; they can act either by moving the pre-fault operating level to a more “secure” one, or through a reduction of the synchronous generator loading by supplying the grid with reactive power.

3.2.2.6 Voltage Stability

The IEEE/CIGRE Joint Task Force on Stability Terms and Definitions in 2004 refers voltage stability as ‘the ability of a power system to maintain steady voltages at all buses in the system after



IRENE-40

Page: 99 of 383




being subjected to a disturbance from a given initial operating condition’ [IEE01]. Any sudden significant loss of generation needs to be matched with load shedding in order to maintain overall voltage stability. If not, transformer tap changers act to boost local voltage levels causing greater voltage instability. The system wide maintenance of voltage stability often entails the loss of pre-determined sacrificial consumer loads. Similarly, any significant load loss needs quick generator response curtailing power output.

For ‘passive’ transmission media, their inherent overloading capability does lead to greater voltage drops than that for ‘active’ media such as HVDC lines, exacerbating any ‘voltage collapse’ scenario. The greater the level of overload capability, the greater the voltage drop, and in essence the more severe the voltage collapse. Under such circumstances, the overloading capability in terms of maintaining voltage stability is disadvantageous, even though it is advantageous for the topics of steady-state security, small-signal stability and transient stability discussed previously. Clearly, being able to obtain greater power flow, for instance with dynamic line rating, does not prove advantageous for voltage stability. On the other hand, the addition of new parallel transmission links, or voltage upgrading, reduces the effective impedance of the transmission corridor, and, as a result, contributes to the decrease of the voltage drop and the increase of voltage stability. As far as the HVDC technologies are concerned, there is a major difference between them, with respect to voltage stability. The classical HVDC consumes up to 60% of the rated active power in reactive power under full load conditions, which have to be compensated. For compensation equipment usually the harmonic filters and additional capacitor banks are used. During voltage reduction, the effect of the compensation is reduced as well, leading to reduced voltages near the HVDC terminal. Also switching actions of filters and capacitors can lead to considerable voltage variations, if the network is weak. The VSC-HVDC is able to supply reactive power to the network up to its MVA rating. Additionally, voltage control at the connection point is possible. Voltage stability near the HVDC terminals is improved considerably. The effect is similar to the operation of STATCOM devices. SVC and STATCOM both can contribute reactive power to the network and can improve steady-state voltage stability considerably. During transients and voltage dips, the capability of the SVC is limited and reactive power output is low. The SVC can be assumed to be a constant impedance, when turned on completely. The reactive power capability of the STATCOM is independent from the voltage in the grid. The STATCOM is only limited in the maximum current output. Thus, voltage stability can be improved considerably in the area surrounding the compensation equipment.



IRENE-40

Page: 100 of 383




The TCSC can improve voltage stability considerably as it reduces the electrical length of an AC transmission line and thus the voltage drop along the line. When increasing the power transfer along a line with series compensation, the voltage stability margins are increased compared to the scenario without TCSC. On the other hand, a PST consumes reactive power which is generally not good when voltages are dropping.

3.2.2.6.1 Conclusions on voltage stability The option of STATCOMs, or VSC-HVDC lines, contributes the most towards the maintenance of voltage stability in the network. SVCs, or even LCC-HVDC, can also assist significantly in maintaining voltage stability under normal operation. However, during extreme conditions, e.g. during a voltage collapse, the contribution of SVCs and LCC-HVDC is limited. Reducing the effective impedance of a transmission corridor has also a significant effect on voltage stability. This can be achieved either by series compensation equipment, TCSCs included, or by the construction of additional ac lines, or through voltage upgrading. The options of HTLS, DLR, and PSTs probably aggravate voltage stability problems, rather than alleviate them.

3.2.3 Network modelling

For the steady-state simulations we use a “single-node per country” model. The model represents the EU-27 Region, with the exception of Malta and Cyprus, which are islands, as well as Luxembourg which is incorporated in Germany. The model represents also the Balkan States, Norway and Switzerland. The modeling approach, in a way represents the flow-based allocation mechanism. Flow-based allocation mechanisms are under a test phase in different regions of Europe [KUR10] and are expected to be the major market operation scheme of the European interconnected system in the coming years. In this mechanism, the capacity allocation for the cross-border flows takes place after the computation of the physical flows. For this reason, the TSOs determine the power transfer distribution factors as well as the border capacities of their interconnections [ERL07] and the allocation office carries out a DC-OPF. In our case, the border capacities are the 1.4*NTC(2010) values, while instead of the PTDFs we have the R,X, and B values of the lines. Finally, instead of a DC-OPF we run an AC-OPF and a Security-Constrained OPF.

3.2.4 Generator modelling

3.2.5 Description of the AC-OPF and the security-constrained OPF algorithms

3.2.5.1 Cost of Security Approach

In order to ensure security, power system operation should always fulfill the N-1 criterion. According to ENTSO-E definition of N-1 criterion [7], a loss of an element within the TSO’s responsibility area must not endanger the security of interconnected operation and “must not lead to



IRENE-40

Page: 101 of 383




the triggering of an uncontrollable cascading outage propagating across the borders or having an impact outside the borders [of the TSO responsibility area]”. Most power systems in practice, not only in ENTSO-E, are bound to operate in an N-1 secure state. Common formulations of OPF-problems usually do not consider the N-1 security criterion. An optimization problem which explicitly takes into account outage events is often termed Security-Constrained Optimal Power Flow (or SC-OPF). The objective is to find a least-cost generation dispatch such that an outage of an arbitrary line or generator will not lead to overloads at any point in the system. It is evident that the dispatch costs determined from SC-OPF calculations will be at least equal or higher than the costs determined through a “standard” OPF. The difference in the generation dispatch costs between a “standard” OPF and an SC-OPF is what we term the ‘Cost of Security’ (see Eq. 3.2). Essentially, the ‘Cost of Security’ reflects the additional costs that are incurred to the system, so that N-1 security is ensured [6].

𝐶𝑜𝑆 = 𝐶𝑆𝐶𝑂𝑃𝐹 − 𝐶𝑂𝑃𝐹 Eq. 3.2 The ‘Cost of Security’ (CoS) provides a quantitative index when comparing the effects of different network reinforcements on the system with respect to power system security. The solution that results in the least ‘cost of security’ guarantees, on the one hand, the secure operation of the power system, and, at the same time, it has the maximum positive effect from a societal viewpoint, as it incurs the least additional costs.

3.2.5.2 AC Optimal Power Flow including VSC-HVDC Lines

The formulation of the optimization problem is divided in two parts. In this section we describe the standard formulation of an AC Optimal Power Flow (OPF) and the necessary extensions in order to include the HVDC line constraints in the algorithm. In the next section we discuss how system security is incorporated in the optimal power flow resulting in a Security-Constrained OPF.

3.2.5.2.1 Standard AC Optimal Power Flow (AC-OPF) The objective of the standard AC Optimal Power Flow (AC-OPF) is usually to minimize total generation costs (see Eq. 3.3). The AC-OPF is implemented as follows:

𝑚𝑖𝑛 � 𝐶𝑗𝑁𝑔𝑒𝑛

𝑗=1�𝑃𝐺𝑗� Eq. 3.3

subject to: 𝑓(𝜃,𝑉,𝑃,𝑄) = 0, Eq. 3.4

𝑃𝑚𝑖𝑛,𝑖 ≤ 𝑃𝑔𝑒𝑛,𝑖 ≤ 𝑃𝑚𝑎𝑥,𝑖, Eq. 3.5 �𝐼𝑖𝑗(𝜃,𝑉)� ≤ 𝐼𝑖𝑗,𝑚𝑎𝑥, Eq. 3.6 𝑄𝑚𝑖𝑛,𝑖 ≤ 𝑄𝑔𝑒𝑛,𝑖 ≤ 𝑄𝑚𝑎𝑥,𝑖, Eq. 3.7 �𝐼𝑗𝑖(𝜃,𝑉)� ≤ 𝐼𝑗𝑖,𝑚𝑎𝑥, Eq. 3.8

𝑉𝑚𝑖𝑛 ≤ 𝑉𝑏𝑢𝑠,𝑖 ≤ 𝑉𝑚𝑎𝑥, Eq. 3.9 𝜃𝑟𝑒𝑓 = 0. Eq. 3.10

Eq. 3.4 represents the power flow equations as described in [8] and [9]. The remaining constraints refer to the active and reactive power limits of the generators (Eq. 3.5 and Eq. 3.7), the voltage



IRENE-40

Page: 102 of 383




limits of the nodes (Eq. 3.9) and the line thermal limits, expressed as constraints on the current (Eq. 3.6 and Eq. 3.8). Eq. 3.10 is added, defining the slack bus, where the phase angle is set to zero.

3.2.5.2.2 VSC-HVDC Modelling and Additional Constraints A VSC-HVDC system consists of two voltage-source converters, connected either Back-to-Back (BtB) or through a common DC-link. The modelling of the VSC-HVDC line for the OPF calculations, as shown in Figure 3.50, is based on [3]. The two HVDC converters are represented by ac voltage sources, which are connected to the ac side through the coupling transformer’s impedance 𝑍𝐶 . The dc-side is described by the relationships Eq. 3.11-Eq. 3.17. 𝑀𝐶𝑖 stands for the PWM amplitude modulation index, 𝑉𝐷𝐶𝑖 is the average dc capacitor voltage at the dc side, and 𝑅𝐷𝐶 the DC cable resistance. From Figure 3.50 it is evident that for every new HVDC line we incorporate in our system, we need to create two new virtual ac nodes, where we connect the equivalent ac voltage sources of the HVDC-link. The HVDC-Voltage Source converters allow for two independent power control loops: the active power and the reactive power control loop [3]. In the active power control loop, the converter in the one side is set to control the power flow, while the converter at the other side controls the dc voltage. In the reactive power control loop, the two converters can act independently, controlling either the voltage at the node, or the reactive power. Combining the two control loops, each converter is able to operate in two possible modes, either in the PV or in the PQ control mode. In the PV control mode, the converter controls the active power and the voltage, while in the PQ mode it controls the active and the reactive power. In both cases, if one converter controls the active power, the other will control the dc voltage.

Figure 3.52: Equivalent circuit of the VSC-HVDC line for OPF calculations

𝑅𝑒�𝑉𝐶𝑘𝐼𝑘∗ + 𝑉𝐶𝑚𝐼𝑚∗ � + 𝑃𝐷𝐶𝑙𝑜𝑠𝑠= 0

Eq. 3.11 𝑉𝐶𝑖 = 𝑀𝐶𝑖

2√2𝑉𝐷𝐶𝑖 , 𝑀𝐶𝑖 ∈ [0,1], 𝑖= 𝑘,𝑚

Eq. 3.12

𝑃𝐷𝐶𝑙𝑜𝑠𝑠 =𝑃𝐶𝑗2 𝑅𝐷𝐶�𝑉𝐷𝐶𝑗�

2 Eq. 3.13 �𝑉𝐷𝐶 = 𝑉𝐷𝐶𝑗 − 𝑉𝐷𝐶𝑖 +

𝑃𝐶𝑗𝑅𝐷𝐶𝑉𝐷𝐶𝑗

= 0

𝑖, 𝑗 = {𝑘,𝑚} , 𝑖 ≠ 𝑗 Eq. 3.14

𝑉𝐶𝑖,𝑚𝑖𝑛 ≤ 𝑉𝐶𝑖 ≤ 𝑉𝐶𝑖,𝑚𝑎𝑥, Eq. 3.15 𝑉𝐷𝐶𝑖,𝑚𝑖𝑛 ≤ 𝑉𝐷𝐶𝑖 ≤ 𝑉𝐷𝐶𝑖,𝑚𝑎𝑥, Eq. 3.16

𝑀𝐶𝑖,𝑚𝑖𝑛 ≤ 𝑀𝐶𝑖 ≤ 𝑀𝐶𝑖,𝑚𝑎𝑥, Eq. 3.17



IRENE-40

Page: 103 of 383




3.2.5.3 Security-Constrained Optimal Power Flow

We will focus mainly on the line outages and their effects on the line flows (including the HVDC links). Incorporating the N-1 security criterion means that additional constraints should be introduced, calculating the line loadings when an unplanned outage of a single transmission line occurs. The objective is to find a least cost generation dispatch such that an outage of an arbitrary line will not lead to overloads at any point in the system. For the implementation of an SC-OPF different approaches exist. Usually, a standard OPF is solved, followed by a contingency analysis which determines the critical overloads. Then these constraints are added in the initial OPF problem and the OPF is solved again [10]. Another possibility is to rely on linear approximations, as in [10] and [6]. An approach followed in [11] and [9], introduces a second set of power flow equations and constraints, in order to incorporate “critical” conditions associated with the maximum loading margin. In this way, the N-1 criterion is being considered by taking into account the overloading, which occurs during the most severe line outage. In this algorithm, we use the current injection method, which allows us to compute the line flows after an outage without having to solve a full power flow. The method was first derived in [12] and [13], and enhanced and applied for FACTS devices in [14]. Here, a further enhancement to the method is proposed in order to incorporate the HVDC corrective control actions after a contingency. In Section 3.2.5.3.1 we describe the basic current injection method, and in Sections 3.2.5.3.2 and 3.2.5.3.3 the enhancement for the HVDC lines, and how this is combined with the method for increased accuracy proposed in [14].

3.2.5.3.1 Current Injection Method The current injection method is able to determine accurately the line currents after an outage through the solution of a linear system of equations, thus avoiding the need of a full power flow calculation. As a result, constraints for the line currents after an outage can be easily incorporated in the optimization problem. Figure 3.51 illustrates the concept of the current injection method. The left part shows the system state before the outage, while in the middle the situation after the outage is represented. As shown in the right part of Figure 3.51, the outage of the line (𝑙 − 𝑝) can be represented by a pair of injection currents 𝐼𝑆𝑙 , 𝐼𝑆𝑝 at the respective nodes, compensating for the current flowing over the line. Essentially, instead of changing the network topology, injection currents eliminate the line flow, thus simulating the line outage.



IRENE-40

Page: 104 of 383




Figure 3.53: Illustration of the Current-Injection Method Following the notation in [14], in the non-faulted situation, the relationship between bus voltages and bus currents is given by Eq. 3.18. 𝑍0 is the bus impedance matrix. When a line outage occurs, bus voltages and bus currents change to 𝑉𝐹 and 𝐼𝐹 respectively, while the new network topology leads to a new bus impedance matrix 𝑍𝐹. Eq. 3.19 holds in this case.

𝑉0 = 𝑍0 ∙ 𝐼0, Eq. 3.18

𝑉𝐹 = 𝑍𝐹 ∙ 𝐼𝐹 Eq. 3.19

Here, we introduce the vector of injection currents 𝐼𝑆. Assuming a non-faulted network topology, as given by 𝑍0, the injection currents should compensate for the flows on the outaged line. The only non-zero elements of the vector 𝐼𝑆 correspond to the nodes {𝑙,𝑝}, where the outage occurred. Hence, Eq. 3.20 is equivalent to Eq. 3.19:

𝑉𝐹 = 𝑍0 ∙ �𝐼𝐹 + 𝐼𝑆�. Eq. 3.20 As a result, the change in the bus voltages after the fault is given by:

𝛥𝑉 = 𝑉0 − 𝑉𝐹 = 𝑍0 ∙ �𝐼0 − 𝐼𝐹��=0

− 𝐼𝑆�. Eq. 3.21

As implied by Eq. 3.21, the basic current injection method assumes that the difference in the bus currents before and after the outage is negligible. Hence, the changes in the line flows can be computed through Eq. 3.22:

𝛥𝐼𝑙𝑖𝑛𝑒 = 𝑌𝐿 ∙ 𝛥𝑉 = −𝑌𝐿 ∙ 𝑍0��=𝐷

∙ 𝐼𝑆 = 𝐷 ∙ 𝐼𝑆. Eq. 3.22

Here, 𝑌𝐿 is the line admittance matrix of the situation without outage. Matrix 𝐷 = −𝑌𝐿 ∙ 𝑍0 is known as the matrix of distribution factors [12], reflecting the influence of the injection currents to the line currents. The injection currents that we calculate have an effect not only on the rest of the line flows but also on the outaged line itself, in a sense of recursion. In other words, any additional current injection at the nodes 𝑙,𝑝 will also change the line flow to 𝐼𝑙𝑝,𝑛𝑒𝑤 = 𝐼𝑙𝑝,0 − 𝛥𝐼𝑙𝑝. As a result, if we wish to eliminate the line flows on the outaged line, the injection currents should be determined from the following equations:

𝐼𝑆𝑙 = 𝐼𝑙𝑝,0 − 𝛥𝐼𝑙𝑝 = 𝐼𝑙𝑝,0 − 𝐷𝑙𝑝,𝑙 ∙ 𝐼𝑆𝑙 − 𝐷𝑙𝑝,𝑝 ∙ 𝐼𝑆𝑝 , Eq. 3.23



IRENE-40

Page: 105 of 383




𝐼𝑆𝑝 = 𝐼𝑝𝑙,0 − 𝛥𝐼𝑝𝑙 = 𝐼𝑝𝑙,0 − 𝐷𝑝𝑙,𝑙 ∙ 𝐼𝑆𝑙 − 𝐷𝑝𝑙,𝑝 ∙ 𝐼𝑆𝑝 , Eq. 3.24 where 𝛥𝐼𝑙𝑝 corresponds to the element associated with the line 𝑙 − 𝑝 in the vector 𝛥𝐼𝑙𝑖𝑛𝑒, while 𝐼𝑙𝑝,0 is the current that flows on the line before the outage. The notation 𝐷𝑙𝑝,𝑙 indicates the element in the row associated with line 𝑙 − 𝑝 and in column 𝑝 of matrix 𝐷. Using the values for 𝐼𝑆𝑙 and 𝐼𝑆𝑝calculated through the above equations, the changes in voltages and line currents after the outage can consequently be calculated with the help of Eq. 3.21 and Eq. 3.22.

3.2.5.3.2 Inclusion of HVDC lines in the Security-Constrained OPF As it is evident from Eq. 3.21, the changes in the bus currents before and after the outage are considered negligible in the basic current injection method. However, as we want to account for the post-contingency control capabilities of the HVDC lines, and based on the HVDC modeling approach in Section 3.2.5.2.2, the bus currents at the nodes where the HVDC line is connected cannot be considered constant. Therefore, we introduce an additional current injection vector, so that the bus currents after the fault to be equal to:

𝐼𝐹 = 𝐼0 + 𝐼𝑇. Eq. 3.25 The vector 𝐼𝑇 has non-zero values only at the elements which correspond to the converter nodes of the HVDC line, i.e. nodes 𝐶𝑘 and 𝐶𝑚 in Figure 3.50. Consequently, Eq. 3.21 changes to:

𝛥𝑉 = 𝑉0 − 𝑉𝐹 = −𝑍0 ∙ �𝐼𝑆 + 𝐼𝑇�. Eq. 3.26 The additional currents have also an effect on the calculation of the injection currents 𝐼𝑆𝑙 and 𝐼𝑆𝑝. Eq. 3.23 and Eq. 3.24 change to:

𝐼𝑆𝑙 = 𝐼𝑙𝑝,0 − 𝐷𝑙𝑝,𝑙 ∙ 𝐼𝑆𝑙 − 𝐷𝑙𝑝,𝑝 ∙ 𝐼𝑆𝑝 − 𝐷𝑙𝑝 ∙ 𝐼𝑇 Eq. 3.27 𝐼𝑆𝑙 = 𝐼𝑙𝑝,0 − 𝐷𝑝𝑙,𝑙 ∙ 𝐼𝑆𝑙 − 𝐷𝑝𝑙,𝑝 ∙ 𝐼𝑆𝑝 − 𝐷𝑝𝑙 ∙ 𝐼𝑇 Eq. 3.28

In the post-contingency state, we need both HVDC converters to operate under a PV control mode in order to accomplish two functions. First, they should keep the voltage constant to a specified value at both line ends. Second, they must control the active power flowing through the HVDC line in an appropriate way, in order to relieve AC lines from possible overloadings. Eq. 3.29 refers to the voltage constraints, while Eq. 3.30 – Eq. 3.32 express the constraints for active power. In Eq. 3.29 – Eq. 3.32, an index 𝑖 to a matrix refers to a whole row, while an index 𝑖 to a vector refers to a single element.

�𝑉𝐹𝑖� = �𝑍𝐹𝑖 ∙ �𝐼0 + 𝐼𝑇�� = �𝑉𝑖,𝑠𝑝𝑒𝑐� ⇒

�𝑉𝐹𝑖�2 = �𝑍𝐹𝑖 ∙ �𝐼0 + 𝐼𝑇��𝑍𝐹𝑖 ∙ �𝐼0 + 𝐼𝑇��

∗ =

= �𝑍𝐹𝑖 ∙ 𝐼0�2 + �𝑍𝐹𝑖 ∙ 𝐼0��𝑍𝐹𝑖 ∙ 𝐼𝑇�

∗ + �𝑍𝐹𝑖 ∙ 𝐼0�∗�𝑍𝐹𝑖 ∙ 𝐼𝑇� + �𝑍𝐹𝑖 ∙ 𝐼𝑇�

2��𝑛𝑒𝑔𝑙𝑒𝑐𝑡𝑒𝑑

=

�𝑉𝑖,𝑠𝑝𝑒𝑐�2.

Eq. 3.29

Here, it should be noted that while Eq. 3.29 refers to the nodes 𝑘,𝑚 of the AC system, Eq. 3.30 – Eq. 3.32 refer to the ‘virtual’ converter nodes 𝐶𝑘 and 𝐶𝑚.



IRENE-40

Page: 106 of 383




𝑃𝐶𝑖,𝑃𝐶𝐶 = 𝑅𝑒 �𝑉𝐹𝐶𝑖 ∙ �𝐼0𝐶𝑖 + 𝐼𝑇𝐶𝑖�∗� = 𝑅𝑒 �𝑍𝐹𝐶𝑖 ∙ �𝐼0 + 𝐼𝑇� ∙ �𝐼0𝐶𝑖 + 𝐼𝑇𝐶𝑖�

∗�

=

𝑅𝑒 ��𝑍𝐹𝐶𝑖 ∙ 𝐼0� ∙ 𝐼0𝐶𝑖∗ + �𝑍𝐹𝐶𝑖 ∙ 𝐼0� ∙ 𝐼𝑇𝐶𝑖

∗ + �𝑍𝐹𝐶𝑖 ∙ 𝐼𝑇� ∙ 𝐼0𝑖∗ + �𝑍𝐹𝐶𝑖 ∙ 𝐼𝑇� ∙ 𝐼𝑇𝐶𝑖

∗��

𝑛𝑒𝑔𝑙𝑒𝑐𝑡𝑒𝑑

�.

Eq. 3.30

In Eq. 3.29–Eq. 3.30, we neglect the values of the terms where the injection currents appear squared, as they are comparably low. The elements of the 𝐼𝑇 vector, not corresponding to a converter should be set to zero. Equivalently to Eq. 3.11 for the non-faulted state, Eq. 3.31 requires that the change in the active power flow after the fault must also be balanced. Here, the effect of the DC line losses, which have a comparably small contribution, is neglected. As implied by Eq. 3.30 and Eq. 3.31, an additional optimization variable 𝑃𝐶𝑘,𝑃𝐶𝐶 is added to the optimization problem for each considered outage, bounded through Eq. 3.32.

𝑃𝐶𝑘,𝑃𝐶𝐶 = −𝑃𝐶𝑚,𝑃𝐶𝐶 Eq. 3.31 −𝑃𝐻𝑉𝐷𝐶,𝑙𝑖𝑚𝑖𝑡 ≤ 𝑃𝐶𝑘,𝑃𝐶𝐶 ≤ 𝑃𝐻𝑉𝐷𝐶,𝑙𝑖𝑚𝑖𝑡 Eq. 3.32 Hence, splitting the equation system Eq. 3.27 – Eq. 3.30 into real and imaginary parts, and forming the linear system of Eq. 3.33, the currents 𝐼𝑆 and 𝐼𝑇 can be determined. Here, 𝐴𝑖𝑐 is derived directly from the linear terms, while 𝑏𝑖𝑐 from the constant terms of Eq. 3.27 – Eq. 3.30. Subsequently, the new line currents in the faulted case can be determined through Eq. 3.34. The entries in 𝐼𝑙𝑖𝑛𝑒 that correspond to the outaged line in both directions are equal to the virtual currents which are eliminated from the injection currents 𝐼𝑆. Thus, for correctness these entries should be set to zero.

𝐴𝑖𝑐

⎝

⎜⎛𝑅𝑒�𝐼𝑆�𝐼𝑚�𝐼𝑆�𝑅𝑒�𝐼𝑇�𝐼𝑚�𝐼𝑇�⎠

⎟⎞ = 𝑏𝑖𝑐 , Eq. 3.33

𝐼𝑙𝑖𝑛𝑒 = 𝐼𝑙𝑖𝑛𝑒,0 − 𝛥𝐼𝑙𝑖𝑛𝑒= 𝐼𝑙𝑖𝑛𝑒,0 + 𝑌𝐿 ∙ 𝑍0 ∙ �𝐼𝑆 + 𝐼𝑇�

Eq. 3.34

3.2.5.3.3 Extension of the Basic Current Injection Method for Increased Accuracy In [14] an extension to the current injection method was proposed in order to improve its accuracy. The reasoning lies on the fact that the bus currents do change after the fault. The authors distinguish between the three different types of buses and form the equations, which are restated here, in Eq. 3.35 –Eq. 3.37, for the sake of completeness. The vector 𝐼𝑆 has, again, non-zero values only at the nodes of the outaged line. On the other hand, due to the extension, the only zero elements of the vector 𝐼𝑇 are the ones corresponding to PQ buses without loads. For the slack bus, Eq. 3.35 should be satisfied, while for the PV buses, Eq. 3.37 and Eq. 3.36 should be fulfilled. Finally, Eq. 3.36 and Eq. 3.38 should hold for PQ buses. For further details the interested reader can refer to [14].

𝛥𝑉𝑖 = −𝑍0 ∙ �𝐼𝑆 + 𝐼𝑇� = 0 Eq. 3.35 𝑃𝑖 = 𝑅𝑒�𝑉𝐹𝑖 ∙ �𝐼0𝑖 + 𝐼𝑇𝑖�∗� Eq. 3.36

�𝑉𝐹𝑖� = �𝑍𝐹 ∙ �𝐼0 + 𝐼𝑇�� = �𝑉0𝑖� Eq. 3.37 𝑄𝑖 = 𝐼𝑚�𝑉𝐹𝑖 ∙ �𝐼0𝑖 + 𝐼𝑇𝑖�∗� Eq. 3.38



IRENE-40

Page: 107 of 383




Due to the extension, the HVDC converters are not bound any more to operate in PV control mode, but we can rather distinguish between two alternatives. If the HVDC converter is connected to a PQ bus, then it should operate in a PV mode, in order to keep the post-fault bus voltage to the specified value. If, however, the converter is connected to a PV bus, then the voltage will be kept steady from the voltage controller (e.g. AVR of a generator) that is connected at that bus. Thus, the converter can operate in a PQ control mode with the objective to keep the reactive power injection/withdrawal steady at the pre-fault level. The formulation of such a constraint is similar to Eq. 3.38. Since operating in a PV control mode after an outage usually requires the injection of reactive power, thus limiting the active power transfer capability of the HVDC converter, a PQ control mode could allow an increased active power transfer.

3.2.5.4 Case Studies

A Case Study #1 and Case Study #2 were run and complemented with Case Study #3: the European Network.

3.2.5.5 Conclusions for the SC-OPF method

In this chapter, a Security-Constrained Optimal Power Flow including post-contingency control of HVDC lines was described. HVDC lines based on the Voltage-Source Converter technology are expected to play a significant role in the future power systems. The algorithm proposed here integrates the control capabilities of the VSC-HVDC technology in a single optimization problem, without the need to solve a detailed power flow or OPF for each contingency. The approach is based on the current injection method and takes full advantage of the VSC-HVDC line flexibility. It allows setting post-contingency voltage setpoints at the HVDC nodes and determines the optimal active HVDC power flow – pre- and post-fault – in order to avoid line overloads. After a contingency, each HVDC converter side can be controlled either in PV or in PQ control mode, depending on the type of nodes it is connected. The algorithm can be used for both power system operation and planning studies. The term ‘Cost of Security’ is introduced for planning studies, which provides a quantitative index for evaluating infrastructure reinforcement measures, with respect to power system security. In [W3ET TN 3004] B three case studies were presented, which examined the performance of the algorithm. It was shown that the current injection method approximates the post-fault line flows sufficiently well, especially towards the higher line loadings. Furthermore, taking advantage of the HVDC post-contingency control, the second case study showed that a significant decrease in the ‘Cost of Security’ can be achieved. The third case study presented the application of the algorithm on a power system model which studies the cross-border power flows of the European power system. Multiple HVDC lines and critical contingencies were included in the SC-OPF computation. Results about the ‘Cost of Security’ and HVDC post-contingency setpoints for a single snapshot were provided. Future work will include the detailed evaluation of different expansion scenarios in



IRENE-40

Page: 108 of 383




the European power system, applying the proposed algorithm and determining the solution which leads to the least ‘cost of security’.

3.2.6 Expansion scenarios – theoretical approach

3.2.6.1 Concerns about the AC expansion

In the initial design of the expansion scenarios, we assumed that the new AC lines should have a length longer than the existing interconnections. In this way, we could bypass certain weak network points within the countries, and increase the interconnection capacity. Such lines represent for example the connection of a generation-rich area in one country to a load-rich area in another; or a connection between two non-congested substations. Such lines could also resemble, for example, an HVAC Supergrid. Therefore, we assumed a line length of 500 km. Nevertheless, the results showed that the new lines were relatively underutilized. In Section 3.2.6.2 we give an explanation about the reasons. In Section 3.2.6.3 we show that due to the simplifications of our single-node per country model, we also need to represent the 500 km line with an equivalent. Table 3.33 shows the annual generation costs for three expansion scenarios. Except for the Base Case, in the second case we study the addition of three new HVDC lines (BritNed, NorGer and SE-LT) and the building of a new line between Lithuania and Poland. In the third case, we assume the further building of 10 new AC lines 400 kV, 3000 MVA of 500 km.

Table 3.33 – Annual Generation Costs for the RES 2050 Scenario. The new AC and HVDC lines have each a length of 500 km.

Base Case 9 HVDC lines (+1 AC LT-PL)

10 AC lines 400 kV + 9 HVDC

106.31 billion € 98.75 billion € 98.53 billion € From Table 3.33 it is evident that the addition of the 10 new lines did not have a significant effect in the sinking of the generation costs, which implies that it did not help towards congestion relief. Preliminary Conclusions: In Sections 3.2.6.2 and 3.2.6.3 we elaborate a little bit more on this case. If we take into account the network impedances within the countries, results improve, but still it seems that a – non-controllable – AC Supergrid is a sub-optimal solution. Our preliminary conclusions could be summarized as follows:



IRENE-40

Page: 109 of 383




- By using a simplified power system model, where only the interconnections are modeled and the connections within a country are assumed to have zero impedance, the assumption of lines longer than the actual interconnections is not valid. In case long AC lines need to be modeled in parallel with the interconnections, an equivalent impedance for the part of the line inside the countries’ network needs to be derived.

- If we need to go for long AC lines, then appropriate compensation mechanisms should be taken into account. As a result, a long line could be modeled as line with length equal to the interconnection length.

- Based on our theoretical considerations and our simulation results, an AC Supergrid is probably not the best solution either with 400kV or with 750kV lines.

- Although a solution could be the significant reinforcement of the power networks within the countries, we show that for building the same kilometers of lines, an AC Supergrid is still more preferable.

- Probably the best solution, from both an economic and security point of view, is controllable flows: a Supergrid based on HVDC lines, or on controllable AC lines (AC with FACTS) seems to be a better solution with significant effect on cost minimization.

3.2.6.2 Theoretical Considerations for the AC expansion

In this part, we study the addition of one new AC line in parallel to the existing interconnections. Here we assume that the length of the new line is similar to the existing interconnections. Such lines can represent long AC lines which connect two non-congested network points with the necessary series compensation equipment. Figure 3.52 represents such a case, where all the existing interconnections are represented by the equivalent impedance 𝑍𝑒𝑥𝑡. Impedance 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 stands for the impedance of the new parallel line.

Figure 3.54 – Representation of the addition of a new AC line 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 in parallel with the existing interconnection 𝑍𝑒𝑥𝑡. If 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 > 𝑍𝑒𝑥𝑡 and 𝐼𝑛𝑒𝑤,𝑙𝑖𝑚 ≥ 𝐼𝑒𝑥𝑡,𝑙𝑖𝑚 = 𝑁𝑇𝐶, then the already existing

interconnection will reach its capacity limit before the new line.

According to the current divider rule, 𝐼𝑒𝑥𝑡 = 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

𝑍𝑒𝑥𝑡+𝑍𝑛𝑒𝑤,𝑒𝑥𝑡𝐼𝑡𝑜𝑡 and 𝐼𝑛𝑒𝑤 = 𝑍𝑒𝑥𝑡

𝑍𝑒𝑥𝑡+𝑍𝑛𝑒𝑤,𝑒𝑥𝑡𝐼𝑡𝑜𝑡. This

yields:

𝐼𝑛𝑒𝑤 = 𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

𝐼𝑒𝑥𝑡 Eq. 3.39



IRENE-40

Page: 110 of 383




Assumptions: 1. The thermal limit of the existing interconnections ( 𝑍𝑒𝑥𝑡 ) is assumed equal to the

Net Transfer Capacity (NTC), i.e. 𝐼𝑒𝑥𝑡 ≤ 𝑁𝑇𝐶 . 2. We add a new line in parallel with the interconnection; the length of this line is

equal to the average length of all the 400 kV lines connecting the two countries. This line represents, in effect, longer AC lines with series compensation.

3. Normally, it should hold 𝐼𝑡𝑜𝑡 ≤ 𝑁𝑇𝐶. In this section we assume that along with the addition of the new line, reinforcements in the internal network are being carried out, so much that the additional flows over the new line to be sustained. However, we still maintain the condition, that over the existing interconnections the power flow should not exceed the NTC value.

4. 𝑍𝑒𝑥𝑡 ≤ 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡; This is because 𝑍𝑒𝑥𝑡 is the equivalent impedance of all the existing parallel interconnecting lines between two countries in 220 kV and 400 kV, while 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 represents only one line in 400 kV.

5. The capacity limit for the already existing 𝑍𝑒𝑥𝑡 is the corresponding NTC value. For the sake of simplicity we will assume in this section that the capacity limit of the new line 𝐼𝑛𝑒𝑤,𝑙𝑖𝑚 is large enough so that 𝑁𝑇𝐶 ≤ 𝐼𝑛𝑒𝑤,𝑙𝑖𝑚.

In realistic cases where assumption (5) may not hold, we will see that the limiting factor in these cases is not the capacity, but the ratio 𝑍𝑒𝑥𝑡/𝑍𝑛𝑒𝑤,𝑒𝑥𝑡. Considering the above assumptions, it holds:

𝐼𝑛𝑒𝑤 ≤ 𝐼𝑒𝑥𝑡 ≤ 𝑁𝑇𝐶 Eq. 3.40

As a result, the total current that can flow from one country to the other, after the transmission expansion, has an upper bound of:

𝐼𝑡𝑜𝑡 ≤ 2 ∙ 𝑁𝑇𝐶 Eq. 3.41

Indeed, as shown in

Table 3.34, all the new 400 kV/3000 MVA lines that we added in the different generation scenarios have impedances which are higher than the equivalent impedance of the corresponding interconnection. The only exception is the line between Hungary and Serbia, which is a single 400 kV line. This should probably occur due to inaccuracies in the interconnection lengths. It should be



IRENE-40

Page: 111 of 383




noted that we have assumed the minimum possible inteconnection length between the two substations (a direct line). This implies that we have determined the minimum possible 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 and, in turn, the maximum possible 𝑍𝑒𝑥𝑡/𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 values. In other words, the calculated currents through the additional lines should be considered optimistic. Even in the specific case nevertheless, it holds 𝐼𝑡𝑜𝑡 = 2.16 ∙ 𝑁𝑇𝐶, so the total capacity is only 8% higher than the upper bound as defined in Eq. 3.

Table 3.34 – Parallel AC lines 400 kV/3000 MVA that were added to relieve overloading during the five generation scenarios

From To 𝑍𝑒𝑥𝑡/𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 Length11 NTC

(MVA;2010) (km)

1 AT CZ 0.86 90 900 2 AT HU 0.59 120 325 3 AT SI 0.48 60 900 4 BE FR 0.14 60 3375 5 BG RO 0.53 150 500 6 EE LV 0.56 50 500 7 FI SE 0.98 250 2050 8 FR DE 0.21 60 3125 9 FR IT 0.21 120 2525

10 FR ES 0.31 90 1250 11 DE PL 0.51 15 1150 12 DE CH 0.14 20 3800 13 GR MK 0.48 70 300 14 HU RS 1.16 50 600

11 The length is the average interconnection length between two 380/400 kV substations in the respective countries. The interconnection length is equal to the length of a direct line between the two substations. This implies that the actual interconnecting line will be possibly longer than the assumed length in Table 3.34. Equally,



IRENE-40

Page: 112 of 383




15 IT SI 0.43 60 380 16 PL SK 0.64 100 600 17 PT ES 0.25 40 1800 18 RO RS 0.60 10 600

In Figure 3.55 we plot the loadings of the additional lines calculated through Eq. 3.39, against the actual loadings from our simulations. We used the RES 2050 Scenario; the plotted actual loadings correspond to the maximum loadings that occur in these lines during the month of January.

Figure 3.55 – Loadings of Additional Lines: Comparison of theoretical results based on Eq. 3.39

(New Line: Theoretical Max Loading) and actual results from simulations (New Line: Actual Max. Loading). The total loading of the interconnection is the sum of the “Existing NTC Value” and the “New Line: Actual Max. Loading” [RES 2050 Scenario]. The “Existing NTC Value” is equal to

1.43*NTC(2010), where NTC(2010) is given in T2.3 Report.

0

500

1000

1500

2000

2500

3000

3500

AT-HU AT-SI BG-RO FR-ES DE-PL IT-SI PT-ES RO-RS

Pow

er (M

VA)

RES 2050 OPF January - Max. Loading of New Lines

Existing NTC Value

New Line: Theoretical Max. Loading

New Line: Actual Max. Loading

New Line Thermal Limit



IRENE-40

Page: 113 of 383




Figure 3.56 – Relationship between the impedance ratio and the net transfer capacity. The dashed

line illustrates the trendline.

In Figure 3.54 the impedance ratio of

Table 3.34 is plotted against the Net Transfer Capacity of the Lines. As the trendline shows, interconnections with higher NTCs usually have a lower impedance ratio. Here, one could claim that lines with lower thermal capacity could be installed, as in any case the capacity would not be used. However, lines with lower thermal capacity usually have increased R and X values. As a result, the impedance ratio decreases further, and less current can flow through the new line. Therefore, lines with higher capacity – and lower R, X, and B values – should be selected for installation.

3.2.6.3 Accounting for the internal reinforcements in the AC expansion

In our example in

0

0,2

0,4

0,6

0,8

1

1,2

1,4

0 500 1000 1500 2000 2500 3000 3500 4000

Zext

/ Zn

ew,e

xt

Net Tranfer Capacity (NTC, 2010) [MVA]

Impedance Ratio vs. Line Capacity



IRENE-40

Page: 114 of 383




Table 3.33, we assumed that all the new AC lines had a length of 500 km, resembling the creation of an AC Supergrid, or generally the bypass of internal weak network points. An equivalent of such a circuit is illustrated in Figure 3.55.

Figure 3.57 – Adding a long parallel AC line to an existing interconnection.

Here, 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 corresponds only to the length of the line parallel to the existing interconnections. The rest of the line is represented by 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡. As the main part of the line is actually within the regional networks it holds 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 > 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡. In the expansion scenario we studied, there were cases where 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 > 10 ∙ 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡, e.g. 460 km vs. 40 km in the case of PT-ES line. It is the easily shown that

𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

≪ 𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

Eq. 3.42

Eq. 3.42, along with Eq. 3.39, explain why all the parallel lines we added in our expansion scenario had a significantly low loading. Until now we have not incorporated the internal impedances (i.e. lines) of the regional networks in our considerations. In what follows, the effect of the network internal impedance will be considered.

Figure 3.58 – Illustration of the additional internal impedance.

For the sake of simplicity, we will assume that the circuit illustrated in Figure 3.56 represents only half of the interconnection. Point A is a substation somewhere within country A, while point B is



IRENE-40

Page: 115 of 383




the midpoint of the interconnecting line. Exactly the same considerations, as the ones that follow, can be made for the other half of the line. Due to the symmetry, all the relationships and the conclusions we derive, can be applied without further adaptation to the whole interconnection. In Figure 3.56, 𝑍𝑖𝑛𝑡 corresponds to the equivalent internal impedance of the existing system from point A till the beginning of the interconnecting line. 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 corresponds to the additional line length which, according to our assumptions, should be built within the country. 𝑍𝑒𝑥𝑡 and 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 have the same values as in Figure 3.55. Assumptions: 6. 𝑍′𝑒𝑥𝑡 ≤ 𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡; here, 𝑍′𝑒𝑥𝑡, 𝑍′𝑒𝑥𝑡,𝑛𝑒𝑤 stand for impedance per unit length. As

already described in the previous section there are several parallel lines interconnecting most countries. As a result, assuming that 𝑍𝑒𝑥𝑡, 𝑍𝑒𝑥𝑡,𝑛𝑒𝑤 have the same length 𝑙𝑒𝑥𝑡, the impedance of the new line is greater.

7. 𝑍′𝑖𝑛𝑡 ≤ 𝑍′𝑒𝑥𝑡 (per unit length); Normally the network within a country is much more meshed than outside the country. So, from a point A within the country, the current would have several paths to follow in order to reach an interconnection substation. As a result, per unit length the impedance 𝑍𝑖𝑛𝑡 should be smaller than the impedance 𝑍𝑒𝑥𝑡. The relationship between 𝑍𝑖𝑛𝑡 and 𝑍𝑒𝑥𝑡in absolute values depends on the length of the interconnections and the distance of point A from the borders.

8. 𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡 = 𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡 per unit length since it is the same line, but 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 ≥𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 in absolute values; assume that we have a 500 km line of which only 100 km represent the interconnection. As we assume symmetry and consider only half of the line in this case, this means that 200 km are within the country, while 50 km is the distance from one country to the midpoint of the interconnecting line. As the line segment within the country is much longer, the respective impedance should be greater.

From assumptions (6), (7), (8), it holds:

𝑍′𝑖𝑛𝑡 ∙ 𝑙𝑖𝑛𝑡𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡 ∙ 𝑙𝑖𝑛𝑡

≤ 𝑍′𝑒𝑥𝑡 ∙ 𝑙𝑒𝑥𝑡𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡 ∙ 𝑙𝑒𝑥𝑡

⇒

𝑍𝑖𝑛𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡

≤ 𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

Eq. 3.43

From known properties of fractions, if Eq. 3.43 is true then it holds:



IRENE-40

Page: 116 of 383





≤ 𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡


Eq. 3.44

Taking into account the circuit in Figure 3.55, it can easily be derived that:

𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

≤ 𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡


Eq. 3.45

Applying again the current divider rule, similar to Eq. 3.39 we obtain:

𝐼𝑛𝑒𝑤 = 𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡


With Eq. 3.45 and Eq. 3.46 it is shown that if we take into account the impedances within the country, the current that will flow over the new line will be greater than what we would calculate with our modelling approach, i.e. assuming a zero internal impedance of the regional networks. This means that, in total, the new lines would be in reality better utilized, leading to an improved congestion relief and lower generation costs than computed. Nevertheless, there is an upper bound of 𝐼𝑛𝑒𝑤 ≤ 𝑍𝑒𝑥𝑡

𝑍𝑛𝑒𝑤,𝑒𝑥𝑡𝐼𝑒𝑥𝑡. So, the resulting generation costs will certainly be more than in the case

where the new line is equally long with the existing interconnections. And for the new line flows, Eq. 3.41 still holds. This means that the total line flows in the “reinforced” interconnections cannot exceed the value of 2 ∙ 𝑁𝑇𝐶. Table 3.35 shows the computed generation costs for the two boundary cases. The true cost from the addition of the 11 HVAC 500km long lines would probably lie somewhere in the middle.

Table 3.35 – Comparing the effect of the additional lines' length on the annual generation costs


(500 km long)

10 HVAC lines + 9 HVDC (HVAC equal to

interconnection length) 106.31 billion € 98.53 billion € 95.29 billion €

Here, compensation equipment should be considered. By adding series compensation to the line, we not only assist in the increase of dynamic security, but also allow more current to flow through the



IRENE-40

Page: 117 of 383




line. A compensated long AC line can achieve similar power flows as a line with length equal to the interconnecting line. This implies that resulting annual generator costs can be similar to the third column of Table 3.35.

3.2.6.3.1 Accounting for the network internal impedance in our calculations If we want to include the effect of the network internal impedance in our calculations, then we should compute a new 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡

𝑒𝑞. , as illustrated in Figure 3.57.

Figure 3.59 – Accounting for the network internal impedance with an equivalent 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑒𝑞. .

The current 𝐼𝑒𝑥𝑡will reach first its capacity limit. Our goal here is to find an equivalent impedance 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑒𝑞. , in order to represent more accurately the relationship between the current that flows in the

existing interconnection and the current that flows in the new line. In other words, the ratio 𝐼𝑛𝑒𝑤/𝐼𝑒𝑥𝑡 should remain unchanged in both cases12

.

From Eq. 3.39 and Eq. 3.46 we obtain:

𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑒𝑞. + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

= 𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

Eq. 3.47

This yields:

𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑒𝑞. = 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 −

𝑍𝑖𝑛𝑡𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡

(𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡) Eq. 3.48

12 Reasons for not calculating the total equivalent impedance: An equal total equivalent impedance would mean that if we applied the same voltage, equal current will flow from both circuits. However, this is not our focus here. In our case, we assume that we have a certain amount of current that we want to pass through the circuits (one could equivalently say that we have a certain amount of power that we want to pass through the interconnection). The question we want to answer is how this current is divided between the existing interconnections and the new line. From that we can identify what are the limiting factors and which solution allows a greater amount of this power to flow through the circuit.



IRENE-40

Page: 118 of 383




Based on equation Eq. 3.48, we attempt a case study. We assume that 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡represents 80% of the total new line impedance 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡. This is equivalent to a new line of 500 km, of which 100 km are the interconnection length (𝑍𝑛𝑒𝑤,𝑒𝑥𝑡) and 400 km are within the regional networks (𝑍𝑛𝑒𝑤,𝑖𝑛𝑡) – 200 km in one side and 200 km in the other side, i.e. 𝑙𝑒𝑥𝑡 = 100 𝑘𝑚, 𝑙𝑖𝑛𝑡 = 400 𝑘𝑚. This results in:

𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑒𝑞. = �0.8 − 𝑍𝑖𝑛𝑡

𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡� 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

0.2 Eq. 3.49

We further assume that 𝑍′𝑖𝑛𝑡 = 0.5 ∙ 𝑍′𝑒𝑥𝑡 (see assumption (7)). This, roughly, is equivalent to the assumption that for each interconnecting line, there exist at least two lines of a similar voltage level inside the country. From Eq. 3.49 we can conclude that:

𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑒𝑞. = 0.67 ∙ 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 Eq. 3.50

From Eq. 3.50 it is obvious that the total equivalent impedance of new long parallel line would be about 𝑍𝑛𝑒𝑤𝑒𝑞. = 1.67 ∙ Znew,ext, and as a result , in reality, it will allow significantly less current than what we assumed in Section 3.2.6.1. In other words, in this specific case (80% of the new line in regional networks), it should hold Itot ≤ 1.6 ∙ NTC for the best case.

3.2.6.4 AC expansion: Long HVAC interconnections or network reinforcements within the country?

In this section we study the case where network reinforcements take place within the regional networks, in order to increase the interconnection capacity with the neighboring countries; in effect, the aim is to increase the NTC, so that Itot,rein ≥ Itot. We then compare this case with the case of long HVAC interconnection lines.



IRENE-40

Page: 119 of 383




Figure 3.60 – Illustration of internal network reinforcements.

Figure 3.58 illustrates the additional network reinforcements in the network. 𝑍𝑖𝑛𝑡 is the equivalent impedance of the internal network, while 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 represents the additional reinforcements. 𝐼𝑒𝑥𝑡 is the current that was flowing until now towards the interconnections. We assume here that the limiting factor in the current capacity is 𝑍𝑖𝑛𝑡 and as a result 𝐼𝑒𝑥𝑡 ≤ 𝑁𝑇𝐶. This is in line with our previous assumptions, as in Section 3.2.6.2 we assumed that all necessary internal reinforcements will be made, and in Section 3.2.6.3 we bypass the internal weak network points through a new line. From Kirchhoff’s laws and the current divider rule, it holds:

𝐼𝑟𝑒𝑖𝑛 = 𝑍𝑖𝑛𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡


𝐼𝑡𝑜𝑡,𝑟𝑒𝑖𝑛 = 𝐼𝑒𝑥𝑡 + 𝐼𝑟𝑒𝑖𝑛 ⇒ 𝐼𝑡𝑜𝑡,𝑟𝑒𝑖𝑛 = �1 + 𝑍𝑖𝑛𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡

� 𝐼𝑒𝑥𝑡 Eq. 3.52

We now compare Eq. 3.52 with the total current that flows when we add a new long transmission line. From Eq. 3.46 we get:

𝐼𝑡𝑜𝑡,𝑙𝑜𝑛𝑔 = 𝐼𝑒𝑥𝑡 + 𝐼𝑛𝑒𝑤 ⇒ 𝐼𝑡𝑜𝑡,𝑙𝑜𝑛𝑔 = �1 + 𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 + 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

� 𝐼𝑒𝑥𝑡 Eq. 3.53

Without loss of generality, we now study the cases where the same current flows through the existing internal connections. This means that the current 𝐼𝑒𝑥𝑡 that flows through 𝑍𝑖𝑛𝑡 is the same in both Eq. 3.52 and Eq. 3.53. Especially at the upper limit, where in both cases 𝐼𝑒𝑥𝑡 = 𝑁𝑇𝐶 this assumption will hold true. From Eq. 3.44 it can be derived that:

𝐼𝑡𝑜𝑡,𝑟𝑒𝑖𝑛 ≤ 𝐼𝑡𝑜𝑡,𝑙𝑜𝑛𝑔 Eq. 3.54



IRENE-40

Page: 120 of 383




Eq. 3.54 will be true, as long as the condition in Eq. 3.43 is fulfilled. For Eq. 3.43 the following holds:



≡ 𝑍′𝑖𝑛𝑡𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡

≤ 𝑍′𝑒𝑥𝑡𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡

,

where the 𝑍′ represent the impedance values per unit length (e.g. per km). With 𝑍′𝑖𝑛𝑡 ≤ 𝑍′𝑒𝑥𝑡 and 𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡 = 𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡 (see Assumptions (7) and (8)), Eq. 3.43 will hold and as a result 𝐼𝑡𝑜𝑡,𝑟𝑒𝑖𝑛 ≤ 𝐼𝑡𝑜𝑡,𝑙𝑜𝑛𝑔. In other words, if the equivalent impedance of the internal reinforcements 𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡 is not less than the impedance of the external segment of the parallel line, then more current will flow in the case of the long transmission line. For 𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡 to be smaller than 𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡, we need to install more than one lines in parallel in the internal network. This means that if we decide for the option of internal network reinforcements, we will need to build more kilometers of lines in parallel, than the kilometers of line we will need in the case of a long AC line, in order to achieve the same result. This is a logical conclusion, if we assume the extreme case, where we do no reinforcements. Then, of course, a long transmission line in parallel will transmit additional amounts of power through the interconnection. What is important here though is that we have shown that the option of internal network reinforcements will always need more kilometers of lines, which should be connected in parallel, in order to achieve the same result as a long AC line – for the same distance. Translating these conclusions to numbers, assume that we want to reinforce the interconnection Germany-Switzerland with a 400 kV/3000 MVA transmission line. The length of the interconnection is 20 km. We also assume that the non-congested substation has a distance of 100 km from the interconnection point (e.g. somewhere inside Germany), while the internal network in Germany, being more meshed than the cross-border lines has an equivalent impedance 𝑍′𝑖𝑛𝑡 = 0.5 ∙𝑍′𝑒𝑥𝑡. Then we should install at least 2 parallel 400 kV lines from the substation to the German border, so that 𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡 = 0.5 ∙ 𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡 and Eq. 3.43 may no longer hold. Otherwise, it is preferable to build a direct line from that substation in Germany to Switzerland. Conclusions (in addition to the preliminary conclusions in Section 3.2.6.1):

1. An AC expansion, which assumes that only the congested interconnecting lines should be reinforced, and the full capacity of the new parallel lines can be used, is probably not valid.

2. The AC expansion should also account for reinforcement within a country’s network, or bypass the weak network points.

3. When selecting the lines to be reinforced, except for the average loading one should also look at the impedance values.



IRENE-40

Page: 121 of 383




4. Controllable flows (HVDC or AC with FACTS) can most probably overcome the limitations imposed from the AC grid.

Results with HVDC lines:

Table 3.36 – Comparison table of generation costs for different expansion scenarios


(500 km long)

10 HVAC lines + 9 HVDC

(HVAC equal to interconnection length)

19 HVDC lines (500 km length)

106.31 billion € 98.53 billion € 95.29 billion € 88.74 billion €

3.2.6.5 Line Upgrades

Based on our considerations, upgrading an existing line, e.g. from 220 kV to 400 kV, is less effective than adding a parallel 400 kV line in terms of total power capacity. Assume an interconnection consists of one 220 kV line and one 400 kV line. By upgrading we get two 400 kV lines. By adding a line, we get one 220 kV and two 400 kV lines. Obviously, the second case will allow more current to flow through. The benefit of upgrading is that it allows a better utilization of the lines. Assuming a given amount of power, below the thermal limits in both cases, then adding a new line results in relatively lower line loadings, especially for the 220 kV line. According to draft calculations, based on the impedance values of the lines, given in the IRENE-40 database, the fraction of the power transmitted over the 220 kV line will be about 11%, while each of the two 400 kV lines will carry 44.5% of the exchanged power. In the case of upgrading, the two 400 kV lines will each share a loading of 50%. In terms of total power capacity though, adding a new 400 kV line will result in a capacity 1.13 times higher that upgrading. If this extra 13% of capacity is not significant, then a line upgrade would make more sense. Nevertheless, it should be here noted that for security considerations, adding a new line will be always preferred against a line upgrade, as this better assists in the fulfillment of the N-1 criterion. More specifically, assuming that there are no more parallel lines, this N-1 calculation should be based on the worst N-1 case which is the loss of one 400kV line. Thus, in the case of two 400kV lines, the transfer capability of the interconnection will be that of one 400kV (with one line outaged). In the case of two 400kV and one 220kV line the transfer capability will be that of one 220kV and one 400kV line (with one 400kV line outaged).

3.2.6.6 AC-750 kV Lines

If instead of 400 kV Lines we used 750 kV lines, the results will be obviously better, but still the improvement will not be incredibly large. Table 3.37 presents the results from an AC-400 kV expansion, an AC-750 kV expansion and an HVDC expansion. The AC-750 kV expansion results in about 1.3 billion Euros additional generation cost savings per year. The HVDC scenario, on the



IRENE-40

Page: 122 of 383




other hand, achieves a cost decrease about five times as much as the AC-750 kV scenario (6.5 billion Euros/year).

Table 3.37 – RES 2050: OPF Results from different expansion scenarios

Base Case 10 EHVAC lines

400 kV 3000 MVA + 9 HVDC

10 UHVAC lines 750 kV 3900 MVA

+ 9 HVDC 19 HVDC lines

106.31 billion € 95.29 billion € 94.01 billion € 88.74 billion € According to draft calculations, based on the impedance values of a double-circuit 400 kV/3000 MVA line, and a single-circuit 750 kV/3900 MVA line, all the limits we have calculated for the additional 400 kV lines can be increased by a factor of 1.61. This means that the average limit of the 18 interconnections will increase from 𝐼𝑡𝑜𝑡 ≈ (1 + 0.5) ∙ 𝑁𝑇𝐶 to 𝐼𝑡𝑜𝑡 ≈ (1 + 0.8) ∙ 𝑁𝑇𝐶 if we add a parallel 765 kV line. Minimum and maximum limits for the 18 interconnections we studied will change from 1.14*NTCÆ1.23*NTC and from 2.16*NTCÆ2.87*NTC. Here, it should be also taken into account that the power which will flow through the additional lines, depends also on the impedance ratio with respect to the parallel paths.



IRENE-40

Page: 123 of 383




3.2.7 Expansion scenarios – simulations

3.2.7.1 AC-OPF with soft constraints on the line flows to ensure no load shedding and convergence

For each generation scenario we started with the base configuration, i.e. the interconnections and based on the NTC values of 2010 (we have increased the NTC values by 140%, for the reasons stated in Section 0). Running an OPF through the whole year, we realized that there were certain hours that the load could not be served through the existing interconnections and generation capacities. Therefore, we modified our OPF algorithm, and transformed the line flow constraints (see also Eq. 3.6 and Eq. 3.8) into soft constraints. We imposed a heavy penalty factor on the soft variables, so that the OPF would be reluctant to exceed the transmission line limits, and only attempt it in case where the load cannot be served. After this set of simulations, we have identified which are the absolutely necessary line reinforcements, in order to ensure that the demand is covered for every hour during the year in Europe of 2050.

Table 3.38 – Line Reinforcements that are necessary in order to ensure that load is served 100% of the time in 2050.

Additional Capacity (MVA)

Assumed Apparent Power Capacity of

2010 (MVA)

Additional Capacity to ensure N-1 criterion

(MVA) BAU Scenario DK-NO 4081 950 4081+2000 = 6081 NL-NO 4854 700 4854+2400 = 7254 CCS Scenario GR-MK 229 429 429+0 = 0 DK-NO 3891 950 3891+2000 = 5891 GR-IT 351 500 351+0 = 701 NL-NO 4414 700 4414+2200 = 6614 DES Scenario DK-NO 3955 950 3955+2000 = 5955 NL-NO 4454 700 4454+2200 = 6654 EFF Scenario DK-NO 1701 950 1701+850 = 2551 NL-NO 2017 700 2017+1000 = 3017 RES Scenario BG-RO 40 715 40+0 = 40 BG-MK 144 643 144+0 = 144



IRENE-40

Page: 124 of 383




FR-ES 1084 1785 1084+0 = 1084 GR-AL 146 215 146+0 = 146 GR-MK 692 429 692+0 = 692 IT-SI 162 543 162+0 = 162 DK-NO 3952 950 3952+2000 = 5952 GR-IT 932 500 932 = 932 NL-NO 4432 700 4432+2200 = 6632 From Table 3.38 – Line Reinforcements that are necessary in order to ensure that load is served 100% of the time in 2050., it is observed that two line reinforcements appear on every generation scenario. These are the interconnections between Norway and Denmark, as well as between Norway and the Netherlands. However, the need for these reinforcements is partly exacerbated due to assumption of the generation scenarios that the hydros in Norway have a yearly availability of 58%. As a result, there are certain times during the year, when the generation capacity in Norway is not sufficient to cover the demand. In any case, however, these interconnections are still one of the most probable candidates to be reinforced until 2050 in order to ensure the security of supply. Another interesting remark regarding Table 3.38 – Line Reinforcements that are necessary in order to ensure that load is served 100% of the time in 2050. is that the RES scenario requires a significantly higher amount of network reinforcemens, in comparison with the other scenarios in order to ensure the security of supply. The RES Scenario projects that 80% of the electricity demand in Europe will be provided by Renewable – and mostly fluctuating – energy sources by 2050. As a result, it becomes evident that there will be a need for increased interconnection capacity between the countries, which will ensure that in case there is a shortage of power in one country, this can be covered through imported power from a region where there is excess generation. The additional line capacities that were determined during the AC-OPF calculations did not take into account the N-1 security criterion. Therefore, in the right column of Table 3.38 – Line Reinforcements that are necessary in order to ensure that load is served 100% of the time in 2050. the total additional transmission capacities are indicated, so that the N-1 criterion can be satisfied. Additional capacities are proposed only for the radial interconnections, namely DK-NO and NL-NO. The rest of the interconnections are part of a meshed network, so an analysis about the critical contingencies and an SC-OPF calculation will be necessary. The system with the additional line capacities, sometimes different for each generation scenario, serves as the base system for each expansion we will apply.



IRENE-40

Page: 125 of 383




3.2.7.2 Adding the necessary interconnection capacities and identifying the line reinforcements

Here we use the standard AC-OPF, as described in Section 3.2.5.2, with the additional line capacities, as shown in Table 3.38 – Line Reinforcements that are necessary in order to ensure that load is served 100% of the time in 2050.. Running the OPF for a whole year, we calculate the loading of each line for every hour and for every generation scenario. In a previous work (ref. [6]), we have concluded that the line reinforcements should be placed in parallel to lines that are most often congested (Note that, for example, this does not hold for FACTS devices. The placement of FACTS devices depends on the network topology and can be more effective on lines that are not necessarily congested). Therefore, we decided to reinforce the lines that are between 99% and 100% loaded for more than 4000 hours a year. Here, it should be noted that we did not select the average line loading as a criterion for line reinforcements. Although both criteria - hours per year above 99% loaded and average line loading – would often result in the same line reinforcements, cases might exist where a line is very often loaded to e.g. 80% of its capacity, but only seldom loaded near 100%. Such lines demonstrate good line utilization, but do not need reinforcement, as they are not congested. As we will see in Section 4.2.1, we do use the average line loading in order to determine the critical contingencies that we will incorporate in the Security-Constrained OPF.



IRENE-40

Page: 126 of 383




Figure 3.61 – BAU 2050 Scenario: Duration of Line Loadings in hours per year. The line loadings

are expressed in %. Lines that are loaded above 99% for more than 4000 hours per year are selected for reinforcement

Table 3.39 – BAU 2050: Location and transmission capacity of the line reinforcements

BAU 2050 Scenario Expansion on Land Capacity [MVA]

(AC-400kV or HVDC / AC-750kV)

Submarine Cables Capacity [MVA] HVDC cables

AT-DE 3000/3900 NL-UK Planned: 1000 AT-HU 3000/3900 DE-NO Planned: 1400 FR-ES 3000/3900 LT-SE Planned: 700 FR-CH 3000/3900 EE-FI 3000 DE-CH 3000/3900 FR-UK 3000



IRENE-40

Page: 127 of 383




GR-MK 3000/3900 IE-UK 3000 IT-SI 3000/3900 DE-SE 3000 RO-RS 3000/3900 GR-IT 3000 LT-PL 3000/3900 NL-NO 3000 PL-SE 3000

Figure 3.62 – CCS 2050 Scenario: Duration of Line Loadings in hours per year. The line loadings


Table 3.40 – CCS 2050: Location and transmission capacity of the line reinforcements

CCS 2050 Scenario Expansion on Land Capacity [MVA] Submarine Cables Capacity [MVA]



IRENE-40

Page: 128 of 383





HVDC cables

AT-DE 3000/3900 NL-UK Planned: 1000 FR-ES 3000/3900 DE-NO Planned: 1400 IT-SI 3000/3900 LT-SE Planned: 700 PT-ES 3000/3900 EE-FI 3000 RO-RS 3000/3900 FR-UK 3000 LT-PL 3000/3900 IE-UK 3000 DK-NO 3000 GR-IT 3000 NL-NO 3000 PL-SE 3000

Figure 3.63 – DES 2050 Scenario: Duration of Line Loadings in hours per year. The line loadings

are expressed in %. Lines that are loaded above 99% for more than 4000 hours per year are



IRENE-40

Page: 129 of 383




selected for reinforcement

Table 3.41 – DES 2050: Location and transmission capacity of the line reinforcements

DES 2050 Scenario Expansion on Land Capacity [MVA]



AT-DE 3000/3900 NL-UK Planned: 1000 AT-SI 3000/3900 DE-NO Planned: 1400 FR-ES 3000/3900 LT-SE Planned: 700 FR-CH 3000/3900 EE-FI 3000 DE-PL 3000/3900 FR-UK 3000 DE-CH 3000/3900 IE-UK 3000 IT-SI 3000/3900 GR-IT 3000 PT-ES 3000/3900 NL-NO 3000 RO-RS 3000/3900 PL-SE 3000 LT-PL 3000/3900



IRENE-40

Page: 130 of 383




Figure 3.64 – EFF 2050 Scenario: Duration of Line Loadings in hours per year. The line loadings


Table 3.42 – EFF 2050: Location and transmission capacity of the line reinforcements

EFF 2050 Scenario Expansion on Land Capacity [MVA]



AT-DE 3000/3900 NL-UK Planned: 1000 BE-FR 3000/3900 DE-NO Planned: 1400 FR-ES 3000/3900 LT-SE Planned: 700 FR-CH 3000/3900 EE-FI 3000 DE-CH 3000/3900 FR-UK 3000



IRENE-40

Page: 131 of 383




GR-MK 3000/3900 IE-UK 3000 IT-SI 3000/3900 DE-SE 3000 RO-RS 3000/3900 GR-IT 3000 LT-PL 3000/3900 NL-NO 3000 PL-SE 3000

Figure 3.65 – RES 2050 Scenario: Duration of Line Loadings in hours per year. The line loadings


Table 3.43 – RES 2050: Location and transmission capacity of the line reinforcements

RES 2050 Scenario Expansion on Land Capacity [MVA]

(AC-400kV or HVDC Submarine Cables Capacity [MVA]

HVDC cables



IRENE-40

Page: 132 of 383




/ AC-750kV) AT-DE 3000/3900 NL-UK Planned: 1000 AT-HU 3000/3900 DE-NO Planned: 1400 FR-ES 3000/3900 LT-SE Planned: 700 FR-CH 3000/3900 EE-FI 3000 DE-NL 3000/3900 FR-UK 3000 DE-PL 3000/3900 IE-UK 3000 DE-CH 3000/3900 GR-IT 3000 IT-SI 3000/3900 NL-NO 3000 RO-RS 3000/3900 PL-SE 3000 LT-PL 3000/3900 In Table 3.44, we summarize the selected network reinforcements for all generation scenarios, and we rank them with respect to the number of scenarios they occur.

Table 3.44 – Classification and ranking of all the selected Network Reinforcements according to the number of generation scenarios they participate

Expansion on Land Submarine Cables AT-DE All EE-FI All FR-ES All FR-UK All IT-SI All IE-UK All RO-RS All GR-IT All LT-PL All (planned) NL-NO All FR-CH All – except CCS PL-SE All DE-CH All – except CCS NL-UK All (planned) AT-HU BAU+RES DE-NO All (planned) DE-PL DES+RES LT-SE All (planned) GR-MK BAU+EFF DE-SE BAU+EFF PT-ES CCS+DES DK-NO CCS AT-SI DES BE-FR EFF DE-NL RES As it can be observed, most of the additional lines are common to all generation scenarios. This implies that these reinforcements would be necessary for the European network, almost irrespective of the future generation mix in Europe. Especially, with the exception of two cables, all the submarine interconnections coincide for all generation scenarios. For the expansion on land, half of



IRENE-40

Page: 133 of 383




the proposed reinforcements should take place on all scenarios. The rest are specific according the generation scenario under study. Based on our considerations in Section 3.2.6, we selected the installation of long transmission lines in all cases. For all lines we assumed a standard length of 500 km, considering that such lines could bypass possible bottlenecks and connect two non-congested substations. The submarine cables are based on the VSC-HVDC technology for all expansion scenarios, as the AC cables can be laid for only limited lengths. For the “Expansion on Land” we distinguish between three expansion scenarios:

- AC 400 kV with compensation - AC 750 kV with compensation - VSC-HVDC

For both AC scenarios, we assumed long AC lines with compensation. The compensation is reflected in our assumption that the effective line length of the AC line – as represented by the R, X, and B values – is not 500 km but equal to the length of the actual interconnecting line (i.e., the average direct line length between the 400 kV substation that connect the two countries). The amount of compensation varies for each line (according to the length of each actual interconnecting line).

3.2.7.3 Series Compensation for the AC-Expansion Scenarios

As mentioned in the previous paragraph, for the AC expansion scenarios we assume that every long AC line, which we add, has series compensation installed. Due to our single-node-per-country model, calculating the exact degree of compensation for each line that will be necessary in reality is not straightforward. Since we have neglected the internal network in our simulations, claiming that this compensation is the impedance difference between the actual line length and the effective line length we have assumed in our studies would not be valid in reality. In fact, the actual degree of compensation should be also dependent on the internal network. As we will see in the following analysis, there are two main factors on which the degree of compensation depends. The first is how much more meshed is the internal network in comparison to the external interconnections. The second is how long is the part of the long AC line that connects the two border substations with respect to the total line length. We split the compensated new AC line in the external part Znew,ext, which corresponds to the interconnection length between the two regions, and the internal part Znew,int, which represents the part of the line inside the two countries. According to Eq. 3.55, without loss of generality, we



IRENE-40

Page: 134 of 383




assume that the whole amount of compensation is incorporated in the internal part, thus resulting in Znew,intcomp . Here Zcomp can be assumed equal to Zcomp = jXcomp.

𝑍𝑛𝑒𝑤𝑐𝑜𝑚𝑝 = 𝑍𝑛𝑒𝑤 − 𝑍𝑐𝑜𝑚𝑝 = 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 + 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡 − 𝑍𝑐𝑜𝑚𝑝��

𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑐𝑜𝑚𝑝

= 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 + 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑐𝑜𝑚𝑝

Eq. 3.55

Eq. 3.48 is rewritten as Eq.3.56, assuming now, however, a compensated line.

𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑒𝑞. = 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡

𝑐𝑜𝑚𝑝 − 𝑍𝑖𝑛𝑡𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡

(𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 + 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑐𝑜𝑚𝑝 ) Eq.3.56

Eq.3.56 can be rewritten in the form of Eq.3.57:

𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑒𝑞. = 𝑍𝑛𝑒𝑤,𝑖𝑛𝑡

𝑐𝑜𝑚𝑝 �1 − 𝑍𝑖𝑛𝑡𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡

� − 𝑍𝑖𝑛𝑡𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡

𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 Eq.3.57

As mentioned in Section 3.2.7.2, in our studies we have assumed that due to the compensation, the whole line has a length equal to the actual interconnection length. Thus, Znew,int

eq. = 0. As a result, for Eq.3.57 it holds:

𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑐𝑜𝑚𝑝 � 𝑍𝑒𝑥𝑡

𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡� = 𝑍𝑖𝑛𝑡

𝑍𝑖𝑛𝑡 + 𝑍𝑒𝑥𝑡𝑍𝑛𝑒𝑤,𝑒𝑥𝑡 Eq.3.58

As (Zint + Zext) ≠ 0, Eq.3.58 can be rewritten as Eq.3.59.

𝑍𝑛𝑒𝑤,𝑖𝑛𝑡𝑐𝑜𝑚𝑝

𝑍𝑖𝑛𝑡= 𝑍𝑛𝑒𝑤,𝑒𝑥𝑡

𝑍𝑒𝑥𝑡 Eq.3.59

In Eq.3.60 we express all the impedances in Eq.3.59 as the product of the “impedance per unit length” (𝑍′) times the length of the line (either the internal part or the external part).

𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡𝑐𝑜𝑚𝑝 ∙ 𝑙𝑖𝑛𝑡𝑍′𝑖𝑛𝑡 ∙ 𝑙𝑖𝑛𝑡

= 𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡 ∙ 𝑙𝑒𝑥𝑡𝑍′𝑒𝑥𝑡 ∙ 𝑙𝑒𝑥𝑡

⇒ 𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡𝑐𝑜𝑚𝑝

𝑍′𝑖𝑛𝑡= 𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡

𝑍′𝑒𝑥𝑡 Eq.3.60

We now make two assumptions. The first is that the internal network, within the countries, is more meshed than the external (the interconnections). This is represented by Eq.3.61, which states that the equivalent impedance per unit length inside the country (𝑍′𝑖𝑛𝑡) is lower than the equivalent



IRENE-40

Page: 135 of 383




impedance per unit length of the lines going outside the country (𝑍′𝑒𝑥𝑡), by a degree 𝛾. The second assumption is that the interconnecting line, which we add, has the same impedance per unit length along the whole line (either in the internal part or the external part). This is reasonable to assume, since it is the same line.

𝑍′𝑖𝑛𝑡 = 𝛾 ∙ 𝑍′𝑒𝑥𝑡, 0 < 𝛾 ≤ 1 Eq.3.61

𝑍′𝑛𝑒𝑤,𝑡𝑜𝑡 = 𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡 = 𝑍′𝑛𝑒𝑤,𝑒𝑥𝑡 Eq.3.62

Taking into account Eq.3.61 and Eq.3.62, Eq.3.60 turns into Eq.3.63:

𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡𝑐𝑜𝑚𝑝 = 𝛾 ∙ 𝑍′𝑛𝑒𝑤,𝑖𝑛𝑡 Eq.3.63

As a result, the necessary impedance for series compensation, as well as the degree of compensation can be computed through Eq.3.64 and Eq.3.65.

𝑋𝑐𝑜𝑚𝑝 = �𝑋′𝑛𝑒𝑤,𝑖𝑛𝑡 − 𝑋′𝑛𝑒𝑤,𝑖𝑛𝑡𝑐𝑜𝑚𝑝 � ∙ 𝑙𝑖𝑛𝑡 = (1 − 𝛾) ∙ 𝑋′𝑛𝑒𝑤,𝑖𝑛𝑡 ∙ 𝑙𝑖𝑛𝑡 Eq.3.64

𝐷𝑐𝑜𝑚𝑝 = 𝑋𝑐𝑜𝑚𝑝𝑋𝑡𝑜𝑡

= (1 − 𝛾) ∙ 𝑋′𝑛𝑒𝑤,𝑖𝑛𝑡 ∙ 𝑙𝑖𝑛𝑡𝑋′𝑛𝑒𝑤,𝑡𝑜𝑡 ∙ 𝑙𝑡𝑜𝑡

= (1 − 𝛾) 𝑙𝑖𝑛𝑡𝑙𝑡𝑜𝑡= (1 − 𝛾) �1 − 𝑙𝑒𝑥𝑡

𝑙𝑡𝑜𝑡� Eq.3.65

As 𝑙𝑖𝑛𝑡 ≤ 𝑙𝑡𝑜𝑡, this means that:

𝐷𝑐𝑜𝑚𝑝 ≤ 1 − 𝛾. Eq.3.66

If we assume that for every interconnecting line there are two parallel paths of similar voltage inside the countries, then 𝛾 = 0.5. This means that the necessary compensation for our studies will not exceed 50% in any case (upper bound). Here, we can observe two effects. The more meshed the internal network is with respect to the external, the higher this upper bound gets, i.e. more compensation would be necessary. The longer the line is with respect to the external interconnection length, the higher the degree of compensation that is necessary (but still remains below the upper bound defined by Eq.3.66). In Table 3.45, the degree of compensation and the necessary series compensation in Mvar for the long AC 400 kV and AC 750 kV lines is presented, taking into account two assumptions: a) there are two parallel paths of similar voltage inside the countries for each interconnection (γ=0.5), and b) all the new lines are 500 km long. The degree of compensation Dcomp is the same for both expansion options, as it not dependent on the reactance 𝑋𝑐𝑜𝑚𝑝. In Table 3.45, 𝑆𝑐𝑜𝑚𝑝 is computed



IRENE-40

Page: 136 of 383




through the relationship 𝑆𝑐𝑜𝑚𝑝 = 𝐼𝑚𝑎𝑥2 ∙ 𝑋𝑐𝑜𝑚𝑝, where 𝐼𝑚𝑎𝑥400𝑘𝑉 = 1.5 𝑝.𝑢. and 𝐼𝑚𝑎𝑥750𝑘𝑉 = 3.9 𝑝. 𝑢. (nominal voltage; baseMVA=1000 MVA; 1500 MVA is the capacity for the AC-400 KV single circuit line and 3900 MVA is the capacity of the 750 kV line). For the AC-400 kV case, 𝑆𝑐𝑜𝑚𝑝 is doubled, in order to account for the double-circuit line.

Table 3.45 – Series compensation for the AC-400 kV and AC-750 kV expansion scenarios, expressed in degree of compensation and amount of reactive power. (Assumptions: lines are 500 km long; for each interconnection there are two parallel paths of similar voltage inside the countries)

AC 400 kV AC 750 kV Line 𝐃𝐜𝐨𝐦𝐩 (%) 𝐒𝐜𝐨𝐦𝐩 (Mvar) 𝐒𝐜𝐨𝐦𝐩 (Mvar)

AT-DE 42.5% 1494 1666 FR-ES 41.0% 1441 1608 IT-SI 44.0% 1547 1725

RO-RS 49.0% 1723 1921 LT-PL 45.0% 1582 1764 FR-CH 46.0% 1617 1804 DE-CH 48.0% 1688 1882 AT-HU 38.0% 1336 1490 DE-PL 48.5% 1705 1902

GR-MK 43.0% 1512 1686 PT-ES 46.0% 1617 1804 AT-SI 44.0% 1547 1725 BE-FR 44.0% 1547 1725 DE-NL 47.0% 1652 1843

Based on the mentioned assumptions, it can be seen from Table 3.45, that the compensation ratio does not exceed 50%. However, if the intra-country network is more meshed, then the necessary compensation ratio would also increase. Here it should be noted that for compensation ratios above 50%, additional studies need to be carried out (e.g., about sub-synchronous resonance phenomena), in order to ensure that the dynamic security of the system is maintained. Based on the results from Table 3.45, in Table 3.46 we present our estimations for the total amount of series compensation that was assumed as installed during our simulation studies for the expansion scenarios AC-400 kV and AC-750 kV. The amount of compensation depends on the number of AC lines that are added in each generation scenario.

Table 3.46 – Total amount of series compensation that is necessary for each generation scenario



IRENE-40

Page: 137 of 383






BAU 2050 13.9 15.5 CCS 2050 9.4 10.5 DES 2050 16.0 17.8 EFF 2050 14.2 15.8 RES 2050 15.8 17.6

3.2.7.4 Total Generation Costs before and after the Expansion Scenarios

Figure 3.64 presents the total generation dispatch costs that result from each generation and expansion scenario. Figure 3.65 complements Figure 3.64 by presenting in percentage the cost reduction that can be achieved with each expansion option in comparison to the base scenario. As a general remark for the generation scenarios, we can observe that the CCS scenario results in the highest generation costs for 2050, while the EFF scenario results in the least costs. The reasons for that lie on the generation scenario design. The EFF scenario projects higher efficiency rates in the electricity consumption, and as a result the annual electricity demand is not as high as in the other scenarios. This, in turn, implies less electricity production and less costs. The CCS projects the construction of several CCS power plants. As the costs per MWh are higher for such plants, the annual generation costs increase and are actually more than double the costs of the DES, EFF and RES scenarios. The RES and DES scenarios demonstrate the least generation costs after the EFF scenario, due to the high renewable energy sources penetration, which have low operating costs. Here, it is interesting to note that both the RES and DES scenarios project an 80% RES participation in the European electricity mix by 2050. The difference is that while RES assumes a distribution of renewables across Europe, DES imports a large part of this energy from North Africa, i.e., there are a lot of CSP plants and significant amounts of energy are transported from South to North. Based on the OPF results, it seems that the DES scenario would result in lower generation costs than the RES scenario. In other words, it seems that large power plants of renewable energy sources and long transmission lines might be more cost-effective than distributed RES power plants across Europe. From a security of supply viewpoint though, a significant dependence of European electricity demand on foreign imports might not be considered optimal.



IRENE-40

Page: 138 of 383




Figure 3.66 – AC-OPF: Annual Total Generation Costs for all scenarios

Concerning the transmission expansion scenarios, as it would have been expected the base scenario results in the higher generation costs for all future possible generation mixes. Having a lower transmission capacity than the expansion scenarios, the base scenario does not allow as high amounts of power exchanges as the other scenarios. As a result, each country should rely on its own resources and “refuse” cheap power, which could be imported from neighbouring countries. Figure 3.65 demonstrates in a clearer way, the cost reductions that each of the expansion scenarios achieve. As it can be observed, the AC-750 kV has a better performance than the AC-400 kV expansion. This was anticipated, as the AC 750 kV lines have a higher transmission capacity (3900 MVA instead of 3000 MVA) and lower impedance. The best performance, however, is achieved if all the new lines were HVDC lines. The cost savings in the HVDC expansion scenario can exceed 20% of the base scenario costs (in DES 2050 generation). It is also interesting to note that the HVDC achieve such performance, although we have assumed significant compensation mechanisms for both AC scenarios. The reason is probably the controllability in the power flow routing that the HVDC lines can offer.

0

50

100

150

200

250


Billi

ons o

f Eur

os

Annual Total Generation Costs (OPF)

Base Scenario

AC 400 kV

AC 750 kV

HVDC



IRENE-40

Page: 139 of 383




Figure 3.67 – AC-OPF: Annual Total Generation Costs Reduction (in %) due to the expansion

scenarios

An additional remark that can be made by observing Figure 3.65 relates to the overall effect of the expansion scenarios on the generation costs. Although cost reductions are observed on all generation scenarios, it seems that the scenarios DES, EFF and RES profit the most from the expansions. This has to do with the fact that the existing European interconnected system was not originally designed for increased power exchanges. As a result the DES and RES scenarios, that require significant power exchanges due to the fluctuating nature of the power generation, profit the most from the transmission expansion. Such expansions allow, in a way, a redesign of the network in order to suit better the needs of a more “fluctuating” generation mix. In comparison, scenarios such as BAU and CCS, which rely on the current generation “model”, where each country has several conventional power plants and does not rely extensively on power exchanges, do not profit as much from the expansion.

3.2.7.5 Valuation Study based on the OPF Results

In this section we carry out a valuation study based on the savings that the European system could have from the reduced total generation costs. Except of the three expansion options we studied during the OPF calculations, i.e. AC-400 kV, AC-750 kV and HVDC, in this section we distinguish between HVDC Overhead lines and HVDC underground cables. For both options we assume that

0%

5%

10%

15%

20%

25%


Cost Reduction of Expansion Scenarios

AC 400 kV

AC 750 kV

HVDC



IRENE-40

Page: 140 of 383




they have the savings as calculated in the HVDC scenario. For the AC expansion scenarios, no cable alternative was studied, due to length limitations. It should also be noted that in all four scenarios, all the submarine interconnections are assumed to be HVDC. This implies, that also the AC expansion scenarios include several HVDC cables. As already explained in Section 3.2.6, we assumed a length of 500km for each line/cable. Except for the three submarine HVDC cables, NL-UK, DE-NO, and LT-SE, which are already under construction with known capacity and costs, for all other lines we assumed costs taken from the IRENE-40 database. For the AC-400 kV OHL, we assumed costs equal to 1.906M€ /km x 500km, while for the AC-750 kV, the estimated costs were equal to 1.220M€ /km x 500km. For the HVDC OHL, two converters with costs 240M€ each were assumed, as well as two parallel Overhead Lines with cost 1 M€/km x 500 km x 2 lines. For the cable alternative, each converter had an estimated cost of 544.5 M€, while three parallel cables were necessary with cost 1.12 M€/km x 500 km x 3 cables. Table 3.47 through Table 3.51 present the valuation results. As it can be easily observed, the HVDC Overhead Line option achieves the best performance. Nevertheless, all possible options are profitable, none with a payback period longer than 5 years; also interesting to note that especially in the DES generation scenario, the expansion leads to such cost savings that the investment is paid back within 1-2 years.

Table 3.47 – Valuation Results for BAU 2050

BAU 2050 AC-400 kV AC-750 kV HVDC OHL HVDC Cables

IRR 7.30% 7.57% 7.66% 7.20% ROI 137.92% 163.49% 171.68% 129.17% Payback period 5 yrs 4 yrs 3 yrs 5 yrs

Table 3.48 – Valuation Results for CCS 2050

CCS 2050 AC-400 kV AC-750 kV HVDC OHL HVDC Cables

IRR 7.42% 7.63% 7.82% 7.44% ROI 148.48% 168.94% 188.89% 151% Payback period 4 yrs 3 yrs 3 yrs 4 yrs



IRENE-40

Page: 141 of 383




Table 3.49 – Valuation Results for DES 2050

DES 2050 AC-400 kV AC-750 kV HVDC OHL HVDC Cables

IRR 8.39% 8.80% 9.02% 8.37% ROI 256.84% 315.01% 349.43% 253.30% Payback period 2 yrs 1 yr 1 yr 2 yrs

Table 3.50 – Valuation Results for EFF 2050

EFF 2050 AC-400 kV AC-750 kV HVDC OHL HVDC Cables

IRR 8.19% 8.51% 8.51% 7.97% ROI 231.17% 273.25% 272.94% 205.36% Payback period 2 yrs 2 yrs 2 yrs 3 yrs

Table 3.51 – Valuation Results for RES 2050

RES 2050 AC-400 kV AC-750 kV HVDC OHL HVDC Cables

IRR 8.13% 8.59% 8.71% 8.08% ROI 223.74% 283.57% 301.27% 218.39% Payback period 2 yrs 2 yrs 2 yrs 2 yrs §4.2 sumarizes the overall security-constrained OPF and cost of security conclusions drawn from above §3.2.



IRENE-40

Page: 142 of 383




3.3 INFRASTRUCTURE SOLUTIONS FOR INCREASED COMPETITIVENESS

3.3.1 Introduction

This paragraph constitutes the report of Task 3.4 “Infrastructure solutions for increased competitiveness” belonging to Work Package 3 “Solutions” of the IRENE-40 project. The document was written by the IRENE-40 partners involved in Task 3.4 reflecting the contribution of their work developed within the task. In this respect, work contributions were made by Imperial College London (ICL), Eidgenössische Technische Hochschule Zurich (ETH Zurich), Siemens AG Power Transmission and Distribution (Siemens) and Energy research Centre of the Netherlands (ECN). Europe’s path towards a low-carbon energy system of the future will inevitably requires a major shift in the structure of electricity generation technologies, primarily targeting renewable energy sources and large-scale low-emission or zero-emission technologies such as nuclear power or carbon capture and storage. The variable and unpredictable nature of the output of renewable energy sources, such as wind and solar power, constitutes a major challenge to the development and operation of the future electric power systems with a large share of renewable energy sources in an economically efficient manner while maintaining security of supply at adequate levels. It is understood that the variable nature of the renewable energy supply source will require significant backup capacity in the form of generation and/or transmission infrastructure in order to integrate the renewable energy sources output fluctuations in a secure and cost-efficient manner. On the other hand, demand side technologies, such as demand response, will also play an important role to the development and operation of the future electric power systems contributing to improve security of supply, utilisation of generation and transmission infrastructure and efficiency of system balancing and ancillary services. In this sense, Task 3.4 report of the IRENE-40 project aims to identify the influence of infrastructural development strategies on competitiveness within the electrical energy system. Specifically, it quantifies and assesses the technical impacts and the associated costs related with different investment alternatives on the future European transmission network infrastructure. Finally, Task 3.4 report presents a technical assessment suggesting present and feasible future technology options available to respond to the identified levels of investment in transmission network infrastructure.



IRENE-40

Page: 143 of 383




3.3.1.1 Questions, aims and objectives

In accordance with IRENE-40 description of work13

, “Task 3.4, infrastructure solutions for increased competitiveness, will identify the influence of infrastructural development strategies on competitiveness within the electrical energy system”.

Broadly, the main objective of Task 3.4 is to quantify and assess the technical impacts and the associated costs related with different investment alternatives on the future European transmission network infrastructure. This should be achieved in three key ways: (i) through the quantification of the secure and cost-optimal levels of transmission network infrastructure required in the European energy system to accommodate a large share of renewable energy resources; (ii) through the quantification of the competitiveness of alternative technologies (demand side flexibility and storage facilities) as options to build transmission infrastructure; and (iii) through the assessment of various technology options as response to the needs for investment in transmission infrastructure. Specifically, Task 3.4 can be divided into three main questions and areas of work each with individual objectives:

− Q1. What is the required level of investment in generation and transmission infrastructure to integrate renewable energy sources in a secure and cost-optimal manner in the future European transmission network?

This question requires the application of a set of comprehensive models to evaluate the technical, security and cost performance of the future electricity system in Europe under the set of generation and demand scenarios developed within IRENE-40 project. Specifically, the response to this question includes the quantification of additional infrastructure developments (generation and transmission infrastructure) required to accommodate various levels of renewable energy sources in a cost-optimal manner, while maintain adequate levels of security of supply. This question is addressed by ICL in Chapter 3.3.3.

− Q2. What is the contribution of competitive technologies, such as demand side flexibility and storage facilities, in the level of investment in generation and transmission infrastructure while maintaining adequate levels of security of supply?

This question details the application of the aforementioned methodology to consider the contribution of competitive technologies, such as demand side flexibility and storage facilities to smooth out the demand profile in order to minimise the overall generation and transmission investment cost and generation operating cost. These technologies act as an additional source of

13 IRENE-40 project, “Infrastructure Roadmap for Energy Networks in Europe”, “Annex I – Description of Work”. Contract number: TREN/FP7EN/218903/IRENE-40.



IRENE-40

Page: 144 of 383




flexibility in the system by reducing the need for flexible thermal (Gas or Coal) generation capacity and transmission network capacity, therefore reducing the overall system costs. This question is addressed by ICL in Chapter 3.3.3.

− Q3. What are the present and feasible future technology options available to respond to the required levels of investment in the future European transmission network infrastructure?

This question builds upon the key messages and conclusions delivered from the generic analysis of the different transmission infrastructure investment alternatives as well as from the optimal cost-efficient and secure levels of transmission investment proposed for the future European transmission network infrastructure. In this sense, this question details the application a technical assessment suggesting various feasible transmission network technology solutions for the future European transmission network. This question is addressed by Siemens in Chapter 3.3.4. The response to this question results from the combination of the analysis undertaken by ETH Zurich and ICL. ETH Zurich evaluated the impact of different transmission infrastructure investment alternatives on economic indicators, such as the overall social welfare, the distribution of market power, congestion cost, and consumer/producer surplus. This analysis requires the development of a methodology to explicitly compare investment strategies targeting the removal of congestion with the objective of increasing market efficiency by means of limiting market power/strategic behaviour of the market participants. The technologies selected, concerning the different investment alternatives, range from the building or reinforcement of “conventional” AC transmission lines or the installation of FACTS devices, in particular Thyristor-Controlled-Series-Compensators (TCSCs). The quantitative assessment was applied to a particular case study representative of the Italy, France and Switzerland main interconnected transmission system. This particular case study is used as a test case for a region of interest, i.e. it represents a subset of the European main interconnected transmission systems which is used within the IRENE-40 project. ICL performed a quantitative assessment to evaluate the generation and transmission infrastructure requirements in a cost optimal manner, facilitating competition in the energy market and taking into account the complex trade-offs among various system cost components, while maintaining security of supply at acceptable levels. The quantitative assessment was applied to the generation and demand scenarios, developed within IRENE-40 project, to quantify the additional infrastructure developments (generation capacity and transmission interconnections) required in the future European transmission network to accommodate various levels of renewable energy sources in a secure and cost-efficient manner. The outcomes from the aforementioned analyses are processed in Chapter 3.3.4 through a technical assessment which translates them into feasible transmission network technologies options for the



IRENE-40

Page: 145 of 383




future European transmission network. The present and feasible future technology options are selected from the technology database14

of IRENE-40 project.

3.3.1.2 Scope

The title of Task 3.4 of IRENE-40 project is “Infrastructure solutions for increased competitiveness”. The work and analysis that this requires is bounded in the following way: Europe’s path towards a low-carbon energy system of the future will inevitably requires a major shift in the structure of electricity generation and demand technologies. On the generation side, the increasing presence of renewable energy sources and large-scale low-emission or zero-emission technologies will constitute a major challenge to the development and operation of the future electric power systems. Due to the variable nature of their output and their technical operating characteristics respectively, the electric power system should evolve to allow the integration of such technologies in an economically efficient manner while maintaining adequate levels of security of supply. On the demand side, responsive demand will also play an important role on the development and operation of the future electric power systems. Demand response will potentially benefit the system contributing to improve the security of supply of the system, the utilisation of generation and transmission infrastructure and the efficiency of system balancing and ancillary services. Under this perspective, IRENE-40 project has developed five scenarios for electricity demand and generation at European country level over a 40 year period from 2010 to 2050. These, in turn, were developed around key determinate drivers for the future transmission network infrastructure in Europe. Amongst the selected key drivers for network development, it is emphasised the evolution of electricity demand, the composition of the generation mix, especially the share of renewable energy sources, the role of the import of renewable energy sources and the impact of demand side response. Based on the electricity demand and generation scenarios, quantitative and qualitative assessments are performed to identify the required level of investment in generation and transmission infrastructure to integrate renewable energy sources in a secure and cost-optimal manner in the future European transmission network. As alternative options to the investment in generation and transmission infrastructure Task 3.4 report explores the contribution of competitive technologies, such as demand side flexibility and storage facilities while maintaining the required system reliability performance at acceptable levels.

14 IRENE-40, Deliverable 2.2, “Technology database and technological development forecast methodology”.



IRENE-40

Page: 146 of 383




Finally, Task 3.4 report performs a technical assessment suggesting present and feasible future transmission network technology options to respond to the required secure and cost-optimal levels of investment in the future European transmission network infrastructure previously quantified.

3.3.1.3 Organisation of the report

Broadly, each sub-paragraph is focussed on responding to one of the main questions outlined earlier. When combined, these paragraphs serve to meet the Task 3.4 aim of identifying the influence of infrastructural development strategies on competitiveness within the electrical energy system. This will be achieved through: (i) quantification and assessment of the technical impacts and the associated costs related with different investment alternatives on the future European transmission network infrastructure; and (ii) technical assessment suggesting present and feasible future technology options available to respond to the identified levels of investment in transmission network infrastructure.

3.3.2 Generic analysis of different transmission infrastructure investment alternatives

In W3IM TN4005C the application of a developed generic analysis of different transmission infrastructure investment alternatives methodology to a case study representative of the Italy, France and Switzerland main interconnected transmission system is demonstrated, i.e. an agent-based model approach to build a market based on Locational Marginal Pricing (LMP). The key conclusions are: Modelling the market participants as adaptive agents in oligopolistic structures enables to consider the possibility of strategic behaviour and the existence/exercise of market power. The market clearing problem of the Independent System Operator (ISO) was formulated to allow studying network investments by “conventionally” reinforcing lines or by installing FACTS devices, in particular Thyristor-Controlled-Series-Compensators (TCSCs).

− Transmission investments – independent of the specific technology – counter the effect introduced by the exercise of market power. In particular it was shown that it is not necessary to develop “isolated” investment strategies. Congestion removal and mitigation of market power are synonymous being influenced in an almost identical way from investment measures. This result may prove relevant for policy makers and investors as it reduces the complexity of prospective decisions making it possible to narrow the investment focus.

− Concerning the use of TCSCs it is possible that the optimal investment location does not coincide with the congested line. This emphasises the fact that transmission investment is a European issue rather than a national one.



IRENE-40

Page: 147 of 383




3.3.3 Future transmission infrastructure investment in Europe

3.3.3.1 Introduction

Europe’s path towards a low-carbon energy system of the future will inevitably require a major shift in the structure of electricity generation technologies, primarily targeting intermittent renewable resources and large-scale low-emission or zero-emission technologies such as nuclear power or Carbon Capture and Storage. Because of variable and unpredictable output of renewable technologies such as wind power, it may become challenging to plan and operate the future electricity systems with high shares of intermittent renewable resources in an economically efficient manner. It is understood that the variability of intermittent renewable supply will require significant backup capacity in the form of generation or transmission network assets able to absorb the renewable output fluctuations; these however come at a cost. In this respect, it is important to quantify adequate levels of backup generation capacity to support security of supply and the cost-effective integration of renewable energy sources. It is equally important and critical to quantify the adequate interconnectors between European countries to allow access to low marginal cost energy sources, including renewable sources, installed in the different European countries. In addition to allow inter-regional energy arbitrage, there are a number of benefits of having interconnections between regional grids. The benefits include:

− Ability to share capacity and reserves between regions;

− Avail benefits of diversity in demand and (renewable) generation; and

− Enable access for control and flexibility from neighbouring systems. Increased level of interconnections among various regions in Europe will lead to more efficient and increased utilisation of the infrastructure and an improved reliability of the overall system. This section documents the methodology and its respective application to quantitative assessments, elaborated based upon the scenarios proposed (Chapter Error! Reference source not found.). Broadly, the objective is to evaluate the cost-optimal power system requirements in the European energy system with a large share of renewable energy resources. The presented methodology focus on quantifying the technical, security and cost performance of interconnected multi-region electricity systems in a cost optimal manner to allow a cost-effective integration of renewable energy sources into the future interconnected pan-European system.



IRENE-40

Page: 148 of 383




Based upon the proposed generation and demand scenarios, the presented quantitative assessment evaluates:

− The additional generation capacity and transmission interconnections requirements to accommodate various levels of renewable energy sources in a cost optimal manner; and

− The contribution of competitive technologies, such as demand side flexibility and storage facilities, in reducing the need for additional generating capacity and inter-regional transmission while maintaining required system reliability performance.

The application of the methodology also identifies the key drivers affecting the secure and cost effective integration of large scale renewable energy sources in into the future interconnected pan-European system. Figure 66 shows the relationship between competing technologies to achieve the objectives aforementioned.

Figure 68: Technology options for efficient integration of low carbon generation systems

Figure 67 presents a generic overview of the modelling approach.

Flexible Generation

Demand Response

Storage

Network

Increasing asset utilisation and efficiency of operation



IRENE-40

Page: 149 of 383




Figure 69: Overview of the modelling approach

3.3.3.2 Methodology

A set of comprehensive models have been applied to evaluate the technical, security and cost performance of the future electricity system in Europe under the proposed set of scenarios (Chapter Error! Reference source not found.). The developed models include:

− Generation and transmission investment model;

− Reliability assessment model; and

− System balancing model. The models work in tandem such that the overall modelling framework seeks to minimise the total system costs comprising: (i) additional generating capacity; (ii) additional inter-regional transmission network capacity; and (iii) annual electricity production cost. A detailed description can be found in [W3IM TN4005C], whereas here we focus on the found quantitative outcomes. Figure 68 shows the schematic representation of the generation and transmission investment model.

Generation and Transmission Investment Model:Minimisation of the overall generation and transmission

investment cost and generation operating cost

Reliability Assessment Model:

Generation capacity adequacy assessment

System Balancing Model:

Minimisation of generation operating cost

Data Base

Adequate level of generation capacity;Overall cost of investment and operation;

Generation and transmission capacity factors (utilisation);Expected energy penetration level of renewable.

Sensitivity Analysis



IRENE-40

Page: 150 of 383




Figure 70: Schematic representation of the generation and transmission investment model

3.3.3.2.1.1 Key inputs:

The main inputs to the “Generation and Transmission Investment Model” are:

− Annualised transmission investment for each transmission corridor and network configuration;

− Generation data including installed capacity, technical and dynamic characteristics, generation operating cost information, and investment cost of additional generating capacity in each region;

− Hourly time series of energy profiles of renewables including wind, solar (PV, CSP) for one year time horizon for each region;

− Seasonal available hydro (run-of-river and reservoir type) energy for each region;

− Hourly time series of electricity demand profiles for one year time horizon for each region; and

− Storage and demand side flexibility (DSF) parameters.

3.3.3.2.1.2 Key outputs: The main outputs from the “Generation and Transmission Investment Model” are:

− Optimal additional generation capacity for each region and optimal transmission capacity for each transmission boundary. Associated infrastructure (generation and transmission) investment costs;

− Optimal scheduling of resources including thermal generators, renewable energy sources, storages and demand side flexibility. Associated system operating costs; and

Cost Characteristics:transmission investment

Generation Data:technical and cost

characteristics

Renewable Generation Data:

wind, solar, hydro, etc.

Storage and Demand Response Data Load Data Network Configuration

Generation and Transmission Investment Model:Optimisation Model

Objective function: Minimisation of the overall generation and transmission investment cost and generation operating cost

Investment decisions: Transmission and generation capacities;

Operating decisions (further assessed in system balancing model):Overall system operation cost;

Generation dispatch including RES and storagePower flows.



IRENE-40

Page: 151 of 383




− Optimal power flows at each interconnector.

3.3.3.2.1.3 Key assumptions: The key assumptions used in the “Generation and Transmission Investment Model” are:

− Reliability criterion: LOLE ≤ 4 hours/years in each region;

− Generating units are assumed to operate independently. For example, it is assumed that an outage of one generator will not directly affect the operation of the other units in the system;

− The thermal generating units are assumed to be either fully available or out of service in accordance with their long-term plant availability statistics;

− Hydro generators are considered to be fully reliable, i.e. these are modelled as units which are available all the time and constrained only by their rated capacities and available hydro energy and reservoir constraints;

− The wind and solar plants are also assumed to be fully available while their output would be limited by the installed capacities and available energy during each hour;

− In order to meet the reliability criterion, additional generation plant, e.g. open cycle gas turbine (OCGT) plant, can be added in each region;

− Optimal secured transmission capacity as an outcome of the model implies that the network design and capacity has conformed to the reliability criterion;

− Demand side flexibility is represented by the amount of power and energy that can be shifted within one day, i.e. 24 hour period;

− Storage units are modelled as units which are available all the time and constrained only by their rated capacities, stored energy and reservoir constraints; and

− Simulations are carried out for hourly periods across one year time horizon.



IRENE-40

Page: 152 of 383




Figure 71: Schematic representation of the system balancing model

3.3.3.2.1.4 Key outputs: The tangible outputs from the “System Balancing Model” include:

− Additional reserve requirements (response, spinning and standing reserves) due to intermittent generation;

− Additional system flexibility requirements to carry out real time system balancing in the form of flexible generation capacity e.g. OCGT or storage;

− Magnitude of renewable energy curtailment (wind, solar, hydro, geothermal etc.), if required, to maintain system security;

− Generation capacity factors (utilisation) and (expected) energy penetration level of renewables;

− Quantification of the role and benefits of demand response and storage in enhancing system balancing capabilities;

− CO2 emission performance;

− Overall system operation costs; and

− Identification of bottleneck in interconnections (flow duration curves).

3.3.3.2.2 Main interconnected transmission system

Resposnse and Reserve Requirements

Generation Data:technical and cost

characteristics

Renewable Generation Data:

wind, solar, hydro , etc.

Storage and Demand Response Data Load Data Network Configuration

Dispatch Optimisation Model:

Objective function: Minimisation of the overall generation operating cost

Operational feasibility No

Dispatch of all generation technologies;Generation capacity factors (utilisation);

Transmission bottlenecks (power flow duration curve);Changes in demand profiles and saving due to demand side flexibility;

Magnitude of energy curtailment (wind, hydro, etc.) and associated costs;CO2 emission performance;

Additional cost of reserve due to renewables;Overall system operation cost.

Yes

Feasibility Model:

Additional ‘flexible’ capacity



IRENE-40

Page: 153 of 383




IRENE-40, Work package 2, Task 2.315

, had the aim of providing a set of reference scenarios on which the forthcoming European transmission infrastructure developments would be quantified and assessed.

Figure 72: Europe – main interconnected transmission system

In the developed network topology, the present interconnectors are maintained and future interconnectors are placed between countries in accordance to the “Ten-Year Development Plan”16

provided by the European Network of Transmission System Operators for Electricity (ENTSO-E).

3.3.3.3 Optimal investment in transmission infrastructure

This section details the application of the system analysis methodology to the developed generation and demand scenarios: It evaluates the additional generation capacity and transmission interconnectors requirements in a cost optimal manner, facilitating competition in energy market taking into account the complex trade-offs among various system cost components, while maintaining system’s security of supply at acceptable levels.

15 IRENE-40, Task 2.3, Internal Report, “Scenario synthesis”. 16 European Network of Transmission System Operators for Electricity (ENTSO-E), “Ten-year network development plan 2010-2020”.

FI

PT

DEPL

SE

ES

IT

FR

UK

NO

NL

LU

CHAT

SI

CZ

RO

Bo&Cro

HU

BG

GR

YU

AL

SK

BE

EE

LV

LT

DK



IRENE-40

Page: 154 of 383




The system analysis methodology is applied to the European network topology, described in the previous section, under the “Business as Usual (BAU)”, “Renewable (RES)”, “Desertec (DES)”, “Carbon Capture and Storage (CCS)”, and “High Efficiency (EFF)” scenarios. The quantitative assessment is initially performed for each scenario for the 2010, 2030 and 2050 decades in order to provide a generic overview of the additional generation and transmission infrastructure requirements over the 40-year period of analysis until 2050.

3.3.3.3.1 Business as Usual scenario Business as Usual scenario is characterised by a diverse mix of the generation portfolio. It is mainly composed of conventional thermal power plants, nuclear power plants and renewable energy sources. Fossil fuel generation technologies remain the favoured to supply energy demand. Thus, BAU scenario is constituted by the less ambitious generation mix for emission reduction targets. The deployment of renewable energy sources meets the 2020 emission reduction targets; however the subsequent uptake is relatively slow. Low carbon generation technologies and renewable energy sources contribute moderately to the electricity of supply in 2050. Figure 71 and Figure 72 show respectively the generation and demand background for the year 2030 and year 2050 for each European country considered in the analysis.

Figure 73: BAU scenario, Year 2030 – generation and demand background

0

50

100

150

200

250

300

AL AT BE BG BO CH CZ DE DK EE ES FI FR GR HU IT LT LU LV NL NO PL PT RO SE SI SK UK YU

[GW

]

New capacity

Other RES

Storage

Hydro reservoir

Hydro ROR

Geothermal

Biomass

CSP

PV

Wind

Oil

Nuclear

Gas CCS

Coal CCS

Gas

Coal

Peak demand



IRENE-40

Page: 155 of 383




Figure 74: BAU scenario, Year 2050 – generation and demand background

It can be seen in Figure 71 and Figure 72 that some countries such as, Germany and United Kingdom, require to build new generation plants, besides the initial generation plants given in the scenario data, in order to maintain security of supply at acceptable levels (LOLE ≤ 4 hours/years). The eventual addition of generation thermal plant is in the form of open cycle gas turbine (OCGT). The “Generation and Transmission Investment Model” and the “Reliability Assessment Model” are then applied to minimise the total system costs comprising: (i) additional generating capacity; (ii) additional inter-regional transmission network capacity; and (iii) annual electricity production cost, while maintaining required system reliability performance. As a result of the application of the system analysis methodology, Table 52 presents a summary of the main characteristics of the system. Generating capacity represents the installed “firm” generating capacity in the system. A discussion on the term “firm” capacity is detailed below. The interconnectors’ transmission capacity [TW-km] represents the total level of transmission network infrastructure that needs to be present in the system, for the year 2030 and 2050, to ensure cost optimality and adequacy of security of supply. The additional interconnectors’ transmission capacity [TW-km] represents the level of transmission network infrastructure that needs to be added in the system, from the year 2010 to the year 2030 and from the year 2030 to the year 2050. The corresponding values are shown in Table 52 between parentheses. The level of transmission

0

50

100

150

200

250

300


[GW

]

New capacity

Other RES

Storage

Hydro reservoir

Hydro ROR

Geothermal

Biomass

CSP

PV

Wind

Oil

Nuclear

Gas CCS

Coal CCS

Gas

Coal

Peak demand



IRENE-40

Page: 156 of 383




network infrastructure used for the reference year of 2010 corresponds to the level specified by the ENTSO-E17

for the same year.

Table 52: BAU scenario – summary of the main characteristics of the system

Year Peak

demand [GW]

Electricity demand [TWh]

“Firm” generation capacity [GW]

Interconnectors transmission

capacity [TW-km]

Additional interconnectors

transmission capacity

[TW-km]

RES contribution

[%]

RES curtailment

[%]

2030 649 4,055 1,108 37 4 (2010–2030) 44 0.3

2050 746 4,674 1,268 48 11 (2030–2050) 47 0.7

BAU scenario for the year 2030 is characterised by a peak demand of 649GW and total energy demand of 4,055GWh, whereas for the year 2050 is characterised by a peak demand of 746GW and total energy demand of 4,674GWh. The minimum installed “firm” capacity, necessary to be present in the system to ensure that the system reliability (LOLE ≤ 4 hours/years) is maintained at acceptable level is 1,108GW and 1,268GW respectively. The aforementioned system characteristics have a direct impact on the transmission network infrastructure in order to deliver the required reliability of electricity of supply in each individual region in a cost optimal manner. Thus, the transmission network infrastructure reinforcements are higher in the year 2050. The term “firm” capacity refers here to generation technologies that are able to present a consistently flat power output profile across the year. It should be emphasised that no generation plant is capable of presenting a consistently flat power output profile across the year mainly due to outages for scheduled maintenance and random forced outages. Thus, the term “firm” refers to generation technologies such as, Coal, Coal CCS, Gas, Gas CCS, Nuclear, Oil, Biomass, Geothermal, Hydro and Storage. The “non-firm” generation technologies refer to those that present a considerable variable power output profile across the year such as, Wind, Solar Photovoltaic and Concentrated Solar Power. It should be stressed that the calculation of the minimum installed “firm” capacity, that is required to be present in the system to attain an acceptable level of security of supply, considers the random nature of forced outages of the “firm” generation technologies, the contribution of the “non-firm” generation technologies which present a limited capacity value and also the presence of

17 European Network of Transmission System Operators for Electricity (ENTSO-E), “Ten-year network development plan 2010-2020”.



IRENE-40

Page: 157 of 383




interconnectors allowing for energy arbitrage and for the share of capacity and reserves between countries. It can be seen in Table 52 that the contribution of renewable energy sources to supply energy demand is relatively similar in both years. This is in agreement with the BAU scenario characteristics which presents a relatively slow deployment of renewable energy sources after 2020. It can also be seen that the amount of renewable energy curtailed in the system is very small. This implies that nearly all the energy available from renewable energy source is used towards supplying energy demand. Table 53 presents a summary of costs components related with the scenarios under analysis. The annual operating cost [b€/Yr] represents the annual electricity production cost from the various generation technologies present in the system. The additional transmission capacity cost [b€] represents the investment cost of the transmission capacity level that needs to be added in the system from the year 2010 to the year 2030 and from the year 2030 to the 2050 respectively. The additional generation capacity cost [b€] represents the investment cost of the generation capacity level that needs to be added in the system from the year 2010 to the year 2030 and from the year 2030 to the 2050 respectively. A discussion on this cost component is detailed below. The additional generation capacity cost (security of supply) [b€] represents the investment cost of the generation capacity level that needs to be added in the system to ensure adequate levels of security of supply. Note that, despite highlighting this cost component in Table 53, it has also been embedded in the additional generation capacity cost figure provide in Table 53.

Table 53: BAU scenario – summary of the main costs components of the system

Year Annual operating

cost [b€/Yr]

Additional transmission capacity

cost [b€]

Additional generating capacity cost

[b€]


(security of supply) [b€]

2030 122 5 (2010–2030)

1,106 (2010–2030)

9 (2010–2030)

2050 190 14 (2030–2050)

470 (2030–2050)

2 (2030–2050)

It can be observed in Table 53 that year 2050 is characterised by a higher electricity demand level than the year 2030, and therefore greater level of installed generating capacity. It should be noted that the generation technologies fuel cost significantly increase from 2010 to 2050 according to the scenario data. Thus, it can be observed in Table 53, that the need of producing more energy from the same generation technologies at higher fuel costs leads to an increase of the annual operating cost from 2030 to 2050. The additional investment cost in transmission capacity also rises throughout the two decades 2010–2030 and 2030–2050 as more transmission infrastructure is required as previously justified. The cost of investment in additional generation infrastructure



IRENE-40

Page: 158 of 383




(security of supply) also increases since there is a need to build more peaking plants in order to maintain security of supply at acceptable levels and to still achieve a cost optimal solution. The additional generating capacity cost [m€] is derived from the “overnight” investment cost of capacity for each particular generation technology (Chapter Error! Reference source not found.). The overnight investment cost of capacity of a generator is typically expressed in [€/kW]. For example, the overnight investment cost of an open cycle gas turbine (OCGT) plant might be €350/kW, so a 1000MW plant would cost €350 million. In economic terms, this is the present-value cost of the plant that it would have to be paid as a lump sum up front to pay completely for its construction. Hence the overnight cost captures the fixed cost of generation. In this work, in order to evaluate the additional generating capacity cost [m€] throughout two decades, say 2010–2030, the installed generating capacity [MW] of each technology is first considered in the year 2030 and then in the year 2010. These values are now subtracted, [MW], and multiplied by the respective overnight investment cost [€/kW] provided in (Chapter Error! Reference source not found.). This approach has an inherent assumption that any increase in generating capacity is solely due to new build of generating plants. On the other hand any decrease in generating capacity is solely due to decommissioning of generating plants. Figure 73 presents the transmission network capacity of the present interconnectors, as given by the ENTSO-E for the year 2010, as well as the optimal levels of investment in transmission network infrastructure for the years 2030 and 2050.



IRENE-40

Page: 159 of 383




Figure 75: BAU scenario – Transmission network capacity requirements

It can be observed in Figure 73 that only scarce number interconnectors require investment in transmission network infrastructure compared to the interconnection levels attained in the year 2010. It can also be seen that higher levels of transmission infrastructure requirement takes place in the year 2050. This conclusion has to some extent been earlier reflected by the higher annualised cost of investment in transmission infrastructure achieved in the two years of analysis. The level of reinforcement required for the interconnectors will be subsequently discussed in more detail. Figure 74 and Figure 75 show the schematic representation of the main interconnected European transmission system with the respective level of transmission infrastructure reinforcement for the years 2030 and 2050 respectively. These represent the additional amounts of transmission network capacity required, compared to the present interconnectors, as given by the ENTSO-E for the year 2010.

0

2

4

6

8

10

12

14PT

-ES

ES-F

RFR

-UK

FR-B

EN

O-D

EFR

-DE

FR-C

HFR

-ITIT

-CH

IT-A

TIT

-SI

DE-D

KDE

-PL

DE-C

ZDE

-AT

DE-C

HPL

-CZ

PL-S

KCZ

-SK

CZ-A

THU

-SK

HU-R

OHU

-YU

HU-B

OHU

-AT

YU-R

OYU

-BG

YU-G

RYU

-AL

YU-B

ORO

-BG

BG-G

RGR

-AL

BO-S

ISI

-AT

CH-A

TU

K-N

OU

K-N

LU

K-BE

NL-

DEN

L-BE

NL-

NO

NL-

DKN

O-D

KN

O-S

ESE

-DE

SE-P

LSE

-FI

SE-D

KLU

-BE

LU-D

EFI

-EE

EE-L

VLT

-LV

LT-P

LLT

-SE

Tran

smiss

ion

capa

city

[GW

]

ENTSO-E 2010BAU 2030BAU 2050



IRENE-40

Page: 160 of 383




Capacity [GW] 0.1 2 5 10 20 30 > 30 Colour code

Figure 76: BAU scenario, Year 2030 – transmission network capacity reinforcements

FI

PT

DE

PL

SE

ES

IT

FR

UK

NO

NL

LU

CHAT

SI

0.0

CZ

RO

Bo&Cro

HU

BG

GR

YU

AL

SK

1.9 0.0

BE

EE

LV

LT

0.0

DK



IRENE-40

Page: 161 of 383





Figure 77: BAU scenario, Year 2050 – transmission network capacity reinforcements

It can be inferred from Figure 74 and Figure 75 that, under BAU scenario, most of the present transmission infrastructure is adequate in most of the cases and therefore there is no need for significant transmission reinforcements. As aforementioned, the presence of interconnectors brings various benefits to the electric power systems, such as energy arbitrage (i.e. to allow countries to access to low marginal cost energy sources installed in other countries) and security of supply (i.e. the ability to share capacity and reserves between interconnected countries). It should be stresses that BAU scenario is constituted by the less ambitious generation mix for emission reduction targets. Low carbon generation technologies and renewable energy sources only contribute moderately to the electricity of supply towards 2050. Thus, significant levels of investment in transmission infrastructure are not required from an energy arbitrage perspective in order to allow countries to access to low marginal cost energy sources installed in other countries. This is because there is only a modest contribution of renewable energy sources to supply energy demand. In terms of security of supply, the fact that all countries have a firm capacity margin above

FI

PT

DE

PL

SE

ES

IT

FR

UK

NO

NL

LU

CHAT

SI

0.0

CZ

RO

Bo&Cro

HU

BG

GR

YU

AL

SK

1.9 0.0

BE

EE

LV

LT

3.6

DK



IRENE-40

Page: 162 of 383




20% and also that there is only a modest contribution from renewable energy sources suggests that such countries are able to maintain their own security of supply within acceptable levels without the need for significant increase of transmission infrastructure reinforcement. Figure 76 presents the firm capacity margin for the year 2050 for every European country considered in the analysis.

Figure 78: BAU scenario, Year 2050 – firm capacity margin

Capacity margin is defined here as the as the percentage difference between the total system generation capacity and the system peak demand with respect to the former. In this particular case, firm plant margin, uses the aforementioned firm generation technologies to obtain the total system generation capacity. Generally, a total system generation capacity requirement is in the neighbourhood of 120% of the system peak demand, in order to attain adequate levels of security of supply18 Figure 76. It is observed from that all countries have a firm capacity margin above 20% suggesting that such countries are able to maintain their own security of supply within acceptable levels without the need for significant increase of transmission infrastructure reinforcement.

3.3.3.3.2 Renewable scenario Renewable scenario is characterised by an ambitious generation technology mix for achieving emission reduction targets and strong policy support is therefore assumed to drive the deployment of renewable energy sources especially after the year 2020. Energy renewable sources replace 18 Resource and Transmission Adequacy Recommendations, North American Electric Reliability Council (NERC), 2004.

264%

92%

64%

119% 123%

48%35%

49%

83%

337%

40%

134%

41% 40%54%

68%

340%

84%

155%

79%

49%

23%

91%

227%

111%

176%

71%

45%

22%

0%

50%

100%

150%

200%

250%

300%

350%

400%


Capa

city

pla

nt m

argi

n [%

]



IRENE-40

Page: 163 of 383




conventional power plants (coal, gas and oil) even before technical end-of-life. Thus, the use of fossil fuels generation technologies decreases from 45% in 2020 to 8% in 2050. The installed capacity of nuclear power plants is also significantly reduced in 2050. The geographical distribution of the renewable energy resources includes large clustered offshore and onshore wind farms in the northwest, solar and wind in the south, hydropower and biomass in central and northern Europe. Figure 77 and Figure 78 show respectively the generation and demand background for the year 2030 and year 2050 for each European country considered in the analysis.

Figure 79: RES scenario, Year 2030 – generation and demand background

0

50

100

150

200

250

300

350

400

450

500


[GW

]

New capacity

Other RES

Storage

Hydro reservoir

Hydro ROR

Geothermal

Biomass

CSP

PV

Wind

Oil

Nuclear

Gas CCS

Coal CCS

Gas

Coal

Peak demand



IRENE-40

Page: 164 of 383




Figure 80: RES scenario, Year 2050 – generation and demand background

It is observed in Figure 77 and Figure 78 the presence of significant levels of installed capacity from intermittent energy sources such as Wind and Solar Photovoltaic, especially in the countries with the most developed power systems such as, Germany, Spain, France, Italy and United Kingdom. As a result of the application of the system analysis methodology, Table 54 presents a summary of the main characteristics of the system.

Table 54: RES scenario – summary of the main characteristics of the system

Year Peak

demand [GW]




capacity [TW-km]



[TW-km]

RES contribution

[%]

RES curtailment

[%]

2030 649 4,216 1,018 35 3 (2010–2030) 45 0.2

2050 875 5,255 1,211 163 128 (2030–2050) 75 3

RES scenario for the year 2030 is characterised by a peak demand of 649GW and total energy demand of 4,216GWh, whereas for the year 2050 is characterised by a peak demand of 875GW and total energy demand of 5,255GWh. The minimum installed “firm” capacity, necessary to be present in the system to ensure that the system reliability (LOLE ≤ 4 hours/years) is maintained at

0

50

100

150

200

250

300

350

400

450

500


[GW

]

New capacity

Other RES

Storage

Hydro reservoir

Hydro ROR

Geothermal

Biomass

CSP

PV

Wind

Oil

Nuclear

Gas CCS

Coal CCS

Gas

Coal

Peak demand



IRENE-40

Page: 165 of 383




acceptable level is 1,018GW and 1,211GW respectively. The aforementioned system characteristics have a direct impact on the transmission network infrastructure in order to deliver the required reliability of electricity of supply in each individual region in a cost optimal manner. The renewable energy sources contribution to energy demand rises from 45% in 2030 to 75% in 2050. Therefore transmission infrastructure needs to be reinforced in order to be able to integrate the significant levels of renewable energy sources in a secure and cost-effective manner. Thus, the transmission network infrastructure reinforcements are considerably higher in the year 2050. It can also be observed that the amount of renewable energy curtailed in the system in 2030 is very small whereas in 2050 there is 3% renewable energy curtailed. Despite all renewable energy sources are assumed to be fully utilised there can be some occasions where other plant’s related or system constraints lead to the curtailment of energy produced from renewable energy sources. For instance, curtailment may happen during periods where there is a coincidence of low demand, high renewable energy sources power output while thermal plant are operating due to their “must run” constraints or to provide reserve. Table 55 presents a summary of the main costs components of the system under analysis.

Table 55: RES scenario – summary of the main costs components of the system


cost [b€/Yr]


cost [b€]


[b€]



2030 200 2 (2010–2030)

1,203 (2010–2030)

11 (2010–2030)

2050 106 150 (2030–2050)

1,858 (2030–2050)

25 (2030–2050)

It can be observed in Table 55 that the year 2050 presents a lower annual operating cost compared to the year 2030. This is because year 2050 does not rely as much in the energy produced by fossil fuel generation technologies due to the significant higher contribution of renewable energy sources towards supplying energy demand. It can also be observed that 2030–2050 presents a higher investment cost in additional transmission capacity compared to 2010–2030 as the transmission infrastructure needs to be reinforced in order to be able to integrate the significant levels of renewable energy sources in a secure and cost-effective manner. It is also interesting to note that 2030–2050 has higher cost of investment in additional generation infrastructure reflecting the need to build peaking plants as backup plant in order to maintain security of supply at acceptable levels in systems with significant levels of renewable energy sources.



IRENE-40

Page: 166 of 383




Figure 79 presents the transmission network capacity of the present interconnectors, as given by the ENTSO-E for the year 2010, as well as the optimal levels of investment in transmission network infrastructure for the years 2030 and 2050.

Figure 81: RES scenario – Transmission network capacity requirements

It can be observed in Figure 79 that the highest levels of transmission infrastructure reinforcement take place in the in the year 2050. This conclusion has to some extent been earlier reflected by the higher annualised cost of investment in transmission presented in the year 2050 compared to the year 2030. The level of reinforcement required for some particular interconnectors will be subsequently discussed in more detail. Figure 80 and Figure 81 show the schematic representation of the main interconnected European transmission system with the respective level of transmission infrastructure reinforcement for the years 2030 and 2050 respectively. These represent the additional amounts of transmission network capacity required, compared to the present interconnectors, as given by the ENTSO-E for the year 2010.

0

5

10

15

20

25

30

35

PT-E

SES

-FR

FR-U

KFR

-BE

NO

-DE

FR-D

EFR

-CH

FR-IT

IT-C

HIT

-AT

IT-S

IDE

-DK

DE-P

LDE

-CZ

DE-A

TDE

-CH

PL-C

ZPL

-SK

CZ-S

KCZ

-AT

HU-S

KHU

-RO

HU-Y

UHU

-BO

HU-A

TYU

-RO

YU-B

GYU

-GR

YU-A

LYU

-BO

RO-B

GBG

-GR

GR-A

LBO

-SI

SI-A

TCH

-AT

UK-N

OUK

-NL

UK-B

EN

L-DE

NL-

BEN

L-N

ON

L-DK

NO

-DK

NO

-SE

SE-D

ESE

-PL

SE-F

ISE

-DK

LU-B

ELU

-DE

FI-E

EEE

-LV

LT-L

VLT

-PL

LT-S

E

Tran

smiss

ion

capa

city

[GW

]

ENTSO-E 2010RES 2030RES 2050



IRENE-40

Page: 167 of 383





Figure 82: RES scenario, Year 2030 – transmission network capacity reinforcements

FI

PT

DE

PL

SE

ES

IT

FR

UK

NO

NL

LU

CHAT

SI

0.0

CZ

RO

Bo&Cro

HU

BG

GR

YU

AL

SK

1.1 0.0

BE

EE

LV

LT

0.0

DK



IRENE-40

Page: 168 of 383





Figure 83: RES scenario, Year 2050 – transmission network capacity reinforcements

Figure 79 showed that the highest levels of transmission infrastructure reinforcement take place in the year 2050. Thus, based on the observation of Figure 81, which details the aforementioned findings, some particular and relevant interconnectors, where the reinforcement in network infrastructure is significantly high, are chosen and the transmission infrastructure reinforcements are scrutinised in more detail for the year 2050.

3.3.3.3.2.1 Interconnector: Spain – France

Figure 81 shows that the interconnector Spain – France requires a transmission infrastructure reinforcement of approximately 20GW when compared to the present interconnector’s capacity, as given by the ENTSO-E for the year 2010. Figure 82 represents the flow utilisation for this particular interconnector. Taking Spain as the reference country, a positive flow utilisation implies that Spain is importing from France, whereas a negative flow utilisation implies that Spain is exporting to France.

FI

PT

DE

PL

SE

ES

IT

FR

UK

NO

NL

LU

CHAT

SI

9.7

CZ

RO

Bo&Cro

HU

BG

GR

YU

AL

SK

3.3 0.6

BE

EE

LV

LT

1.5

DK



IRENE-40

Page: 169 of 383




Figure 84: Interconnector Spain – France: flow utilisation

Figure 85: Spain – normalised wind power output

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Flow

util

isat

ion

[%]

Interconnector: Spain - France

0%

20%

40%

60%

80%

100%


Nor

mal

ised

pow

er o

utpu

t [%

]

Spain: Wind



IRENE-40

Page: 170 of 383




Figure 86: Spain – normalised solar power output

It can be observed in Figure 82 that Spain has a very high flow utilisation (negative) during winter/spring season meaning that Spain exports energy to France during most of these seasons of the year. Figure 83 shows a high availability of wind power output during winter/spring season which is complemented by also high availability of solar power output especially during the spring season, as seen in Figure 84. There is therefore a significant correlation between the availability of the renewable energy sources in Spain and the exports of energy from Spain to France. Conversely, the most of the imports from Spain, leading to 100% flow utilisation, occur during summer season which coincides with low availability of wind power output in Spain and a lower availability of solar power output when compared to spring season in Spain. Figure 85 and Figure 86 show the daily correlation between the energy produced by renewable energy sources and the flow utilisation in the Spain – France interconnector.

Figure 87: Spain – correlation between renewable energy sources and maximum energy exports

0%

20%

40%

60%

80%

100%


Nor

mal

ised

pow

er o

utpu

t [%

]

Spain: Solar

0

20

40

60

80

100

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Pow

er [G

W]

Flow

util

isat

ion

[%]

Time [hour]

Flow utilisation Load Wind Solar



IRENE-40

Page: 171 of 383




It can be seen in Figure 85 that there is a significant degree of correlation between high level of energy produced by renewable energy sources and the maximum flow utilisation during periods where Spain is exporting energy to France. From eleven o’clock in the morning until midnight the interconnector Spain – France is being fully utilised in order to export high levels of energy produced from renewable energy sources from Spain to France. Figure 86 shows the opposite effect. There is a significant degree of correlation between low level of energy produced by renewable energy sources and the maximum flow utilisation during periods where Spain is importing energy from France.

Figure 88: Spain – correlation between renewable energy sources and maximum energy imports

It can be observed in Figure 86 that in periods characterised by low energy produced by renewable energy sources the flow utilisation of the Spain – France interconnector becomes 100%, reflecting the need of Spain importing energy from France.

3.3.3.3.2.2 Interconnector: France – Belgium Figure 81 shows that the interconnector France – Belgium requires a transmission infrastructure reinforcement of approximately 27.5GW when compared to the present interconnector’s capacity, as given by the ENTSO-E for the year 2010.

0

20

40

60

80

100

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Pow

er [G

W]

Flow

util

isat

ion

[%]

Time [hour]

Flow utilisation Load Wind Solar



IRENE-40

Page: 172 of 383




Figure 89: Interconnector France – Belgium: flow utilisation

Figure 90: Belgium – normalised wind power output

Figure 91: Interconnector Spain – France: flow utilisation

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%


Flow

util

isat

ion

[%]

Interconnector: France - Belgium

0%

20%

40%

60%

80%

100%


Nor

mal

ised

pow

er o

utpu

t [%

]

Belgium: Wind

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%


Flow

util

isat

ion

[%]

Interconnector: Spain - France



IRENE-40

Page: 173 of 383




Belgium is characterised by a firm plant margin of 41% which makes a considerable contribution to security of supply. Besides the significant levels of installed firm capacity, Belgium has 14GW of installed wind capacity and 8GW of installed solar capacity facilitating the export of energy to neighbouring countries. France is characterised by a firm plant margin of -1%, which in order to maintain security of supply within acceptable levels, France has to rely on its relatively high level of renewable energy resources and also on its interconnection with the neighbouring countries. It can be seen in Figure 87 that France imports (100% flow utilisation) from Belgium during periods of high wind power output in Belgium, Figure 88. It also interesting to note, in Figure 87 and Figure 89, that there is a period from April to June and another period from August to September where France exports to Belgium due to the incoming flow to France from the renewable energy sources placed in Spain.

3.3.3.3.2.3 Interconnector: United Kingdom – Belgium Figure 81 shows that the interconnector United Kingdom – Belgium requires a transmission infrastructure reinforcement of approximately 12GW when compared to the present interconnector’s capacity, as given by the ENTSO-E for the year 2010.



IRENE-40

Page: 174 of 383




Figure 92: Interconnector UK – Belgium: flow

utilisation Figure 93: UK – normalised wind power output

Figure 94: Interconnector France – Belgium: flow

utilisation Figure 95: Belgium – normalised wind power output

It can be seen in Figure 91 that from January to April there are relatively good levels of wind power in the UK, therefore UK exports to Belgium, Figure 90. It can be also observed Figure 93 that, during the same period, Belgium also presents relatively good levels of wind power, which combined with UK exports to Belgium, leads in turn Belgium export to France. The last set of Figures show that there is a period from April to June and another from August to September, Figure 90, where UK imports from Belgium which in turn, Belgium imports from France, Figure 92, energy relative to the renewable energy sources placed in Spain that come through France. The last set of Figures also show that in the period from November to December the presence of relatively high wind power output in the UK coinciding with also relatively high wind power output in Belgium results in energy exports from the UK to Belgium and then from Belgium to France.

3.3.3.3.2.4 Interconnector: Germany – Denmark

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%


Flow

util

isat

ion

[%]

Interconnector: UK - Belgium

0%

20%

40%

60%

80%

100%


Nor

mal

ised

pow

er o

utpu

t [%

]

UK: Wind

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%


Flow

util

isat

ion

[%]

Interconnector: France - Belgium

0%

20%

40%

60%

80%

100%


Nor

mal

ised

pow

er o

utpu

t [%

]

Belgium: Wind



IRENE-40

Page: 175 of 383




Figure 81 shows that the interconnector United Kingdom – Belgium requires a transmission infrastructure reinforcement of approximately 17GW when compared to the present interconnector’s capacity, as given by the ENTSO-E for the year 2010.

Figure 96: Interconnector Germany – Denmark: flow utilisation

Figure 97: Denmark – normalised wind power output

-100%

-80%

-60%

-40%

-20%

0%

20%

40%

60%

80%

100%


Flow

util

isat

ion

[%]

Interconnector: Germany - Denmark

0%

20%

40%

60%

80%

100%


Nor

mal

ised

pow

er o

utpu

t [%

]

Denmark: Wind



IRENE-40

Page: 176 of 383




Figure 98: Denmark – normalised solar power output

Figure 94 shows that Germany imports energy from Denmark throughout all seasons of the year. Denmark is characterised by a peak demand of 11GW and an installed “firm” capacity of 16GW, leading to a relatively high firm plant margin of 46% contributing significantly to security of supply. Besides the significant levels of installed firm capacity, Denmark has also a significant level of installed capacity of renewable energy sources, 21GW, of which 13.5GW are of relatively high wind power output, Figure 95. Such attributes makes Denmark to export significant levels of energy to the neighbouring countries throughout all the seasons of the year. Nevertheless, please note that there will be occasions where Denmark needs to import energy from neighbouring countries. Germany has a relatively modest firm plant margin of 13%, which in order to maintain security of supply within acceptable levels; Germany has to rely on its relatively high level of renewable energy resources and also on its interconnections with the neighbouring countries.

3.3.3.3.3 Desertec scenario Desertec scenario is characterised by an ambitious generation technology mix for achieving emission reduction targets and strong policy support is therefore assumed to drive the deployment of renewable energy sources especially after the year 2020. Thus, renewable energy sources dominate the generation of electricity of supply in 2050. The Desertec scenario assumes a strong development of renewable energy sources very similar to the Renewable scenario. The main difference is that part of Europe’s electricity of supply will be generated in North Africa, and then transported via electricity “high-ways” to Europe entering at Spanish and Italian borders. The Desertec scenario aims to supply 3% of Europe’s electricity demand by 2030 and 15% in 2050 from solar power generation located in North Africa. The power plants installed in North Africa as part of Desertec are physically located outside of Europe, but are assumed to be fully dedicated for the supply of electricity to Europe.

0%

20%

40%

60%

80%

100%


Nor

mal

ised

pow

er o

utpu

t [%

]

Denmark: Solar



IRENE-40

Page: 177 of 383




Figure 97 and Figure 98 show respectively the generation and demand background for the year 2030 and year 2050 for each European country considered in the analysis.

Figure 99: DES scenario, Year 2030 – generation and demand background

0

50

100

150

200

250

300

350

400

450

500


[GW

]

New capacity

Other RES

Storage

Hydro reservoir

Hydro ROR

Geothermal

Biomass

CSP

PV

Wind

Oil

Nuclear

Gas CCS

Coal CCS

Gas

Coal

Peak demand



IRENE-40

Page: 178 of 383




Figure 100: DES scenario, Year 2050 – generation and demand background

It is observed in Figure 97 and Figure 98 the presence of significant levels of installed capacity of Concentrated Solar Power (CSP) in Spain and Italy. The DES scenario assumes that part of Europe’s electricity will be generated in North Africa, and then transported via electricity “high-ways” to Europe entering at the Spanish and Italian boarders19

. DES scenario also assumes that all electricity produced from CSP technologies located in Morocco will be imported from Spain (“one way flow”) and all electricity produced from CSP technologies located in Tunisia will be imported from Italy (“one way flow”).

As a result of the application of the system analysis methodology, Table 56 presents a summary of the main characteristics of the system.

Table 56: DES scenario – summary of the main characteristics of the system

Year Peak

demand [GW]




capacity [TW-km]



[TW-km]

RES contribution

[%]

RES curtailment

[%]

2030 650 4,212 1,011 650 9 (2010–2030) 52 2

2050 827 5,188 1,194 817 147 (2030–2050) 78 4

19 IRENE-40, Task 2.3, Internal Report, “Scenario synthesis”.

0

50

100

150

200

250

300

350

400

450

500


[GW

]

New capacity

Other RES

Storage

Hydro reservoir

Hydro ROR

Geothermal

Biomass

CSP

PV

Wind

Oil

Nuclear

Gas CCS

Coal CCS

Gas

Coal

Peak demand



IRENE-40

Page: 179 of 383




DES scenario for the year 2030 is characterised by a peak demand of 650GW and total energy demand of 4,212GWh, whereas for the year 2050 is characterised by a peak demand of 817GW and total energy demand of 5,188GWh. The minimum installed “firm” capacity, necessary to be present in the system to ensure that the system reliability (LOLE ≤ 4 hours/years) is maintained at acceptable level is 1,011GW and 1,194GW respectively. The aforementioned system characteristics have a direct impact on the transmission network infrastructure in order to deliver the required reliability of electricity of supply in each individual region in a cost optimal manner. The renewable energy sources contribution to energy demand rises from 52% in 2030 to 78% in 2050. Therefore transmission infrastructure needs to be reinforced in order to be able to integrate the significant levels of renewable energy sources in a secure and cost-effective manner. Thus, the transmission network infrastructure reinforcements are considerably higher in the year 2050. It can also be observed that the amount of renewable energy curtailed in the system is 2% 2030 and 4% in 2050. Despite all renewable energy sources are assumed to be fully utilised there can be some occasions where other plant’s related or system constraints lead to the curtailment of energy produced from renewable energy sources. Table 57 presents a summary of the main costs components of the system under analysis.

Table 57: DES scenario – summary of the main costs components of the system


cost [b€/Yr]


cost [b€]


[b€]



2030 165 10 (2010–2030)

1,938 (2010–2030)

11 (2010–2030)

2050 85 173 (2030–2050)

2,063 (2030–2050)

13 (2030–2050)

It can be observed in Table 57 that the year 2050 presents a considerable lower annual operating cost compared to the year 2030. This is because year 2050 does not rely as much in the energy produced by fossil fuel generation technologies due to the significant higher contribution of renewable energy sources towards supplying energy demand. It can also be observed that 2030–2050 presents higher cost of investment in additional transmission capacity compared to 2010–2030 as the transmission infrastructure needs to be reinforced in order to be able to integrate the significant levels of renewable energy sources in a secure and cost-effective manner. It is also interesting to note that 2030–2050 has higher cost of investment in additional generation infrastructure reflecting the need to build peaking plants as backup plant in order to maintain security of supply at acceptable levels in systems with significant levels of renewable energy sources.



IRENE-40

Page: 180 of 383




Figure 99 presents the transmission network capacity of the present interconnectors, as given by the ENTSO-E for the year 2010, as well as the optimal levels of investment in transmission network infrastructure for the years 2030 and 2050.

Figure 101: DES scenario – Transmission network capacity requirements

It can be observed in Figure 99 that the highest levels of transmission infrastructure reinforcement take place in the in the year 2050. This conclusion has to some extent been earlier reflected by the higher annualised cost of investment in transmission presented in the year 2050 compared to the year 2030. The level of reinforcement required for some particular interconnectors will be subsequently discussed in more detail. Figure 100 and Figure 101 show the schematic representation of the main interconnected European transmission system with the respective level of transmission infrastructure reinforcement for the years 2030 and 2050 respectively. These represent the additional amounts of transmission network capacity required, compared to the present interconnectors, as given by the ENTSO-E for the year 2010.

0

5

10

15

20

25

30

35

PT-E

SES

-FR

FR-U

KFR

-BE

NO

-DE

FR-D

EFR

-CH

FR-IT

IT-C

HIT

-AT

IT-S

IDE

-DK

DE-P

LDE

-CZ

DE-A

TDE

-CH

PL-C

ZPL

-SK

CZ-S

KCZ

-AT

HU-S

KHU

-RO

HU-Y

UHU

-BO

HU-A

TYU

-RO

YU-B

GYU

-GR

YU-A

LYU

-BO

RO-B

GBG

-GR

GR-A

LBO

-SI

SI-A

TCH

-AT

UK-N

OUK

-NL

UK-B

EN

L-DE

NL-

BEN

L-N

ON

L-DK

NO

-DK

NO

-SE

SE-D

ESE

-PL

SE-F

ISE

-DK

LU-B

ELU

-DE

FI-E

EEE

-LV

LT-L

VLT

-PL

LT-S

E

Tran

smiss

ion

capa

city

[GW

]

ENTSO-E 2010DES 2030DES 2050



IRENE-40

Page: 181 of 383





Figure 102: DES scenario, Year 2030 – transmission network capacity reinforcements

FI

PT

DE

PL

SE

ES

IT

FR

UK

NO

NL

LU

CHAT

SI

0.0

CZ

RO

Bo&Cro

HU

BG

GR

YU

AL

SK

0.9 0.0

BE

EE

LV

LT

0.0

DK

Documents

DELIVERABLE 3.1chatziva.com/publications/W3SI_DV_8008_part_I.pdf¾ Power system security aspects are best served by VSC-HVDC reinforcements • 10 to 18 Gvar2 Economic competitiveness: