at SciVerse ScienceDirect

Energy 55 (2013) 716e727

Contents lists available

Energy

journal homepage: www.elsevier .com/locate/energy

Exergy analysis of the oil and gas processing on a North Sea oilplatform a real production day

Mari Voldsund a,*, Ivar Ståle Ertesvåg b, Wei He c, Signe Kjelstrup a

aDepartment of Chemistry, Norwegian University of Science and Technology, NO-7491 Trondheim, NorwaybDepartment of Energy and Process Engineering, Norwegian University of Science and Technology, NO-7491 Trondheim, Norwayc Statoil ASA, NO-5254 Sandsli, Norway

a r t i c l e i n f o

Article history:Received 15 November 2012Received in revised form8 February 2013Accepted 11 February 2013Available online 10 April 2013

Keywords:ExergyEfficiencyGas compressionOffshore platform

* Corresponding author. Tel.: þ47 90748259; fax: þE-mail addresses: mari.voldsund@chem.ntnu.no

(M. Voldsund), ivar.s.ertesvag@ntnu.no (I.S. Ertesvågsigne.kjelstrup@chem.ntnu.no (S. Kjelstrup).

0360-5442/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.energy.2013.02.038

a b s t r a c t

We explore the applicability of exergy analysis as an evaluation and monitoring tool for the oil and gasprocessing on an offshore platform. A real production day on a particular North Sea platform is analysed.A process flowsheet is simulated using measured process data. We distinguish between temperaturebased exergy, pressure based exergy and the mixing part of the chemical exergy. It is shown that physicalexergy in the material streams mainly is pressure based exergy, and most exergy destruction is related todecrease or increase in pressure. The sub-processes with most destructed exergy are the productionmanifold (4600 kW), the recompression train (4150 kW) and the reinjection trains (10,400 kW). At thisplatform 260 kW separation work is done, where a considerable part is done in the compression trains inaddition to in the separation train. The specific power consumption is 179 � 3 kWh/Sm3 and the exer-getic efficiency is 0.13 � 0.02. We propose measures to decrease exergy destruction, and that exergyanalysis should be taken into regular use by the oil and gas industry. This study serve as a showcase onhow to do an exact analysis of an existing offshore platform using measured process data.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

In 2009, gas turbines and diesel engines on oil and gas platformswere responsible for 21% of Norway’s total CO2-emissions [1]. Mostplatforms generate their own power with gas turbines, and thetypical power consumption at a North Sea offshore platform variesfrom around 10 MW to several hundred MW. There is a generalagreement that the world’s CO2 emissions should be reduced andthat the world’s resources should be utilised in a sustainable way.Improvement of energy efficiency is a challenge in the petroleumsector, as in the industry in general. The sector is therefore in needfor a tool to monitor the energy performance of the platformprocesses.

Today, specific CO2 emissions (CO2 emission per unit producedoil) are often used as a performance parameter by the oil and gasindustry. This parameter reflects the aim of reducing the world’s CO2emissions e it encourages energy efficiency and use of renewableenergy sources. However, it does not account for the varying oper-ating conditions for offshore platforms. Different platforms have

47 73550877., mari.voldsund@gmail.com), weih@statoil.com (W. He),

All rights reserved.

different well stream conditions, different pressure requirements forexport oil and injection/export gas etc. The same platform alsooperates under varying conditions, e.g. well streams that changeover time.

In Norway, as well as in a number of other countries, the industryhas to pay tax for CO2 and NOx emissions. At the same time increasedrecovery and extended lifetimes in mature fields is encouraged.However, measures designed to improve recovery often requiresignificant amounts of power and may entail additional emissions toair [2]. The taxes do then punish measures that are encouraged bythe authorities [3].

Exergy analysis is a thermodynamic method which is not yetsystematically used by the oil and gas industry. The exergetic effi-ciency takes into account theminimum theoretical work that has tobe done for a given process, and gives thus another perspectivethan specific CO2 emissions. By using exergy analysis, one can alsocalculate the destructed exergy in different parts of the process andindicate possibilities for improvement. We want to explore the useof exergy analysis as a tool for platform performance benchmarkingand as an everyday tool to evaluate performance. We thereforeanalyse the oil and gas processing a real production day on a NorthSea oil platform. We use measured process data for the specific dayto simulate a process flowsheet for that day and calculate des-tructed exergy and exergetic efficiency. To gain more insight about

M. Voldsund et al. / Energy 55 (2013) 716e727 717

the process, we also distinguish between temperature basedexergy, pressure based exergy and the mixing part of the chemicalexergy. This is the first study using exergy analysis on a North Seaplatform, which is the main motivation for this study.

There is one other exergy analysis of an offshore platformknown in the literature. This is an analysis of the oil and gas pro-cessing of a Brazilian platform performed by de Oliveira Junior andvan Hombeeck [4]. On this platform, power was consumed in orderto heat the petroleum for separation, to compress natural gas and topump oil to the coast. A heat recovery systemwas installed in orderto recover heat from exhaust gases for the petroleum heating. Boththe petroleum heating, despite the heat recovery system, and thecompression operations gave considerable exergy destruction. Theseparation process, that required heating, was pointed out as aplace to do improvements.

The North Sea oil platform we analyse was built more than 20years ago. The platformwas chosen because it is a relatively simpleone, but contains still all processes typical for such platforms. Itexports oil and reinjects gas in the reservoir for pressure mainte-nance. The power consumption of the entire platformvaries around34 MW and in the oil and gas processing part, which is analysedhere, it varies around 24 MW. The boundary conditions of thesystem are dictated by the conditions of the well streams, specifi-cations on vapour pressure and water content for the produced oil,pressure required for the export through the export pipelines,pressure required in the injection gas and pressure- and tempera-ture specifications for the fuel gas. Initial studies of the oil and gasprocessing at this platformwere presented by Voldsund et al. [5,6].

This is the first study based onmeasured process data, and it canserve as a showcase on how to do an exact assessment of a specificplatform using exergy analysis.

2. Theoretical background

2.1. Exergy

The exergy of a system is defined as the maximum theoreticalwork obtainable when the system interacts with the environmentto reach equilibrium. This maximum theoretical work is obtainedwhen all processes involved are reversible. In all real processessome exergy will be destructed. In an exergy analysis of a process,thermodynamic inefficiencies can be identified. For a comprehen-sive introduction to exergy analysis, see the textbook of Kotas [7] orMoran and Shapiro [8]. For a thorough review of applications ofexergy analysis, see the textbook of Dincer [9]. Important quantitiesin exergy analysis are:

� The product exergy, EP, is the desired result expressed in termsof exergy.

� The utilized exergy, EU, is the resources in terms of exergy usedto provide the product exergy.

� Exergy loss, EL, is thermodynamic inefficiencies of a systemassociatedwith the transfer of exergy with energy andmaterialstreams to the surroundings.

� Exergy destruction, ED, is thermodynamic inefficiencies of asystem associated with the irreversibilities (entropy genera-tion) within the system boundaries.

For a system in steady state the destructed exergy for a certaintime period, ED, is the exergy entering the systemminus the exergyleaving the system:

ED ¼ W þXk

ZTk

�1� T0

Tk

�dQk þ

Xj

njej; (1)

whereW is the work added, Qk the heat transferred into the systemat temperature Tk, T0 the ambient temperature, nj the number ofmoles in material stream j and ej the molar exergy in materialstream j. Some of the terms in Eq. (1) correspond to product exergy,some correspond to utilized exergy and some correspond to exergyloss, depending on the system considered, ED ¼ EU�EP�EL.

Exergy in a material stream can be split into physical (thermo-mechanical) exergy, chemical exergy, kinetic exergy and potentialexergy. Given on molar form, we have: e ¼ eph þ ech þ ekin þ epot.

The molar physical exergy accounts for deviation from thermaland mechanical equilibriumwith the environment and is given by:

eph ¼ h� h0 � T0ðs� s0Þ; (2)

where h and s are the molar enthalpy and entropy, and h0 and s0 arethe molar enthalpy and entropy for the same stream but at T0 andp0, where p0 is the ambient pressure.

The physical exergy can be divided into temperature based andpressure based exergy, eT and eP, respectively. The most commonway to do this is:

eT ¼ h� hðT0Þ � T0ðs� sðT0ÞÞ; (3)

and

eP ¼ hðT0Þ � h0 � T0ðsðT0Þ � s0Þ; (4)

where h(T0) and s(T0) is enthalpy and entropy evaluated at theinitial pressure of the stream and T0.

The molar chemical exergy accounts for deviation from a statewith only thermal and mechanical equilibrium with the environ-ment to a state with also chemical equilibrium with the environ-ment, and is given by:

ech ¼Xi

xiei|fflfflfflffl{zfflfflfflffl}I

þ h0 �

Xi

xihi;0 � T0

s0 �

Xi

xisi;0

!!|fflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}

II

; (5)

where term I accounts for the molar chemical exergy of each of thecomponents when they are pure and term II accounts for mixingeffects. The symbol xi is the mole fraction of chemical component i,ei the molar chemical exergy of pure i, hi,0 the molar enthalpy ofpure i at T0 and p0 and si,0 the molar entropy of pure i at T0 and p0.

The kinetic and potential exergy of a material stream is equal tothe kinetic and potential energy of the material stream.

2.2. Process performance parameters

There exists a variety of ways to define performance parametersfor industrial processes based on energy and exergy. We presentsome parameters that are useful for oil and gas processing:

� The specific power consumptionwe define as consumed powerper unit oil produced. As long as all power comes from the samefossil fuel source, this is proportional to specific CO2 emissions.

� The exergetic efficiency, ε, is:

ε ¼ EPEU

: (6)

This parameter takes into account the minimum theoreticalwork that has to be done for a given process. The interpretatonof EP and EU vary. We define EP and EU for our system inSection 4.2.

M. Voldsund et al. / Energy 55 (2013) 716e727718

� The efficiency defect, di, of subsystem i is the fraction of theinput exergy to the whole system which is lost through irre-versibilities in the subsystem [7]:

di ¼ED;iEU

: (7)

This parameter shows how different subsystems contribute toreduction of the exergetic efficiency.

3. System description

The studied North Sea platform is an oil producing platform thathas been in production for more than 20 years. Two power turbinesproduce electric power that covers all power needed at the plat-form. Fuel gas for the power turbines is taken from the gas pro-duced at the platform (approximately 3%). In the oil and gasprocessing, reservoir fluids are separated into oil, gas and water.The produced oil is pumped 18 km to a nearby platform for export,the water is rinsed and released to the sea, and the produced gas isrecompressed and reinjected into the reservoir for pressure main-tenance. Pressure maintenance is also achieved by injection of highpressure water, but this water is provided by a nearby platform. Anoverview of the oil and gas processing and characteristics for thestudied day are given below.

3.1. Process overview

A schematic overview of the oil and gas processing at thestudied platform is given in Fig. 1.

A number of production wells are connected to the platformand the well streams have different temperatures, pressures andoile, gase and water fractions. The wells are connected to the oiland gas processing section via a production manifold. In themanifold a selection of thewells are set into production. For the daywe study, the pressures of the producingwells vary between 80 and170 bar. The pressures are reduced to approximately 70 bar beforethe streams are mixed. The resulting stream consists of reservoirfluids with 78 mol% gas, and is sent to a separation train.

In the separation train gas and water are separated from thecrude oil using gravitational separators and an electrostatic coa-lescer. The train consists of three stages where in the first twostages there are three-phase separators, and in the third stage thereis a two-phase separator and an electrostatic coalescer. For eachseparator, the pressure is reduced, so that more gas is released fromthe oil. The oil shall meet specifications of basic sediment content,water content and vapour pressure. In total the pressure is reducedfrom 71 to 2.8 bar during this section. Oily water from the sepa-rators is sent to a water treatment process where traces of oil areremoved. This process is not included in this analysis. Water fromthe electrostatic coalescer is pumped back to the 2nd-stage sepa-rator. In contrast to the platform studied by de Oliveira Junior andvan Hombeeck [4], no heating is required in the separation process.

The remaining, stabilised oil is pumped via two pumps withcooling of the oil between, to meet the pressure conditions in thetransportation pipeline. A minimum flow is required through thepumps, and to achieve this, some of the oil is recycled back to rightafter the 2nd-stage separator.

The gas that is released in each stage in the separation train, issent to the recompression train. The train consists of three stages,each with a cooler, a scrubber and a compressor. The gas is cooledfor a more efficient compression. The scrubber is a separator thatremoves small amounts of condensed liquid. Scrubbing protectsthe compressor and allows more optimal compression. In the end

of the train the pressure has reached the 1st-stage separationpressure.

Since the platform is more than 20 years old, and the gas to oilratio (GOR) in the feed has changed over time, the flow rates in therecompression train is lower than what the train was designed for.A minimum flow of gas is required through the compressors toprevent surging, and some of the gas is therefore recycled aroundeach stage (anti-surge recycling). The fractions of gas that arerecycled are 92%, 69% and 72% in the 1st, 2nd and 3rd recom-pression stages, respectively.

After recompression, the gas enters the reinjection trains. Theseare three parallel trains where the gas is compressed up to injectionwell pressure. In each of these trains there are two stages, eachwitha cooler, a scrubber and a compressor e the same way as in therecompression train. The train is run at maximum capacity, so thereis no need for anti-surge recycling. High pressure gas leaves thesystem for injection back into the reservoir through 5 injectionwells.

Fuel gas is taken from the 1st-stage separator and is cooled andfed through a pressure reducing control valve to a scrubber forliquid removal. After the scrubber, the gas is heated with an elec-trical heater before a last possible liquid removal, before it is sent tothe power turbines. Gas for a pilot flame at the flare is also takenfrom this section. For normal process conditions, the amount of gasto the flare from other parts of the processes is negligible.

Condensate from scrubbers throughout the processes is sentback to the 2nd separation stage, either directly or through a drainsystem. The drain system consists of several tanks and smallpumps, but is for simplicity looked upon as only one mixer and onepump in this study.

3.2. Process characteristics

We had available process data measured from 2009 to 2011.Many process variables change from day to day due to variation inthe well conditions and since the operators switch betweendifferent wells during production. To obtain a consistent andrepresentative flowsheet we decided to use the average values for aperiod of one day with stable conditions. We defined the 85% of thedays that were closest to the median value for selected processvariables as ‘normal production days’, since it was found that withthis criterion, days with shutdowns or other major disturbanceswere excluded. The day studied here is one of these ‘normal pro-duction days’. The process conditions were stable throughout thisday in the sense that the standard deviation in measured producedoil flow rate is less than 10 Sm3/h (for an average flow rate of132.5 Sm3/h) and the injected gas flow rate measured in each of thein total 5 injection wells is less than 103 Sm3/h (for an average totalflow rate of 369 � 103 Sm3/h).

The varying well conditions lead to variation in several param-eters, which are important for the performance of the oil platform.The variation in these parameters are summarised in Table 1. Thestudied day has a medium gas injection flow rate, a low oil pro-duction flow rate and a high gas injection pressure.

The adiabatic efficiencies calculated from measured inlet andoutlet temperatures and pressures for the real production day aregiven in Table 2.

4. Methodology

4.1. Simulation of the process flowsheet

The chemical processes were simulated using the processsimulator Aspen HYSYS. We used the property package where thePengeRobinson equation of state [10] is used to calculate

Drain system

Separation

Production manifold

Export pumping

Recompression

Reinjection

Fuel gas treatment

80-

170

bar

70 bar2.8 bar

32 bar132.5Sm3/h

2.8 bar

70 bar

75 ˚C

70 bar147 ˚C

236 bar369 103

Sm3/h

123 ˚C

103 ˚C

92 ˚C92 %69 %72 %

10 · 103

Sm3/h

369 · 103

Sm3/h

9 · 103

Sm3/h

93 ˚C

96 ˚C

94 ˚C

67 Sm3/hProduced

water

Oilexport

Gasinjection

Toflares

Topowerturbines

EC

Fig. 1. Process flowsheet of the oil and gas processing for the studied production day. Gas streams are orange, water streams are blue and oil, condensate or mixed streams arebrown. Well streams (well 7, 16, 23, 24 and 26) enter the process in a production manifold where the pressures are decreased and the streams are mixed. The resulting mixed streamenters the separation train where it is separated into gas, oil and water. The water is sent out of the process, the oil is sent to the export pumping section where it is pumped forexport, high pressure gas is sent to the reinjection trains and low pressure gas is sent to the recompression train where it is compressed before it is sent to the reinjection trains. Inthe reinjection trains the gas is further compressed before it is reinjected into the reservoir. Fuel gas is taken from the 1st separation stage, and treated in the fuel gas system. It iscombusted in power turbines and in pilot flames in the flare system. There is a drain systemwhere some small liquid streams are collected and pumped back to the separation train.It consists of several units, but a simplified version is used in this study. (For interpretation of the references to colour in this figure legend, the reader is referred to the web versionof this article.)

M. Voldsund et al. / Energy 55 (2013) 716e727 719

thermodynamic properties. This package is the recommendedproperty package for oil and gas applications [11]. The HYSYS PRoption was chosen and liquid densities were calculated with theCOSTALD method. The HYSYS PR option has several enhancementsto the original PengeRobinson equation and the COSTALD methodcalculates better liquid densities than the equation of state [11].

Table 1Process parameters important for the performance of the oil platform for all normalproduction days from 2009 to 2011. Maximum and minimum levels of the param-eters are given, together with values for the studied production day.

Process parameter Max/Min Studied day

Gas injection flow rate, 103 Sm3/h 323/392 369Oil production flow rate, Sm3/h 121.6/302.6 132.5Gas injection pressure, bar 210/240 236

Interaction coefficients were taken from the HYSYS library and theinteraction parameters unavailable from the library were set asestimated by HYSYS. Hypothetical components were used tosimulate the heavy oil fractions. Hypothetical components aremade-up components that represent oil fractions that can consistof a number of different real components.

Mean values for measured process variables like flow rates,temperatures and pressures for the studied production day wereused as input variables in the simulation together with some valuesfound in documentation of equipment (pump efficiencies and apressure drop over a scrubber). The hypothetical components usedto describe the heavy oil fractions and the composition of theoile, gase and water phases were taken as developed by the oilcompany. It was assumed that these compositions do not changeover time. Other measured process variables were used to validate

Table 2Adiabatic compressor efficiencies for the production day analysed. The efficienciesare calculated from the measured inlet and outlet temperatures and pressures of thecompressors, given in Tables A.12 and A.13 in Appendix A.

Compressor Adiabatic efficiency

Recompression1st stage 47%2nd stage 69%3rd stage 56%Reinjection A1st stage 64%2nd stage 54%Reinjection B1st stage 64%2nd stage 57%Reinjection C1st stage 69%2nd stage 64%

M. Voldsund et al. / Energy 55 (2013) 716e727720

the model. Details about the hypothetical components, thecomposition of the reservoir fluids, and the measured processvariables, used as input variables and for validation, are given inAppendix A.

4.2. Exergy analysis

The ambient temperature was set to 8 �C, as this is the averagetemperature for the North Sea throughout the year [12].

The exergy destruction in each process unit was found using theexergy balance of the unit. Exergy in the material streams werecalculated creating user variables in HYSYS programmed with Vi-sual Basic. Physical exergy was calculated as described byAbdollahi-Demneh et al. [13]. Temperature- and pressure basedexergy was calculated by modifying the same code. New code wasdeveloped that calculated the mixing part of the chemical exergy(term II in Eq. (5)). The component chemical exergy (term I in Eq.(5)) was not taken into consideration, since this exergy only passesthrough the system. Contributions from kinetic and potentialexergy were neglected.

All cooling in the system is done with cooling water dischargedirreversibly to the sea. The exergy leaving the system in form ofthermal energy was therefore regarded as destructed exergy. Thismeans in practice that the system boundaries were drawn aroundthe points where cooling water and sea water are mixed. The smallamount of heating that is required is done with an electric fuel gasheater. Since no thermal energy enters the system and all thermalenergy leaving the system was regarded destructed, the exergybalance (Eq. (1)) reduced to:

ED ¼ W þXj

njej: (8)

For the calculation of the exergetic efficiency, EP was defined asthe exergy difference between process streams leaving andentering the system and EU as the power delivered to the processunits (after electrical and mechanical losses). This is identical to theexergetic efficiency defined by de Oliveira Junior and van Hom-beeck for their system, except for that they used the exergy of thefuel gas as utilised exergy, not the power consumption, meaningthat they included the power turbines in the analysis. The exergeticefficiency defined this way corresponds to the theoretical mini-mum exergy input required to drive the process with the currentboundary conditions for the material streams, divided by the actualexergy input. No streams leaving the system were considered asexergy loss, giving EL ¼ 0.

4.3. Uncertainty analysis

The dominant contributions to uncertainty in the calculationswere the uncertainty in measured process variables and uncer-tainty from inaccuracies in the equation of state.

Uncertainties (with 95% confidence interval) were determinedfor all measured process variables. At the platform, some mea-surements (fiscal measurements) are subject to requirements foruncertainty set by the authorities (The Norwegian PetroleumDirectorate and The Climate and Pollution Agency). For these var-iables we assumed that the real measurement uncertainties wereequal to the set limits. For the remaining measured variables, un-certainties were taken as set in the oil company’s own guidelinesfor accuracy. However, for some of these variables the uncertaintieswere adjusted after discussions with the operators of the platformand in accordance with the authors’ experience with the dataset.

When all errors are independent and random, the magnitude ofthe uncertainty sA in variable A is given by:

sA ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiXni¼1

�vAvxi

sxi

�2vuut ; (9)

where A depends on n other variables xi, each with uncertainty sxi[14]. Propagation of uncertainty from measured variables to thecalculated performance parameters and destructed exergies wascalculated using Eq. (9) with the following approximation:

vAvxi

¼ Aðxi þ sxi Þ � Aðxi � sxiÞ2sxi

: (10)

The values for A(xi þ sxi) and A(xi�sxi) were found for each xi bysimulating the process with the value of xi changed and everythingelse kept constant.

Inaccuracies in calculated destructed exergy, power consump-tion and performance parameters originating from the chosenequation of state (PengeRobinson) were found by focusing onmethane, which is a key component in the process. Enthalpy andphysical exergy for methane at relevant temperatures and pres-sures calculated with the PengeRobinson equation of state wascompared with enthalpy and physical exergy for methane at thesame temperatures and pressures calculated with a presumablymore accurate equation of state developed by Setzmann andWagner [15]. The magnitude of the uncertainty of calculated powerconsumption was assessed by the difference in calculated enthalpychange for methane with the two equations of state for the tem-perature and pressure in and out of the unit. The magnitude of theuncertainty of calculated destructed exergy was assessed by thedifference in calculated destructed exergy, taking physical exergyin, physical exergy out and (if relevant) power in to the unit intoaccount, for methane with the two equations of state.

The uncertainty reported, sA,comb, was a combination of theuncertainty frommeasurements, sA,m, and the uncertainty from thestate equation, sA,eos:

sA;comb ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffis2A;m þ s2A;eos

q: (11)

5. Results

5.1. Exergy flows entering and leaving the process

Different types of exergy entering and leaving the process aregiven in Table 3. The component part of the chemical exergy, term Iin Eq. (5), is not included in the analysis. This is a high value due to

Table 3Mass and exergy flows entering and leaving the oil and gas processing at the plat-form for the analysed production day. For chemical exergy only the mixing part isgiven (term II in Eq. (5)). The power exergy is power added to the process units afterelectric losses and mechanical losses (DH for the units).

Exergy stream Mass flow,ton/h

Chemical exergy,mixing part, kW

Physicalexergy, kW

Power, kW

InWell 7 90 �1220 7940Well 16 99 �1840 11,200Well 23 112 �1680 11,260Well 24 103 �1780 10,150Well 26 89 �1500 8810Power 23,800Total in 494 �8020 49,350 23,800OutOil export 108 �891 298Gas injection 310 �6680 50,580Produced water 68 �1 539Flare gas 0.3 �6 21Fuel gas 8 �174 763Total out 494 �7750 52,210 0

Table 5Power consumption, destructed exergy and efficiency defect for each sub-process inthe oil and gas processing at the platform for the analysed production day. ‘Powerconsumption’ is power added to the process units (DH), after electric losses andmechanical losses.

Sub-process Power consumption,kW

Destructedexergy, kW

Efficiency defect, e

Production manifold 0 4600 � 400 0.195 � 0.017Separation train 0.117 � 0.017 800 � 200 0.034 � 0.008Export section 320 � 150 240 � 190 0.009 � 0.008Recompression train 4700 � 30 4150 � 190 0.173 � 0.008Reinjection trains 18,640 � 180 10,400 � 500 0.43 � 0.02Fuel gas system 156 � 2 508 � 5 0.0214 � 0.0004Total 23,800 � 400 20,700 � 500 0.867 � 0.013

M. Voldsund et al. / Energy 55 (2013) 716e727 721

the high energy content in the oil and the natural gas, but it has noimpact on calculated exergy differences, since the chemical com-ponents entering and leaving each control volume are the same.Term II in Eq. (5) expresses the mixing effects on the chemicalexergy and is shown for each flow. The difference in chemicalexergy for the system is due to mixing effects. When positive, achange in the mixing exergy is the minimum exergy (work) theo-retically required to conduct the separation in the system. Anegative value means that mixing takes place.

We see from Table 3 that around 49 MW enters the system inform of physical exergy in the well streams. This physical exergy ismainly due to high well stream pressures. The percentage of tem-perature- and pressure based exergy in the physical exergyentering and leaving the process is given in Table 4, and we see thatthe pressure based exergy dominates the physical exergy in thewell streams. Through process units like pumps and compressors23.8 MW power enter the process. The gas injection stream con-tains around 51 MW physical exergy. Due to the high injectionpressure, this stream is also dominated by pressure based exergy,see Table 4. Most of the exergy that enter the system, and that is notdestructed, leave the system with the injection gas. The otherstreams leaving the system contain less than 1MWphysical exergy.

5.2. Exergy transformations and exergy destruction in eachsubsystem

Power consumption, exergy destruction and efficiency defectsin the different sub-processes are given in Table 5. The change in

Table 4Percentage of temperature- and pressure based exergy components in the physicalexergy entering and leaving the process.

Exergy stream Temperature based exergy, % Pressure based exergy, %

InWell 7 10 90Well 16 7 93Well 23 9 91Well 24 7 93Well 26 7 93Total in 7 93OutOil export 55 45Gas injection 4 96Produced water 97 3Flare gas 2 98Fuel gas 2 98Total out 5 95

different types of exergy in the sub-processes is given in Table 6.The major contributions to exergy destruction distributed on thetype of process units in each sub-process are given in Fig. 2.

Table 5 and Fig. 2 show that exergy destruction takes placemainly in the reinjection trains, the recompression train and in theproduction manifold. Efficiency defects, cf. Eq. (7), for these sub-systems are 0.43, 0.17 and 0.19 respectively. If we compare with deOliveira Junior and van Hombeeck [4] we see the same high exergyloss due to compression. In addition they calculated a high exergyloss due to heating, which we do not have at this platform, sincethere is no need for heating to separate the well streams. They didnot include the production manifold in their analysis.

In the reinjection trains the pressure is increased from 70 bar to236 bar. In Tables 5 and 6 we see that from the 18,640 kW powerused in this section, 7990 kW is used to increase pressure basedexergy, 300 kW is used to increase chemical exergy and tempera-ture based exergy while 10,400 kW is destructed. Fig. 2 shows thatthe destructed exergy in the reinjection train is mainly exergydestruction in compressors and due to cooling, while a smallamount of exergy is destructed in mixers. Exergy is destructed inthe compressors because the compressor efficiencies are not 100%,so all power input is not used to increase the pressure. Some poweris transformed into thermal energy, giving exergy destruction.Higher compressor efficiencies will give less exergy destruction,but this is only possible up to a certain point with today’s tech-nology. A large part of this exergy destruction can therefore belooked upon as ‘unavoidable’. In the coolers thermal energy isremoved, and since the cooling water is mixed irreversibly with thesea, the associated exergy is destructed.

The pressure is increased from 2.8 bar to 70 bar in the recom-pression train. In Tables 5 and 6 we see that for the 4700 kW powerused in this section, the chemical and physical exergy is onlyincreased with 550 kWwhile the remaining 4150 kW is destructed.Exergy is destructed for the same reasons as in the reinjection train,but in addition there is exergy destruction due to the anti-surgerecycling around the compressors. This recycling gives the largestcontribution to the total exergy destruction in the reinjection train,

Table 6Change in chemical exergy and temperature- and pressure based exergy for eachsub-process. The power consumption minus the sum of the change in each of thesethree exergy types equals the exergy destruction given in Table 5.

Sub-process Chemical exergychange, kW

Temperaturebased Exergychange, kW

Pressure basedexergy change,kW

Production manifold �0.014 �960 �3660Separation train 50 �280 �590Export section 0 �50 130Recompression train 20 140 390Reinjection trains 190 110 7990Fuel gas system 4 �30 �320

0

1000

2000

3000

4000

5000

6000

Valv

es

Hea

tlos

san

dm

ixin

g

Valv

es

Mix

ing

Com

pres

sors

Coo

lers

Rec

ycle

stre

ams

Mix

ing

Com

pres

sors

Coo

lers

Mix

ing

Pum

ps

Elec

trica

lhea

ter

Valv

es

Production tropxEnoitcejnieRnoisserpmoceRnoitarapeSdlofinampumping

Fuel gas

Des

truct

edex

ergy

, kW

Fig. 2. The major contributions to exergy destruction distributed on type of process unit in the different sub-processes of the oil and gas processing at the platform for the dayanalysed.

Table 7Performance parameters for the oil and gas processing at the platform for theanalysed production day. ‘Power consumption’ is power added to the process units,after electric losses and mechanical losses in gears etc.

Parameter Value

Specific power consumption, kWh/Sm3 179 � 3Exergetic efficiency, e 0.13 � 0.02

M. Voldsund et al. / Energy 55 (2013) 716e727722

mainly due to lost pressure in the recycle streams. In addition, theexergy destruction in the compressors and coolers are much higherthan necessary, since much more gas than necessary is cooled andcompressed. This exergy destruction is a direct result of the factthat there are different flow rates in this sub-process compared towhat the process equipment was designed for. This is the samesituation as observed in a study of an Italian onshore gas treatmentfacility by Margarone et al. [16], where the system for power pro-duction was dimensioned for a higher production than the current,due to the natural depletion of the gas reservoirs.

In the production manifold, the pressures of the well streamsare reduced, and this is the reason for the destructed exergy. Themajor part of the exergy destruction is destruction of pressurebased exergy; 3660 kW.

The efficiency defect of the separation train is only around 0.03,see Table 5. The purpose of the separation train is to separate oil,gas and water, and thus increase the chemical exergy of the ma-terial streams. This is done in several stages with reducing pressure,so physical exergy is used to do separation work. In Tables 5 and 6we see that 870 kW physical exergy is used, and out of this 800 kWis destructed and 50 kW is transformed into chemical exergy. InFig. 2 the destructed exergy labeled ‘valves’ is exergy destructedonly due to pressure reduction in the main oil stream. The des-tructed exergy labeled ‘mixing’ is both exergy destruction due topressure reduction, since pressure is reduced in high pressurestreams that are mixed with low pressure streams, and exergydestruction due to mixing of streams with different temperaturesand concentrations. The small amount of power consumption inthe separation train is due to the small water pump.

The efficiency defect of the fuel gas system is near 0.2 and theamount of destructed exergy is 508 kW, see Table 5. We see in Fig. 2that this exergy destruction mainly take place in the electricalheater and in valves. In the electrical heater 156 kW high qualitypower is transformed into thermal energy.

In the export pumping section 320 kW power is used to increasethe pressure of the export oil. The efficiencies of the pumps are low,see Table A.16, but since the fluid is liquid, not gas, and the flowrates are low in this section, the exergy destruction here is lowcompared to other parts of the process and this sub-process has thelowest efficiency defect.

5.3. Separation work

We can see from Table 6 that approximately 260 kW separationwork is done. Both the reinjection trains and the recompressiontrain give significant contributions to this increase in chemicalexergy, in addition to the separation train. The recompression trainsgive the highest contribution with 190 kW. The chemical exergyincreases of the separation train and the recompression train arealso considerable, with 50 and 20 kW respectively. The scrubbing inthe compression trains is of high relative importance for the overallseparation. The high gas to oil ratio (GOR) at this particular plat-form gives a high significance to the exergy transformations andexergy destruction in the compression sections. With high gas flowrates, high flow rates of condensate are separated in these sections,and this gives an important contribution to the overall separation ofcomponents.

5.4. Performance parameters

Calculated performance parameters for the process are given inTable 7. The specific power consumption is proportional to the partof the specific CO2 emissions originating from the oil and gas pro-cessing part of the platform when natural gas is used for powerproduction. The parameter evaluates the power consumptionwithout taking the boundary conditions of the system into account.The exergetic efficiency compares the power used in the processwith the power needed if all internal processes were reversible. Themanagement of the process with set boundary conditions for thematerial streams is thus evaluated. The specific power consump-tion and the exergetic efficiency show different features of theprocess, and both parameters should be kept low.

M. Voldsund et al. / Energy 55 (2013) 716e727 723

5.5. Accuracy

The calculated uncertainties of destructed exergy from mea-surements and inaccuracies in the equation of state were of thesame order of magnitude. For some subsystems uncertainty frommeasurements dominated, while for other subsystems uncertaintyfrom inaccuracies in the equation of state dominated.

The absolute difference between enthalpy calculated formethane with the equation of state developed by Setzmann andWagner [15] and the PengeRobinson equation of state varied be-tween 0 and 13 kJ/kg for relevant temperatures and pressures(pressures between 1 atm and 500 bar and temperatures between25 �C and 175 �C). The same difference for physical exergy variedbetween 0 and 14 kJ/kg. The relative deviation of physical exergyvaried between 0 and 1.6 � 10�2.

For the mass- and energy balances, the relative deviations be-tween mass and energy flow in and out of the system were1.2 � 10�5 and 1.6 � 10�5, respectively. Errors from the mass- andenergy balance were negligible compared to the calculateduncertainties.

The neglect of contributions from potential exergy has noimpact on the overall exergy destruction and exergetic efficiency,since the feed- and product streams of the process (except for thesmall fuel gas stream) enter and leave the process at the sameelevation. However, it can be seen in Table A.18 that the neglect ofheight differences has led to a small error in the simulation. Someexergy destruction might have been assigned to wrong processunits, since measured process data is used as input in the simula-tion, and we have pressure increases and decreases due to heightdifferences. The process units are spread on 3 different levels with atotal height difference of approximately 30 m. Physical exergy thatin reality is transferred into potential exergy will give a highercalculated destructed exergy, and vice versa. In most cases thiserror is negligible compared to calculated uncertainties. Theexception is the separation train and the export pumping section,where calculated uncertainties are lower than the contributionfrom neglect of potential exergy. However, these sub-processeshave small efficiency defects, see Table 5, and the accuracy ofthese values is thus not critical for the overall results.

It was assumed that the compositions of the different phases inthe well streams were constant over time, so the compositions andhypothetical components developed by the oil company could beused in the simulation of the process the studied day. To test theimpact on the results, compositions and hypothetical componentsdeveloped by the oil company for a nearby reservoir was used.The differences in the results were mostly within calculateduncertainties.

6. Discussion

Exergy analysis is a systematic approach that localises thermo-dynamic losses and quantifies theoretical saving potential. It makespossible to compare the magnitude of different types of losses. Theresults presented above can be used directly to improve the processat this platform as discussed in Section 6.1, or they can be used forevaluation of performance as discussed in Section 6.2.

6.1. Reduction of exergy destruction

The exergy destruction map in Fig. 2 raises some importantissues.

Measures should be done to increase compressor efficienciesand to reduce the anti-surge recycling, for instance by modifyingthe existing compressors or by replacing them with new com-pressors. Higher efficiencies and design for flow rates relevant for

current and future production are of big importance. Althoughsome exergy destruction is unavoidable, Table 2 shows room forimprovement.

The high exergy destruction due to cooling of compressed gasand dissipation of cooling water (with temperature higher than T0)into the sea indicates that the potential to exploit this exergyshould be examined. The temperatures of the warm streams arerelatively low (between 70 and 150 �C, see Fig. 1) and the exergyrelated to this is therefore hard to utilize due to technical reasons.The demand for thermal energy is limited at this specific platform,but it should be possible to eliminate the electric heater in the fuelgas system by heat integration with one of the warmest com-pressed process streams. However, waste heat from the powerturbines is available at higher temperatures, andmight be preferredfor heating purposes.

Attention should be drawn to the exergy destruction due topressure reduction in the production manifold. The possibility toexploit this exergy in ejectors and expanders should be reviewed.Multiphase ejectors can use the exergy in high pressure wells toenhance recovery and flow rates in depleted wells [17]. Expandersfor multiphase flow are under development [18e22], and thismight be an application. Similarly, the exergy destructed in thevalves in the fuel gas system and in the separation train could havebeen recovered in expanders. In the fuel gas system, we have lowerpotential for savings, but it is easier since we have expansion of gas,which is a mature technology. For the separation train, it might beinteresting to assess the concept of separator turbines, which inaddition to production of power gives separation with morecompact equipment [18,19].

The suggestions above are of different nature; updating existingprocess units, modifying the configuration of the process anddeveloping new process units. The benefits of the measures pro-posed are dependent on the remaining lifetime of the installation,and will always be evaluated against investment costs. The feasi-bility of the suggested measures will also be influenced by thegeneral demand for simple and compact processes on offshoreinstallations.

6.2. Performance parameters and exergy analysis in the oil industry

In this study we have calculated the exergetic efficiency of theoile and gas processing of an existing offshore platform. Thisparameter can add to the industry’s own measures of performance,like the specific CO2 emissions. The exergetic efficiency can be usedto both quantify and justify best practices. It can also be used by thepublic sector to set standards for performance, that all should adhereto. Such standards may eventually lead to developments of moreenergy efficient technologies and to the best operation of these.

Under the introduction of other power sources for offshoreplatforms, for instance electric power from land or power fromoffshore windmills [23], specific CO2 emissions are no longer pro-portional to specific power consumption. Tax on specific CO2emissions encourages the use of renewable power sources, butonce such a source is taken into use, this parameter does not sayanything about the performance of the process anymore. Specificpower consumption will always evaluate the power demand andexergetic efficiency will always evaluate the management of theprocess with given boundary conditions.

We have so far only looked at the oil and gas processing at oneplatform. We propose that more platforms should be analysed, toexplore the applicability of exergy analysis when comparingdifferent platforms. We also propose that one platform should bemonitored over time, to see how exergy analysis can be used toevaluate efforts on adapting to changing process conditions or onincreasing the process efficiency.

Table A.8Molecular weight,M, normal boiling point, Tb, and ideal liquid density, rid.liq., for thehypothetical components used to describe the heavy oil fractions.

Name M, g/mol Tb, �C rid.liq., kg/m3

HypoA-1 81 73 721.2HypoA-2 108 99 740.1HypoA-3 125 152 774.6HypoA-4 171 230 817.1HypoA-5 247 316 859.3HypoA-6 388 437 906.2HypoA-7 640 618 988.5

Table A.9Composition of the three phases of the reservoir fluids, given as molar fractions.

Component Gas Liquid Water

CO2 9.18e-3 8.61e-3 0Methane 0.83 0.78 0Ethane 6.81e-2 6.41e-2 0Propane 3.74e-2 3.55e-2 0i-Butane 5.71e-3 5.52e-3 0n-Butane 1.34e-2 1.30e-2 0i-Pentane 4.28e-3 4.39e-3 0n-Pentane 5.51e-3 5.80e-3 0H2O 0 0 1N2 9.18e-3 8.61e-3 0HypoA-1 9.07e-3 1.34e-2 0HypoA-2 3.47e-3 1.17e-2 0HypoA-3 7.14e-4 1.49e-2 0HypoA-4 0 1.24e-2 0HypoA-5 0 9.01e-3 0HypoA-6 0 5.22e-3 0HypoA-7 0 3.44e-3 0

M. Voldsund et al. / Energy 55 (2013) 716e727724

7. Conclusion

An exergy analysis has been performed on the oil and gas pro-cessing on a North Sea oil platform for a real production day. Theday was simulated using measured process data. The magnitude ofchanges in different types of exergy was quantified. It was shownthat physical exergy in the material streams are mainly due to highpressures, and most exergy destruction take place in processes thatincrease pressure (compressors and cooling in the compressiontrains) or decrease pressure (in pressure reduction valves andrecycling). It was also shown that at this particular platform a largepart of the separation work is done in the recompression andreinjection trains (20 and 190 kW), not only in the separation train(50 kW). The overall process consumes 23.8 MW electric power.The sub-processes with most destructed exergy are the productionmanifold (4600 kW), the recompression train (4150 kW) and thereinjection trains (10,400 kW). The specific power consumption is179 kWh/Sm3 and the exergetic efficiency is 0.13. We proposemeasures to decrease exergy destruction, and thus increase theexergetic efficiency, of the process. We show the usefulness ofperforming exergy analysis on offshore platforms, and propose thatthis should be taken into regular use by the oil and gas industry.

Acknowledgements

The motivation from Statoil’s new-idea project of reducing CO2emissions from offshore oil and gas platforms is essential to thisstudy. Kirstin Hosaas, Laila Anita Gangstad, Alf Gunnar Edwardsenand Hanne Lied Larsen from Statoil are thanked for help with issuesregarding the process flowsheet. Tuong-Van Nguyen is thanked forvaluable discussions. The Faculty of Natural Sciences and Technol-ogy at the Norwegian University of Science and Technology isacknowledged for financial support.

Appendix A. Process flowsheet

The simulation of the process flowsheet of the oile and gasprocessing at the studied platform is explained in detail in thisappendix. Fig. 1 in Section 3 showed the simulated process flow-sheet. Details regarding the simulation are listed in Appendix A.1,while input data are given in Appendix A.2 and the validity of thesimulation is discussed in Appendix A.3. The reported uncertaintiesfor measured process variables are the 95% confidence intervals forthe measurements, determined as described in Section 4.3.

Appendix A.1. Simulation details

When simulating the process flowsheet, the following simpli-fications and manipulations were done:

� In the real process there are two separators in the 1st separationstage; one normal and one test separator (both are continuouslyin use). For simplicity they were merged into one separator.

� The HYSYS separators overpredicted the separation of waterand oil in the separation train, so in the simulation a part of theseparated water from each of the separators was split off andadded to the oil stream, to correct for this.

� In the drain system small amounts of liquid from knock outdrums in the flare system and from scrubbers with low liquidflow rates are collected in a reclaimed oil sump.When the liquidhere reaches a certain level, it is pumped to the 2nd separationstage. For simplicity this was simulated as a small pump thatcontinuously pumps liquid from the scrubbers to the 2nd sep-aration stage. Since we study a normal production day with astable production, we neglected liquid from the flare system.

Appendix A.2. Input data

The feed of the system is reservoir fluids, and hypotheticalcomponents were used to simulate these fluids. The properties ofthe hypothetical components are given in Table A.8. The composi-tion of each phase in the reservoir fluids are given in Table A.9. Thisdescription of the reservoir fluids is developed by the oil company.

The reservoir fluids enter the process from thewells through theproduction manifold. The flow rates of oil, gas and water in eachwell stream were initially set as the allocated flow rates calculatedby the oil company, given in Table A.10. After all the input param-eters of the rest of the process were set, these flow rates werescaled so that the simulated flow rates of export oil, injection gasand produced water fitted with measured flow rates for thesestreams, given in Table A.11. The values of these measured flowrates are more exact than the allocated flow rates.

Temperatures, pressures and flow rates that were set asmeasured throughout the process are given in Tables A.12eA.14. Inthe reinjection trains, the total gas flow rate is determined by themeasured gas injection rate. Flow rates are in addition measuredseveral places through each of the injection trains, and the flow rateof each train is set to make the simulated flow rates as close to all ofthe measured flow rates as possible (least squares). Measured andsimulated flow rates in this sub-process are given in Table A.15. Inthe export pumping section and fuel gas system, not enough pro-cess variables were measured, so the efficiencies of the exportpumps were found from the performance curves of the pumps, andthe pressure drop over the fuel gas cooler was taken from the coolerdata sheet, see Table A.16. Efficiencies of the small pump in thedrain system and the water pump were set to the assumed value75%, and pressure drops over all separators where this was notgiven by measured pressures were set to 0 kPa.

Table A.10Allocated flow rates of gas, oil and water for each well stream for the studied pro-duction day.

Well Gas 103 Sm3/h Oil Sm3/h Water m3/h

7 57.6 20.6 13.816 87.5 27.2 1.523 80.5 21.1 13.924 81.9 40.1 1.926 71.3 23.5 5.4

Table A.11Measured flow rates in process streams leaving the platform.

Produced fluid Unit Flow rate

Export oil Sm3/h 132.5 � 0.4Injection gas 103 Sm3/h 369 � 17Produced water Sm3/h 67 � 5

Table A.12Measured temperatures for the studied production day. Where values otherthanmeasured values are used in the simulated process flowsheet, for instancebecause measured values were not available for the studied day, this is indi-cated in footnotes.

Description Temperature, �C

Production manifoldFrom well 7, valve, in 85.8 � 1.0From well 16, valve, in 84.7 � 1.0From well 23, valve, in 87.1 � 1.0From well 24, valve, in 81.0 � 1.0From well 26, valve, in 79.6 � 1.0From well 7, after heat loss 76.6 � 1.0From well 16, after heat loss 75.6 � 1.0From well 23, after heat loss 71.3 � 1.0From well 24, after heat loss 76.8 � 1.0From well 26, after heat loss 74.3 � 1.0Separation trainGas from 1st separator, after heat loss 73.6 � 1.0a

Gas from 2nd separator, after heat loss 59.2 � 1.0Gas from 3rd separator, after heat loss 46.9 � 1.0Recompression train1st cooler, out 39.9 � 1.01st compressor, out 104.9 � 1.02nd cooler, out 21.0 � 1.02nd compressor, out 111.8 � 1.03rd cooler, out 24.0 � 1.03rd compressor, out 146.5 � 1.0Export pumping2nd pump, in 48.1 � 1.0Reinjection, Train A1st cooler, out 28.0b

1st compressor, out 94.0 � 1.02nd cooler, out 28.0 � 1.02nd compressor, out 77.1 � 1.0Reinjection, Train B1st cooler, out 28.0 � 1.01st compressor, out 95.6 � 1.02nd cooler, out 28.0 � 1.02nd compressor, out 74.4 � 1.0Reinjection, Train C1st cooler, out 30.0c

1st compressor, out 93.4 � 1.02nd cooler, out 30.0d

2nd compressor, out 80.7 � 1.0Fuel gas system1st scrubber, gas out 35.0 � 1.02nd scrubber, in 63.0 � 1.0

a The weighted mean based on mass flow rate for the two separators that inthe simulated flowsheet is merged into one.

b This temperature is not measured for the studied production day, so theset point of the cooler is used.

c This temperature is not measured and the set point for the cooler is notknown for the studied production day, so the set point for the cooler a fewweeksearlier is used.

d This temperature is not measured and the set point for the cooler is notknown for the studied production day, so the set point for the cooler a fewweeks earlier is used.

Table A.13Measured pressures for the studied production day. Where valuesother than measured values are used in the simulated process flow-sheet, for instance because measured values were not available for thestudied day, this is indicated in footnotes.

Description Pressure, bar

Production manifoldFrom well 7, valve, in 130.1 � 1.3From well 16, valve, in 113.0 � 1.1From well 23, valve, in 165.1 � 1.7From well 24, valve, in 87.6 � 0.9From well 26, valve, in 88.8 � 0.9From well 7, valve, out 73.0 � 0.7From well 16, valve, out 73.0 � 0.7From well 23, valve, out 73.1 � 0.7From well 24, valve, out 72.7 � 0.7From well 26, valve, out 72.3 � 0.7Separation train1st separator, in 70.4 � 0.7a

2nd separator, in 8.50 � 0.083rd separator, in 2.80 � 0.03Water pump, out 8.77 � 0.09Recompression train1st compressor, in 2.41 � 0.021st compressor, out 5.72 � 0.062nd compressor, in 5.20 � 0.052nd compressor, out 18.75 � 0.193rd compressor, in 18.29 � 0.183rd compressor, out 70.0 � 0.7Export pumping1st pump, out 13.30 � 0.132nd pump, in 12.81 � 0.132nd pump, out 32.1 � 0.3Reinjection, Train A1st compressor, in 68.8 � 0.71st compressor, out 137.4 � 1.42nd compressor, in 137.5 � 1.4b

2nd compressor, out 236 � 2Reinjection, Train B1st compressor, in 68.9 � 0.71st compressor, out 139.8 � 1.42nd compressor, in 139.1 � 1.42nd compressor, out 236 � 2Reinjection, Train C1st compressor, in 66.1 � 0.71st compressor, out 131.9 � 1.32nd compressor, in 129.2 � 1.32nd compressor, out 236 � 2Fuel gas system1st scrubber, in 38.8 � 0.42nd scrubber, in 38.4 � 0.42nd scrubber, gas out 38.0 � 0.4To flare 9.30 � 0.09To turbine 18.25 � 0.18Drain systemDrain pump, out 8.52 � 0.09c

a The weighted mean based on mass flow rate for the measuredvalues in the gas flow from the two separators that in the simulatedflowsheet is merged into one.

b This pressure was measured to 137.5 � 1.4 bar, but can not behigher than the pressure out from the 1st separator, so in the simu-lation it is instead set to 137.4 bar.

c This is the pressure measured in the most recent pumping period.

M. Voldsund et al. / Energy 55 (2013) 716e727 725

Table A.14Measured flow rates set in the simulated process flowsheet for the studied pro-duction day.

Description Unit Flow rate

Separation trainWater from 1st separator Sm3/h 54 � 5Water from 2nd separator m3/h 12.6 � 1.3Water pump, out m3/h 0.53 � 0.05Recompression train1st compressor, in m3/h 7100 � 7002nd compressor, in m3/h 5800 � 6003rd compressor, in m3/h 1560 � 160Export pumping section1st pump, out m3/h 230 � 202nd pump, out m3/h 176 � 18Fuel gas systemTo flares Sm3/h 335 � 14a

To power turbines Sm3/h 9630 � 170

a Sum of pilot flame for high pressure and low pressure flare, where it is assumedthat half of the measured flow rate for low pressure flare is for pilot flame while therest is from other places in the system, and is negligible these places.

Table A.15Measured and simulated flow rates of gas for the studied production day in thereinjection trains.

Description Measured flow rate, m3/h Simulated flowrate, m3/h

Train A1st compressor, in 1210 � 120 11401st compressor, out 750 � 70 7502nd compressor, in 510 � 50 503Train B1st compressor, in 1300 � 130 12131st compressor, out 770 � 80 7902nd compressor, in 530 � 50 529Train C1st compressor, in 2400 � 200 23482nd compressor, in 1040 � 100 1059

Table A.16Efficiencies, h, of pumps in the export pumping section found from the performancecurves of the pumps, and pressure drop, DP, of cooler found from cooler data sheet.

Process unit Variable Value

Booster export pump h, % 55Main export pump h, % 48Fuel gas cooler DP, kPa 50

Table A.17Measured and simulated flow rates of gas for the studied production day in theseparation train. The uncertainties for the two last gas flows in the separation trainare not known, because the flow rates are lower than what the flowmeters aredesigned for.

Description Measured flowrate, ton/h

Simulated flowrate, ton/h

Gas from 1st separator 320 � 30a 318Gas from 2nd separator 8.1 10.4Gas from 3rd separator 2.2 2.2

a The sum of the gas flow from the two separators that in the simulated flowsheetis merged into one.

Table A.18Measured and simulated pressure for the studied production day in the separationtrain.

Description Measuredpressure, bar

Simulatedpressure, bar

Oil from electrostatic coalescer 4.25 � 0.04 2.80

Table A.19Measured power consumption in compression trains and sum of simulated enthalpychange, DH, over the compressors for each train.

Compressor train Measured powerconsumption, kW

Sum of simulatedDH, kW

Recompression 5200 � 100 4703Reinjection A 5550 � 110 4781Reinjection B 5940 � 120 5008Reinjection C 9800 � 200 8847

M. Voldsund et al. / Energy 55 (2013) 716e727726

Appendix A.3. Validation

Measured flow rates are compared with simulated flow rates inTables A.15 and A.17. The simulated flow rates were within the

uncertainty of the measured flow rates, when the uncertainty wasknown.

A measured pressure in the separation trainwas compared witha simulated pressure in Table A.18. The deviation between thesenumbers is due to the fact that height differences are not includedin the simulation, as discussed in Section 5.5. The pressure differ-ence corresponds to a height difference of 17 m within the sepa-ration train. This is however of minor importance for the overallresults.

In Table A.19 the measured power consumption of eachcompression train is compared with the summed enthalpy change,DH, over the compressors in each train. The differences betweenthe power consumption and the enthalpy changes are electric andmechanical losses. The numbers indicate that 84e90% of the powerconsumed in each train end up in the process streams, and this isconsidered realistic.

References

[1] Statistics Norway, Greenhouse gases, by source, energy product and pollutant(1980e2011), http://www.ssb.no/emner/01/04/10/klimagassn/ [November2012].

[2] Nordvik FM, Verlo SB, Zenker E, editors. Facts 2011: the Norwegian petroleumsector, Ministry of petroleum and energy and Norwegian petroleum Direc-torate 2010.

[3] Helgesen OK. Klimameldingen kan gi mindre utvinning. Teknisk Ukeblad2012;16:14.

[4] de Oliveira Junior S, van Hombeeck M. Exergy analysis of petroleum separa-tion processes in offshore platforms. Energy Conversion and Management1997;38(15e17):1577e84.

[5] Voldsund M, Ertesvåg IS, Røsjorde A, HeW, Kjelstrup S. Exergy analysis of the oiland gas separation processes on a north sea oil platform. In: Favrat D, editor.Proceedings of ECOS 2010: 23rd international conference on efficiency, cost,optimization, simulation, and environmental impact of energy systems 2010.

[6] Voldsund M, He W, Røsjorde A, Ertesvåg IS, Kjelstrup S. Evaluation of the oiland gas processing at a real production day on a north sea oil platform usingexergy analysis. In: Desideri U, Manfrida G, Sciubba E, editors. Proceedings ofECOS 2012: 25th international conference on efficiency, cost, optimization,simulation, and environmental impact of energy systems 2012.

[7] Kotas TJ. The exergy method of thermal plant analysis. 2nd ed. Krieger Pub-lishing Company; 1995.

[8] Moran MJ, Shapiro HN. Fundamentals of engineering thermodynamics. 5th ed.USA: John Wiley & Sons, Inc; 2004.

[9] Dincer I, Rosen MA. Exergy: energy, environment and sustainable develop-ment. 1st ed. Oxford, UK and Burlington, USA: Elsevier; 2007.

[10] Peng D-Y, Robinson DB. A new two-constant equation of state. Industrial &Engineering Chemistry Fundamentals 1976;15:59e64.

[11] Aspen HYSYS simulation basis guide. Burlington: Aspen Technology, Inc.;March 2011.

[12] Norwegian Meteorological Institute. Mean temperature e ordered report.http://eklima.met.no; October 2012.

[13] Abdollahi-Demneh F, Moosavian MA, Omidkhah MR, Bahmanyar H. Calcu-lating exergy in flowsheeting simulators: a hysys implementation. Energy1991;20:1061e155.

[14] Taylor JR. An introduction to error analysis. 2nd ed. Sausalito, California:University Science Books; 1997.

[15] Setzmann U, Wagner W. A new equation of state and tables of thermody-namic properties for methane covering the range from the melting line to625 k at pressures up to 1000 mpa. Journal of Physical and Chemical Refer-ence Data 2011;36:5320e7.

M. Voldsund et al. / Energy 55 (2013) 716e727 727

[16] Margarone M, Magi S, Gorla G, Biffi S, Siboni P, Valenti G, et al. Revamping,energy efficiency, and exergy analysis of an existing upstream gas treatmentfacility. Journal of Energy Resources Technology 2011;133:012001-1e012001-9.

[17] Andreussi P, Sodini S, Faluomi V, Ciandri P, Ansiati A, Paone F. Multiphaseejector to boost production: first application in the gulf of mexico. In: Pro-ceedings of offshore technology conference, 5 Maye8 May 2003, Houston,Texas; 2003.

[18] Rawlins C, Ross G. Design and analysis of a multiphase turbine for compactgas-liquid separation. SPE Production & Facilities 2002;17:47e52.

[19] Oxley K, Bennett J, Fremin L, Taylor J, Ross G. Rst’s mission to mars e thefirst commercial application of rotary separator turbine technology. In:

Proceedings of offshore technology conference, 5 Maye8 May 2003, Houston,Texas; 2003.

[20] Charron Y, Pagnier P, Marchetta E, Stihle S. Multiphase flow helico-axialturbine: applications and performance. In: Proceedings of the 11th AbuDhabi international petroleum exhibition and conference; 2004.

[21] Falcimaigne J, Decarre S. Multiphase Production: pipeline transport, pumpingand metering.

[22] Patel VP, KimmelHE. Fifteen years offield experience in lngexpander technology.In: Proceedings of the first middle east turbomachinery symposium; 2011.

[23] He W, Jacobsen G, Anderson T, Olsen F, Hanson TD, Korpås M, et al. The po-tential of integrating wind power with offshore oil and gas platforms. WindEngineering 2010;34(2):125e38.