— ENERGY PORTFOLIO MAN AGEMENT

Hydro and pump storage modeling in

ABB Ability™ PROMOD®

Benchmark report

TABLE OF CONTENTS

Table of Contents

1 ABSTRACT .......................................................................................................................................... 1

2 INTRODUCTION ................................................................................................................................. 1 2.1 Hydro Modeling ................................................................................................................................... 1 2.2 Pump Storage Modeling ................................................................................................................... 3

3 ANALYSIS............................................................................................................................................ 4

4 CONCLUSION ..................................................................................................................................... 9

5 ABOUT THE AUTHOR ........................................................................................................................ 9

List of Figures

Figure 1 Hydro Dispatch Logic ........................................................................................................................... 3 Figure 2 PJM Hourly Hydro Generation Benchmark Comparison @ System Level ............................... 6 Figure 3 PJM Hourly Pump Storage Generation Benchmark Comparison @ System Level ................. 7 Figure 4 PJM Average Hourly Trading Hub LMP Benchmark Comparison............................................... 7 Figure 5 PJM Average LMP by Trading Hub {Average (e7 – Legacy) = 0.33 $/MWh}............................. 8 Figure 6 PJM Average LMP by Bus {Average (e7 – Legacy) = -0.54 $/MWh} ........................................... 8

List of Tables

Table 1 Benchmark Result at System Level .....................................................................................................5

ABSTRACT HYDRO & PUMP STOR AGE MODELING

9AKK107046A5365 -US 1

1 ABSTRACT The PROMOD Hydro & Pump Storage optimization modules provide the capability to determine

the optimum release schedules and storage strategies for systems of water based generating

units. These scheduling optimizations help with improving system operation and efficient

resource utilization. These modules have the ability to model the limitations of water using

energy availability limits and capacity levels while automatically bias the generating schedules to

peak load hours to reduce the load shape served by other units.

New enhancements have been developed in PROMOD on the e7 data management platform

which leverages cutting edge technology to enable faster development of new features and

reach high quality standards. The Introduction section of this report covers the foundation of

hydro and pump storage modeling capabilities that exist in both legacy and PROMOD on e7. The

Analysis section discusses the differences between these two versions for hydro and pump

storage modules. The benchmark results using MISO’s MTEP16, PJM RTEP, and SPP databases are

also included in the Analysis section. A summary of the modeling differences between legacy and

e7 are provided in the Conclusion section.

Future work will include further enhancements that will be implemented in the PROMOD on e7.

These enhancements will target more sophisticated business needs.

2 INTRODUCTION PROMOD analysis has been widely received with high acknowledgement by industry leaders and

experts over the last 41 years. For example, PROMOD analysis is a critical component of the PJM

Regional Transmission Expansion Process (RTEP). It drives the Market Efficiency RTEP planning

component. Also, MISO leverages PROMOD to provide MTEP reports not only for planning but

also for setting the rules to coordinate with their members.

To use the latest technology platform and address the changing needs of the modern customer,

PROMOD on e7 is developed utilizing Xpress Solver and Microsoft® SQL® databases for input and

output data along with providing web services and grid computing. It provides striking interface

capabilities such as in-application reporting, charting, formula tool (to create custom variables),

and flexible visualization of the resolved data or input/output together. It can perform

daily/weekly commitment with support for hourly or sub-hourly (30, 15, 10, 5 minute) dispatch. It

can also automatically execute a defined set of tasks to reduce repeatable steps, data input

errors, and the time to obtain valid results.

This report focuses on the hydro and pump storage modeling aspects of PROMOD and highlights

the differences between the legacy and e7. The following two subsections covers the detail

modeling for these two modules separately.

2.1 Hydro Modeling

Hydro power is used worldwide to provide relatively low marginal cost and low emission

electricity. Hydro power facilities are some of the most flexible types of generators in the power

sector. Flexibility is a key resource for the electric grid in balancing demand with generation

despite normal and unexpected variability in load or resource availability and from forecast

ABB ABILIT Y PROMOD INTRODUCTION

2 9AKK107046A5365 -US

errors. In general, large hydro power facilities have the capability to provide significant flexibility

to the electric grid because of their fast ramping capabilities and innate storage in the reservoirs

behind large dams.

In PROMOD, the hydro units are scheduled before the thermal units, subject to minimum and

maximum capacity constraints, energy limits, ramp rate constraints, and minimum up and down

time limits. Hydro units may be scheduled against either a location area, control area (pool), or

system load. To schedule the energy, PROMOD provides options to define hydro type as Run of

River or Peak Shave. For the Run of River option, the hydro units will be scheduled such that the

provided energy limit is spread uniformly. For Peak shave option, the hydro energy will be

allocated to cover the upper-most part of the load. Basically, the hydro unit will be scheduled

with a run of river portion (based on the min capacity) and the rest of the energy will be applied

to level the peak hourly load.

To adjust the hydro dispatch for economic purpose i.e. hydro-thermal co-optimization (HTC),

PROMOD provides an option to specify hydro unit response P factor. The multiplication of this

factor and the difference between the maximum and the minimum capacity will adjust the hydro

unit schedule based on the locational marginal price (reference a bus or hub) and the unit’s cost.

The hydro energy remains the same after HTC schedule.

PROMOD also allows the users to configure the weekly max energy limit beside the monthly max

energy limit. If the weekly max energy limit is set to zero, Proportional Load Following (PLF)

schedule will be performed according to the provided monthly max energy. The PLF logic

determines the load target based on the benefit to control area (pool) or bene-fit to location area

depending on the defined dispatch location selected by the user. This logic is used in calculating

the schedule based on an iterative approach. This will result in using less hydro energy during low

load periods compared to the energy used during the high load periods.

If the minimum capacity is set to zero, the minimum capacity percentage is multiplied by the

weekly energy limit (either provided by direct input or calculated from Monthly energy limit using

PLF) and used as input to the run of river component, and the rest of the weekly energy

contributes to the peak levelling (Peak Shave). The sum of the resulted sub period hydro schedule

will equal the weekly and monthly energy limits.

Figure 1 illustrates how hydro dispatch in PROMOD is modeled where:

– “PLF” stands for “Proportional Load Following”

– “R” stands for “Run of River”

– “PS” stands for “Peak Shave”

INTRODUCTION HYDRO & PUMP STOR AGE MODELING

9AKK107046A5365 -US 3

Figure 1 Hydro Dispatch Logic

2.2 Pump Storage Modeling

A pump storage unit consumes power to pump water into a reservoir during off-peak hours and

generates during on-peak hours based on the price signals. PROMOD provides flexibility to

dispatch these storage units against the locational marginal price (LMP). This dispatch takes

place after the preliminary dispatch of the thermal units in the system (end of the first iteration).

Preliminary scheduling is always done at the end of the first iteration of the unit dispatch to get

the price signals. During this step, unit operating constraints such as min runtime and min

downtime are factored into the resulting unit commitment. In each hour/sub-hour, for both

pumping and generating, the incremental cost is determined in terms of $/MW change in

generation, in a discrete function (LMP cost curve). User can define the number of steps to be

included in the LMP cost curve for pump storage optimization. For ex-ample with input of 4 LMP

Cost Curve Steps, for a pump storage generator with maximum capacity of 200MW, there will be

ABB ABILIT Y PROMOD ANALYSIS

4 9AKK107046A5365 -US

specific MW levels of -200, -150, -100, -50, 0, 50, 100, 150, and 200. If there are multiple units in

the same station mapped to the same power flow bus, those units will share the same LMP cost

curve. The MW levels will be determined by the sum of all those pump storage units; maximum

capacities divided by the number of MW levels.

The obtained pump storage schedule after the preliminary dispatch is then used in the following

iterations to adjust the corresponding load. The operating constraints for storage include a

requirement that the value of the energy displaced on-peak must off-set the cost of storing the

energy off-peak. PROMOD allows user to define a final energy storage schedule at the end of n-1

iteration against the corresponding LMP which corresponds to the most accurate locational

marginal price. In other words, the dispatch of pump storage units takes place before the final

thermal dispatch of remaining units in the system. This schedule is essentially “locked in” and

treated as a modification to the load for the thermal dispatch of all remaining units and

subsequent transactions.

In the pump storage optimization process, the following constraints will be enforced:

– The reservoir’s starting net level at midnight of the prescribed calendar day.

– Ending reservoir level is assumed to be the same as the starting level.

– Unit can have both daily refill cycle and weekly refill cycle. If the reservoir level at 6 AM is

entered, it is treated as daily refill cycle and the reservoir level at 6 AM must be enforced.

– The generating maximum and minimum capacities for each unit of the re-source.

– The minimum and maximum reservoir levels.

– The pumped storage reservoir may undergo water/energy inflow, ex-pressed in GWh per week

or month. This energy will be generated in the course of the week in which it arrives even if

the pumped storage station is not otherwise operated.

– The maximum weekly generation will not exceed the input value.

The objective of pump storage dispatch is to minimize the cost of meeting load while:

– In each hour, the pumping level must be no greater than the maximum pumping capacity and

no less than the minimum pumping capacity. Similarly, the generating level must be no

greater than the maximum generating capacity and no less than the minimum generating

capacity.

– At no time may the quantity of stored energy exceed the maximum or fall below the minimum.

– Between any two hours, the change in stored energy must be equal to the difference between

the energy input to the storage system and the energy output from it, accounting for the

round-trip input-output efficiency.

In the final dispatch for nodal transmission simulation, the pumped storage generation is

included as the fixed generation at a bus. Therefore, it will influence active constraints such as

flows, interfaces, and control area (pool) load balance. The pumped storage pumping is factored

in as a fixed load at a bus, and will contribute to the underlying flows due to load.

3 ANALYSIS This section covers the modeling differences between the legacy and e7 for hydro and pump

storage modules. To identify these differences, the benchmark results using MISO’s MTEP, PJM

ANALYSIS HYDRO & PUMP STOR AGE MODELING

9AKK107046A5365 -US 5

RTEP, and SPP databases are examined in terms of hydro and pump storage generations. For

these comparisons, the third week of July (peak week) forecast results are used. The 2024

forecast for MISO, 2020 forecast for SPP, and 2019 forecast for PJM are used for these

comparison. Table 1 shows total energy comparisons between the legacy and e7 for these

databases in terms of total generation (hydro, pump storage, thermal, and price stations) and

total cost. The price stations cover renewables and transactions. The total hydro generations are

provided to highlight the databases with the highest difference.

Table 1 Benchmark Result at System Level

Since PJM hydro comparison shows the largest difference, the remaining of this section is

focused on PJM database. As shown in Figure 2, hydro units at the system level are running more

during peak hours in e7 except for the weekends (low load hours versus to the high load hours of

weekdays) comparing to the legacy PROMOD. Note that the X-axis shows the number of hours

from 1 to 168 to cover the whole week starting at Monday 7/15/2019. The Y-axis shows the hydro

generation in MW.

ABB ABILIT Y PROMOD ANALYSIS

6 9AKK107046A5365 -US

Figure 2 PJM Hourly Hydro Generation Benchmark Comparison @ System Level

Here are the main modeling differences between the e7 and the legacy for hydro dispatch:

1. In the legacy PROMOD, hydro units are dispatched from high capacity factor to low capacity

factor. In PROMOD on e7, they are dispatched from high capacity to low capacity.

2. In PROMOD on e7, Proportional Load Following logic works similar to the legacy PROMOD

with the exception that, instead of saving the hourly schedule for the month, it will sum up the

hydro energy for each week and use that weekly energy through the normal weekly logic to

come up with the hourly schedule for the week.

3. The cost of hydro unit is zero, unless the user input set as an adder ($/MWh) provided which

will be applied to the dispatch lambda of all committed minimum segments including hydro

units in the Legacy PROMOD. PROMOD on e7 also has this input but it is by unit. Note that for

this database, this user input was zero.

4. In the legacy PROMOD, there isn’t initial loading and ramp rate applied for the start of the

study (hour 1). In PROMOD on e7, the initial status determined based on generation from hour

1. If in hour 1, unit is off, hours off is min down time + 1. If in hour 1, unit is on, hour on is min

up time + 1. Note that the provided benchmark comparison is for the third week of the study

and this differences shouldn’t impact the result differences.

For pump storage generation differences, Figure 3 shows that e7 is running pump storage more

than the legacy. In terms of average LMP hub price difference between the e7 and the legacy for

this week, e7 price is 0.33$/MWh higher than the legacy obtained price.

ANALYSIS HYDRO & PUMP STOR AGE MODELING

9AKK107046A5365 -US 7

Figure 3 PJM Hourly Pump Storage Generation Benchmark Comparison @ System Level

Figure 4 shows the hourly locational marginal prices average over all the hubs. In average, the e7

LMP is 33.78$/MWh and the legacy LMP is 33.45$/MWh. Note that in general the e7 LMP for high

load hours is less than the legacy LMP. The X-axis shows the number of hours from 1 to 168 to

cover the whole week starting at Monday 7/15/2019. The Y-axis shows LMP in terms of $/MWh.

The LMP price differences by trading hub are provided in Figure 5. Figure 6 shows average LMP

differences at different bus. In average over all the buses, the e7 LMP is 34.28$/MWh and the

legacy e7 LMP is 34.82$/MWh.

Figure 4 PJM Average Hourly Trading Hub LMP Benchmark Comparison

ABB ABILIT Y PROMOD ANALYSIS

8 9AKK107046A5365 -US

Figure 5 PJM Average LMP by Trading Hub {Average (e7 – Legacy) = 0.33 $/MWh}

Figure 6 PJM Average LMP by Bus {Average (e7 – Legacy) = -0.54 $/MWh}

CONCLUSION HYDRO & PUMP STOR AGE MODELING

9AKK107046A5365 -US 9

The only difference between the e7 and the legacy PROMOD for pump storage dispatch is an

extra option provided for the user to do final pump storage scheduling at the end of n iteration

against the corresponding price. The legacy PROMOD will do final pump storage scheduling at

the end of n-1 iteration. Note that e7 has this option as well and the provided benchmark results

are based on scheduling at the end of n-1 iteration. The number of iteration for these studies is 3,

i.e., n = 3.

Since pump storage scheduling is a function of the obtained prices, these results need to be

examined further in terms of profits, flows and flowgate sensitivities for monitored flowgates

and the prices at key generator buses as well as for hubs. Further investigation is in progress and

will be reported in near future.

4 CONCLUSION This report provides the basic modeling algorithm for scheduling hydro and pump storage units

in PROMOD. To enable faster development of new features and reach high quality standards for

addressing the changing needs of the modern customer, the PROMOD is upgraded to PROMOD

on e7 utilizing Xpress Solver and Microsoft SQL databases for input and output data along

providing web services and grid computing capabilities.

To identify modeling differences between the legacy and PROMOD on e7, the benchmark results

using MISO’s MTEP16, PJM RTEP, and SPP databases for hydro and pump storage modeling are

provided. The benchmark results are examined in terms of total generation and cost. The cur-

rent results show a very comparable generation and cost at the system level for these different

databases between the e7 and the legacy. Hourly hydro schedule and pumped storage run more

especially during peak hours in e7 comparing to the legacy PROMOD. Also the resulting LMP

prices for high load hours from the e7 is lower compared to the output from the legacy.

Further investigation is in progress which examines profits, flows and flowgate sensitivities for

monitored flowgates and the prices at key generator buses while looking at the generation by

different unit types (other than hydro and pump storage units). Meanwhile, enhancements to

PROMOD on e7 will continue with the purpose of meeting more sophisticated business needs.

5 ABOUT THE AUTHOR Batool Nekooie, Ph.D., Product Manager, ABB Enterprise Software

Dr. Nekooie is a market analysis product manager at ABB Enterprise Software. She has held

various management and senior engineering positions in the energy industry during the last 25

years. She has worked with large regulated investor owned utilities as well as unregulated

merchant energy companies in the United States. Her research interests and experience includes

optimization, forecasting, and data analysis applied to utility system planning. Batool holds a

Doctorate degree in Electrical Engineering from the Georgia Institute of Technology.

NOTICE This document contains information about one or more ABB products and may include a description of or a

reference to one or more standards that may be generally relevant to the ABB products. The presence of any such

description of a standard or reference to a standard is not a representation that all of the ABB products referenced

in this document support all of the features of the described or referenced standard. In order to determine the

specific features supported by a particular ABB product, the reader should consult the product specifications for

the particular ABB product.

ABB may have one or more patents or pending patent applications protecting the intellectual property in the ABB

products described in this document.

The information in this document is subject to change without notice and should not be construed as a

commitment by ABB. ABB assumes no responsibility for any errors that may appear in this document.

Products described or referenced in this document are designed to be connected and to communicate information

and data through network interfaces, which should be connected to a secure net-work. It is the sole responsibility

of the system/product owner to provide and continuously ensure a secure connection between the product and

the system network and/or any other networks that may be connected.

The system/product owners must establish and maintain appropriate measures, including, but not limited to, the

installation of firewalls, application of authentication measures, encryption of data, installation of antivirus

programs, and so on, to protect these products, the network, its system, and interfaces against security breaches,

unauthorized access, interference, intrusion, leakage, and/or theft of data or information.

ABB performs functionality testing on the products and updates that we release. However, system/product owners

are ultimately responsible for ensuring that any product updates or other major system updates (to include but not

limited to code changes, configuration file changes, third-party software updates or patches, hardware change out,

and so on) are compatible with the security measures implemented. The system/ product owners must verify that

the system and associated products function as expected in the environment in which they are deployed.

In no event shall ABB be liable for direct, indirect, special, incidental or consequential damages of any nature or kind

arising from the use of this document, nor shall ABB be liable for incidental or consequential damages arising from

use of any software or hardware described in this document.

This document and parts thereof must not be reproduced or copied without written permission from ABB, and the

contents thereof must not be imparted to a third party nor used for any unauthorized purpose.

The software or hardware described in this document is furnished under a license and may be used, copied, or

disclosed only in accordance with the terms of such license. This product meets the requirements specified in EMC

Directive 2014/30/EU and in Low Voltage Directive 2014/35/EU.

TRADEMARKS PROMOD is a registered trademark of ABB. ABB Ability is a trademark of ABB.

Microsoft and SQL are registered trademarks of Microsoft Corporation in the US and other countries.

All rights to copyrights, registered trademarks, and trademarks reside with their respective owners.

Copyright © 2018 ABB.

All rights reserved.

Release: August 2018

Document Number: 9AKK107046A5365-US

—

Visit us

new.abb.com/enterprise-software/energy-portfolio/management

info.pges@abb.com

Do

cu

me

nt

Nu

mb

er:

9A

KK

107

04

6A

53

65

-US