Flexible operating characteristics and versatility of services provided by pumped-storage hydropower technologies make them ideally suited to support large-scale integration of variable renewable resources and improve the reliability of power systems, a recent study concludes.
By Vladimir Koritarov, Tao Guo, Erik Ela, Bruno Trouille, James Feltes and Michael Reed
Vladimir Koritarov is deputy director of the Center for Energy, Environmental and Economic Systems Analysis at Argonne National Laboratory. Tao Guo is regional director of west coast with Energy Exemplar LLC. Erik Ela is senior engineer with the National Renewable Energy Laboratory. Bruno Trouille is vice president of the international projects group at MWH Americas. James Feltes is senior manager in the consulting department at Siemens PTI. Michael Reed is program manager/chief engineer in the Wind and Water Power Technologies Office of the Office of Energy Efficiency and Renewable Energy at the U.S. Department of Energy.
This article has been evaluated and edited in accordance with reviews conducted by two or more professionals who have relevant expertise. These peer reviewers judge manuscripts for technical accuracy, usefulness, and overall importance within the hydroelectric industry.
A study was recently undertaken to determine the role and value of advanced pumped-storage hydropower (PSH) in the U.S.1 Work involved developing detailed simulation models of advanced PSH technologies (adjustable speed, or AS, and ternary units) in order to analyze their technical capabilities to provide grid services and to assess the value of these services under different market structures and for different levels of renewable generation resources integrated in the power system. Compared with fixed-speed (FS) PSH units, AS and ternary units provide greater operating flexibility and efficiency but are more expensive to build due to additional equipment and space requirements. An especially important characteristic of AS and ternary technology is their ability to provide regulation service in the pumping mode of operation.
The study was funded by the U.S. Department of Energy’s (DOE) Office of Energy Efficiency and Renewable Energy (EERE) through a program managed by EERE’s Wind and Water Power Technologies Office. The study was led by Argonne National Laboratory, and other project team members included Siemens PTI Inc., Energy Exemplar LLC, MWH Americas Inc. and the National Renewable Energy Laboratory (NREL).
The scope of work had two main components:
— Develop vendor-neutral dynamic simulation models for advanced PSH technologies; and
— Analyze production cost and revenue to assess the value of PSH in the power system.
The project team established task force groups (TFGs) to focus on specific aspects of the modeling and/or analysis. The advanced technology modeling TFG developed vendor-neutral models of advanced PSH technologies, which were then integrated into Siemens PTI’s PSS®E software and tested using the standard PSS®E test cases, as well as the dynamic PSS®E cases for the Western Interconnection developed by the Western Electricity Coordinating Council (WECC). The new models were added to the PSS®E library. In addition, the vendor-neutral models are available for integration into other software packages.
The simulations addressed a range of power system operational issues and time frames (see Figure 1 on page 52). The analysis aimed to capture PSH behavior and operational characteristics across different time scales, from a fraction of a second for dynamic responses to annual simulations for production cost runs. The team used four computer models (PSS®E, FESTIV, CHEERS, and PLEXOS) to simulate system operation and analyze operational issues occurring on different time scales.
Both cost- and market-based approaches were applied. The cost-based approach allows for the evaluation of benefits PSH plants provide to the power system and is typically applied for PSH projects operating in traditional vertically integrated utilities. The market-based approach allows for the calculation of revenues a PSH project can realize in a restructured electricity market, where the plant competes to provide energy and ancillary services. The main distinction between the approaches is that in the cost-based approach the value of a project is measured by the overall benefits it provides to the system, while the market-based approach focuses on providing data for the analysis of the financial viability of the project in a competitive market.
The simulations of system operations were performed for a future year largely based on WECC’s long-term projections for 2022. WECC’s Transmission Expansion Planning Policy Committee 2022 Common Case served as the foundation for building modeling cases and scenarios, but certain parameters and data varied depending on the scenario assumptions. Simulations of power system operations were performed for two levels of renewable energy penetration:
1. Baseline Scenario – Corresponding to mandated renewable portfolio standard levels of renewable energy generation, amounting to about 14% of total generation within the U.S. part of the Western Interconnection in 2022; and
2. High Wind Scenario – Corresponding to the high wind scenario from the Western Wind and Solar Integration Study – Phase 2,2 amounting to about 33% of renewable energy generation within the U.S. part of the interconnection in 2022.
Summary of key findings
The study involved numerous simulations and model runs across various time scales. The key findings and conclusions are summarized below.
Advanced technology modeling
Dynamic models for AS and ternary units were developed as vendor-neutral models and are described.3,4,5,6,7
The project team used these dynamic models to conduct power system performance studies and analyze dynamic behavior of these technologies and their impact on the power system. Also, analyses of FS and AS PSH technologies and their dynamic responses were studied for various system disturbances, including over- and under-frequency events due to sudden loss of load or generation in the system, as well as to changes in the power generated by variable renewable energy sources. Compared to FS PSH plants, the advanced PSH technologies provide greater flexibility and faster response to system disturbances.
Testing demonstrated that the new dynamic models can be used for typical dynamic simulation analyses required by transmission planning and interconnection studies. The tests also demonstrated the new capabilities, such as the use of AS and ternary PSH plants to provide regulation service in pumping mode. For all scenarios and disturbances, the new models showed expected performance and allowed demonstration of the expected advantages of the advanced PSH technology, specifically the capability of AS and ternary pumps to participate in secondary frequency control.
Production cost simulations using PLEXOS
Energy Exemplar’s PLEXOS model was used to perform production cost and revenue simulations for the base and high wind renewable energy scenarios, with and without FS and AS PSH plants modeled in the system. The day-ahead simulations were performed on an hourly basis for 2022 for all cases. Higher-resolution PLEXOS three-stage simulations with a five-minute time step were performed in each case for four typical weeks in 2022, i.e., the third week in January, April, July and October.
The analysis focused on three areas: Western Interconnection, California and Sacramento Municipal Utility District (SMUD). In the WECC Transmission Expansion Planning Policy Committee database,8 the SMUD load region represents the Balancing Authority of Northern California (BANC).
Both cost- and market-based ap-proaches were used in the analysis. While the cost-based approach was applied for simulation of the entire Western Interconnection and for the SMUD footprint, a market-based approach (as a bid-based electricity market) was applied for simulation of the California footprint.
Annual simulation results
The following sections present some of the key results obtained from the annual PLEXOS simulations of the Western Interconnection, California and SMUD for three cases, all run for the base and high wind scenarios:
— Without PSH plants;
— With the existing FS PSH plants; and
— With the existing FS and additional AS PSH plants.
Production cost savings
Table 1 on page 55 summarizes the savings in total system production cost in 2022 that can be attributed to PSH capacity and demonstrates that these savings are greater for higher penetration of renewable energy resources in the system (high wind scenario).
Simulation results for the Western Interconnection show that the existing FS PSH plants reduce total system operating cost in 2022 by about 1.1% (about $167 million) under the base scenario or about 2% (about $248 million) under the high wind scenario. The addition of three proposed AS PSH plants — Eagle Mountain, Iowa Hill and Swan Lake North — could reduce total production cost by an additional 1%, or $144 million, under the base scenario and by an additional 1.8%, or $229 million, under the high wind scenario.
Percentagewise, even larger cost savings could be achieved in California, where FS and AS PSH capacity reduces total system operating costs by 3.4%, or $171 million, under the base scenario and by a total of 9.1%, or $376 million, under the high wind scenario.
Results for the SMUD area show that the addition of the Iowa Hill AS PSH plant could result in annual production cost savings of about $23 million, or 8.6% of the total SMUD production cost, under the base scenario, and in savings of about $51 million, or 16.45%, under the high wind scenario.
PLEXOS simulations of the California system were performed using the market-based approach, which allows for detailed analysis of the value of energy arbitrage based on the locational marginal price (LMP) of electricity in each hour. PLEXOS simulations were performed using the co-optimization of energy and ancillary services, so the results for energy arbitrage with ancillary services are likely different from the results obtained if the PSH operations were optimized to maximize the energy arbitrage revenues only. Table 2 on page 56 presents key PLEXOS results for the base and high wind scenarios.
The high penetration of variable energy resources under the high wind scenario keeps average LMPs low and even negative when there are curtailments of excess variable generation. The cost of pumping energy for FS PSH plants under the high wind scenario is negative because the pumping energy is mostly supplied by the wind generation that would have been curtailed. Table 2 also shows that the capacity of existing FS PSH plants would not be sufficient for the high level of renewable resources in the system. With the addition of AS PSH plants, the overall pumping cost under the high wind scenario becomes positive, but its relatively low value indicates that the PSH pumping energy is still mostly comprised of the wind generation that would have been curtailed.
Table 2 also shows that, under the high wind scenario, the addition of AS PSH plants increases the total annual net revenues from energy arbitrage; however, the net revenues per kilowatt of PSH capacity are smaller because of the much larger PSH capacity in the system.
Given that the combined capacity of FS and AS PSH plants represents less than 3% of the total Western Interconnection system capacity in 2022, it can be observed that PSH plants provide a significant amount of operating reserves to the system, especially in cases when both FS and AS PSH plants are operating. Also, PSH contributions to operating reserves increase significantly with the addition of AS PSH plants to the system.
An especially large increase is observed for the regulation down and flexibility down reserves because the AS PSH can provide these services in the pumping mode of operation as well.
With regard to the monetary value of PSH contributions to operating reserves, PLEXOS simulations for California were performed using a market-based approach, which allowed for individual pricing and revenue analysis of ancillary services. Table 3 on page 58 provides a summary of PSH total annual revenues attributable to their contributions of operating reserves in 2022. The results show that the highest annual revenues are from the provision of regulation down service.
Integration of variable energy resources
PSH plants enable larger penetration of variable energy resources (VER) in the power system by providing a large quantity of flexible capacity that can be used to compensate for the variability and uncertainty of VER generation. In addition, the operating characteristics of PSH plants, which have quick ramping capabilities and can provide large quantities of operating reserves, make them ideally suited to support VER generation.
PLEXOS simulation results for the Western Interconnection under the base scenario show that the FS PSH plants reduce curtailments of VER generation by 565 GWh, or about 29% of total curtailments if there were no PSH plants in the system. With both FS and AS PSH plants operating in the Western Interconnection system, the curtailments are reduced by 958 GWh, or about 50% of total curtailments. The amount of curtailed VER generation under the high wind scenario amounts to 56,885 GWh in the case without PSH plants in the system. The FS PSH plants reduce this curtailment by 8,482 GWh, or 15%, while when both FS and AS PSH plants are operating in the system, the curtailments are reduced by 12,675 GWh, or 22%. Assuming a 30% capacity factor, the savings of 12,675 GWh roughly correspond to an average annual generation of almost 5,000 MW of wind capacity.
In California, under the base scenario, curtailments of VER generation are reduced from 155 GWh in the case without PSH plants to 46 GWh (70%) if FS PSH are in the system, and to 14 GWh (91%) if both FS and AS PSH are operating. Under the high wind scenario, the curtailments are reduced from 618 GWh in the case without PSH plants to 380 GWh (39%) if FS PSH plants are operating in the system, to 275 GWh (55%) if both FS and AS PSH plants are operating.
The results for the SMUD footprint show that the addition of the Iowa Hill AS PSH plant reduces renewable energy curtailments from 19 GWh to 1 GWh (95%) under the high wind scenario. There were no curtailments of VER generation under the base scenario.
Reduced cycling of thermal generating units
The flexibility of PSH capacity, its fast ramping characteristics and its load-leveling operation create a flatter net load profile for thermal units, which allows them to operate in a steadier mode, thus reducing the need for their ramping and frequent startups and shutdowns.
— Reduced startup costs. Because startups and shutdowns of thermal units involve substantial operating costs, as well as increased wear and tear on the units, a reduction in the number of startups provides for significant savings in system operating costs. PLEXOS results show that under both scenarios, the number of starts and startup costs of thermal generators are reduced substantially as more PSH capacity is introduced into the system.
If both FS and AS PSH plants are operating in the system, the annual thermal startup cost savings for the Western Interconnection are $44 million (about a 28.6% reduction in system startup costs) under the base scenario and $31 million (about 17.7% savings) under the high wind scenario.
In California, the savings in startup costs are similar under both scenarios and amount to about $10 million if only the existing FS PSH plants are operating in the system and to about $20 million if both FS and AS PSH plants are operating.
In the case of SMUD, the addition of the Iowa Hill AS PSH plant reduces annual startup costs by about $2 million under both scenarios. As a percent of total system startup costs in 2022, the cost savings ($2 million) represent about 45% of total startup costs under the base and about 42% under the high wind scenario.
— Reduced thermal generator ramping. PLEXOS simulations for the Western Interconnection in 2022, under the base scenario, show that FS PSH plants reduce the total ramp-up needs of thermal generators by 1,786 GW and ramp-down needs by 2,560 GW. These values represent aggregated ramping travel of all units in all hours of the year. If both FS and AS PSH plants are in the system, the ramp-up needs of thermal generators are reduced by 3,420 GW and ramp-down needs by 4,817 GW.
Similarly, the results for California in 2022, under the high wind scenario, show that FS PSH plants reduce the ramp-up and ramp-down needs of thermal generators by 531 GW and 945 GW, respectively. If both FS and AS PSH plants are in the system, the ramp-up and ramp-down needs of thermal generators are reduced by 1,214 GW and 1,943 GW, respectively.
In the case of SMUD, the proposed Iowa Hill AS PSH plant reduces ramp-up and ramp-down needs by 136 GW and 197 GW, respectively, under the base scenario, and by 119 GW and 174 GW, respectively, under the high wind scenario.
PSH impacts on power system emissions
Simulation results for the Western Interconnection show an increase in CO2, NOx, and SO2 emissions under the base scenario, but the operation of PSH plants decreases overall system emissions under the high wind scenario. This is primarily due to a higher percentage of wind energy that is available for PSH pumping and the PSH impacts on reducing the curtailments of wind energy, which offset the increased emissions of conventional thermal generating units.
Results for California show a decrease in CO2 and NOx emissions and an increase in SO2 emissions under both scenarios.
The most significant emission reductions are observed for the SMUD system. The introduction of the proposed Iowa Hill AS PSH plant reduces pollutant emissions in the SMUD system under both scenarios.
Three-stage DA-HA-RT simulation results
To capture the uncertainty of renewable energy forecasting and intra-hourly variability of VER, as well as to evaluate system needs for operating reserves and flexible ramping capacity, three-stage DA-HA-RT (Day Ahead – Hour Ahead – Real Time) sequential simulations with a five-minute time step in RT were performed for four typical weeks in different seasons of the year. Simulations were performed for the three system footprints for the third weeks in January, April, July, and October of 2022.
Table 4 on page 60 presents key results obtained from simulations for the three systems under the high wind scenario. SMUD plans the addition of the Iowa Hill AS PSH plant to its power system, so conventional FS PSH plants were not modeled in the simulations of the SMUD footprint.
The results of these high-resolution simulations show that the overall production cost savings due to operation of FS and AS PSH plants in the system amount to about 3.6% of the total production costs in the Western Interconnection, 7.3% in California and 14.3% in the SMUD system. Although these are average cost savings over the four typical weeks in 2022, the average annual values would be in a similar range. PLEXOS annual simulation runs using the hourly time step also provide similar results.
The impacts of PSH plants on reduction of startup and shutdown cost are also significant. The operation of FS and AS PSH plants in the system reduces overall startup and shutdown costs from about 11% in SMUD up to almost 42% in California.
Similarly, the operation of both FS and AS PSH plants reduces the need for ramping of thermal units. Over the four typical weeks in 2022, ramping up and down of thermal units decreased by about 22% to 25% in the Western Interconnection and SMUD areas. In California, the ramping down of thermal units decreasing by more than 60%. These results demonstrate that PSH plants can manage a significant number of ramping duties to counterbalance the intra-hourly variations in loads and variable renewable generation.
In the three-stage simulations, the results of RT simulations show higher operating costs and ramping needs than those of the DA simulations. This is because the RT simulations capture the intra-hourly variability of VER generation, which is not captured by DA simulations that use hourly time steps. The higher operating cost and ramping needs of thermal generators in RT simulations indicate that they require additional ramping to meet the sub-hourly variability and uncertainties of load and variable renewable generation.
Analysis of reliability and costs using the FESTIV model
NREL’s FESTIV model was used to analyze in high temporal detail how conventional and advanced PSH can reduce total system production costs and improve steady-state reliability. The FESTIV model was used to simulate BANC, where the SMUD system is located, for two time periods — one with highly volatile variable generation and relatively low load in April and one with reduced variable generation but significant load in July. In both time periods, use of a FS PSH plant reduced the total system production costs. With the addition of an AS PSH plant rather than the FS PSH plant, production costs were further reduced. These results mirror those obtained from PLEXOS simulations and the analysis of detailed power system operations at multiple time scales demonstrates that conventional and advanced PSH provide significant benefits to systems of this size by reducing production costs.
The FESTIV model was also used to evaluate the contributions of PSH plants to the reliability of power system operation. The simulations were performed using a four-second time step to model the real-time operation of the power system and calculate the area control error and energy imbalances. A case without a PSH plant operating in the system served as the reference case.
The results show that FS and AS PSH plants reduce the number of Control Performance Standard 2 (CPS2) violations and improve the CPS2 score in the April and July weeks of 2022, but the effects are more significant during the July week. The results for the July week also show improvements in the absolute amount of area control error and the standard deviation of ACE.
These results also illustrate the impacts of PSH provisions of regulation reserve on improving system reliability, which allows a balancing authority to better meet steady-state reliability standards. From Figure 2, it can be seen that AS PSH units frequently provide regulation service in the pumping mode and for this reason often pump with less than full capacity.
PSH plants provide a variety of benefits to the power system. In the past the benefits of PSH plants were usually associated only with the energy arbitrage and contingency reserves, but this study clearly shows these are just a fraction of the total value PSH plants provide to the system. Many of the PSH services and contributions are taken for granted, and for many of them there are no established mechanisms to provide revenues to PSH plants for providing those services or contributions to the power system.
The study shows that the value of PSH plants increases with higher penetration of VER in the system. In addition to enabling larger integration of VER technologies and reducing the curtailments of excess variable generation, PSH plants reduce overall system generation costs, provide flexibility and various operating reserves necessary for system operation, reduce cycling of thermal units and associated startup/shutdown and ramping costs, reduce transmission congestion, increase the reliability of system operation, and provide many other benefits. In addition, with a larger share of VER in the system, PSH plants tend to have a positive impact on system emissions, as a larger share of pumping energy is provided by VER generation.
Compared to the conventional FS PSH plants, the analyses showed that the advanced AS PSH technologies provide greater flexibility and faster response to system disturbances, allow for greater savings in overall system production costs, provide larger amounts of various operating reserves, and generally provide more value to the power system.
This work is supported by the U.S. Department of Energy’s Office of Energy Efficiency and Renewable Energy, Wind and Water Power Technologies Office, under contract DE-AC02-06CH11357.
1Modeling and Analysis of Value of Advanced Pumped Storage Hydropower in the United States, ANL/DIS-14/7, Argonne National Laboratory, Lemont, Ill, 2014.
2The Western Wind and Solar Integration Study Phase 2, NREL/TP-5500-55588, National Renewable Energy Laboratory, Golden, Colo., 2013.
3Review of Existing Hydroelectric Turbine-Governor Simulation Models, ANL/DIS-13/05, Argonne National Laboratory, Lemont, Ill., 2013.
4Modeling Adjustable Speed Pumped Storage Hydro Units Employing Doubly-Fed Induction Machines, ANL/DIS-13/06, Argonne National Laboratory, Lemont, Ill., 2013.
5Modeling Ternary Pumped Storage Units, ANL/DIS-13/07, Argonne National Laboratory, Lemont, Ill., 2013.
6Testing Dynamic Simulation Models for Different Types of Advanced Pumped Storage Hydro Units, ANL/DIS-13/08, Argonne National Laboratory, Lemont, Ill., 2013.
7Simulation of the Secondary Frequency Control Capability of the Advanced PSH Technology and its Application to the SMUD System, ANL/DIS-13/10, Argonne National Laboratory, Lemont, Ill., 2013.
Koritarov, Vladimir, et al., “Modeling and Simulation of Advanced Pumped-Storage Hydropower Technologies and their Contributions to the Power System,” Proceedings of HydroVision International 2014.