TODAY’S STUDY: THE BENEFITS OF PUMPED HYDRO STORAGE CALCULATED

Quantifying the Operational Benefits of Conventional and Advanced Pumped Storage Hydro on Reliability and Efficiency

I. Krad, E. Ela, V. Koritarov, July 2014 (National Renewable Energy Laboratory)

Abstract

Pumped storage hydro (PSH) plants have significant potential in providing reliability and efficiency benefits in future electric power systems. New PSH technologies, like adjustable-speed PSH, have also been introduced and can present further benefits. An understanding of these benefits on systems with high penetrations of variable generation (VG) is a primary focus. This paper will demonstrate and quantify some of the reliability and efficiency benefits afforded by pumped storage hydro plants utilizing the Flexible Energy Scheduling Tool for Integrating Variable generation (FESTIV), an integrated power system operations tool which evaluates both reliability and production costs. A description about the FESTIV tool and how it simulates PSH operations at multiple timescales will be given. Impacts of PSH on area control error, production costs, and system operation are quantified on a high VG scenario in the Balancing Area of Northern California. We also perform a study on how advanced PSH can provide a fast form of regulation to improve reliability and potentially reduce costs.

click to enlarge

Introduction

The benefits of energy storage systems are desirable and well documented. They can help reduce production costs by providing power during expensive peak periods, while purchasing the power and storing it during cheap off-peak periods. They can provide numerous types of active power control support including contingency reserve, primary frequency control, automatic generation control, and load following. It can also provide benefits for reducing capacity needs, congestion management, and voltage and reactive power support. The response time, synchronization time, and ability to provide energy as both a generator and a load give energy storage unique qualities for both improving reliability and reducing production costs. Currently, the most common form of utility-scale energy storage is pumped storage hydro (PSH).

PSH first began gaining popularity in the 1970s in response to a sharp rise in natural gas and oil prices. In the USA, the Power Plant and Industrial Fuel Use Act was enacted which would limit the amount of oil and natural gas that can be consumed via new power plants [1]. The construction of new PSH plants was justified by comparing the net cost of a PSH plant and an equivalently sized fossil fuel plant [1-3]. This method disregarded the operational benefits that PSH can provide and failed to provide a level comparison. The financial justification of PSH was based on the potential for energy arbitrage. As a result, the allure of PSH has slowly diminished over the years due to falling oil and gas prices and improved thermal generator operating characteristics. However, by incentivizing and recognizing the other benefits afforded by PSH, namely their ability to aid in system reliability, there is potential for PSH to again return to the forefront of emerging grid technologies.

There has been some research performed in an attempt to better model PSH plants. The authors of [4] developed a mixed integer linear programming (MIP) model of PSH that considers the operating characteristics of PSH such as ramp transition constraints and pumped-storage operating mode constraints. They also introduce a method to model the head effect through approximations to capture the relationship between power, volume of the water, and the flow of the water. The authors of [5] developed an aggregate hydro plant, mixed-integer model that considers minimum on and off times, unit availability constraints, start-up constraints, change of water flow limit constraints, water flow constraints, reservoir balance constraints, reservoir volume limits, and reservoir spill constraints.

Conventional PSH units typically utilized synchronous machines to generate electricity. As a result, the generators’ speed is fixed at the corresponding synchronous frequency. Adjustable-speed PSH can utilize a doubly-fed induction machine (DFIM) rather than the synchronous generator. As a result, the speed of the generator can be varied and a power electronics converter can be used to control the output power [6].

There has been considerable research performed in an attempt to demonstrate the value of PSH in facilitating the penetration of variable generation such as wind and solar. The authors of [7] indicate that new PSH plants could significantly improve grid reliability while reducing the need for new thermal generation in areas with high levels of wind and solar generators. The authors of [8] devise a co-optimized coordination of wind power and PSH. They develop a stochastic MIP-based solution method that minimizes expected operating costs and corrective action costs. Wind forecast uncertainties and component outages are treated as stochastic variables. The authors of [9] investigated the use of PSH in small, islanded systems with high wind penetration.

Their study showed that PSH is particularly valuable in such scenarios due to their ability to provide primary frequency control. The authors of [10] develop an operating strategy of a hybrid wind-hydro system with the goal of ensuring wind generation output for 24 hours. The operational strategy is determined via a 24-hour stochastic, operational profit maximization optimization problem incorporating the operational constraints of the wind-hydro system. The authors conclude that by co-optimizing wind and PSH, operational profits can be increased anywhere from 12% to 22% depending on the deviation penalty level.

click to enlarge

Today’s electricity markets may not be designed in a way that would allow market regions to obtain all the benefits that energy storage owners can provide. Potential ways that ISOs and RTOs could extract more of the benefits and avoid limitations are described in [11]. For example, some entities argued in the past that in some wholesale electricity markets, some market participants may have faced undue discrimination in the way that frequency regulation was procured. The Federal Energy Regulatory Commission (FERC) acknowledged that previous compensation methods for frequency regulation did not recognize the performance benefit of faster-ramping and more accurate resources. In order to correct this, FERC Order 755 requires all system and transmission operators to pay resources based on their actual performance, including a capacity payment that covers opportunity costs and a performance payment that reflects the generator’s ability to follow the control signal [12].

Traditionally, resources were sent a smoothed, low frequency control signal for frequency control, but now it may be beneficial to allow certain generators that have the ability to follow the unfiltered or high-frequency control signal.

Power systems will become more susceptible to variability and uncertainty as the amount of VG installed increases. Variability can be seen as the expected changes in system variables while uncertainty is the unexpected changes in system variables [13]. Variability and uncertainty occurs at multiple timescales and it is important to understand these characteristics vary at different timescales. As more and more VG is installed, net load forecasting can become less accurate and system ramping events can become more prevalent. As a result, systems must adapt and become more flexible in order to maintain reliability at least cost. Due to their fast ramping and response time, PSH can be a useful tool in mitigating these problems.

The rest of this paper is organized as follows: section II introduces the model used, section III describes the test system and assumptions used, section IV describes the results of the simulations, and section V concludes the paper…

click to enlarge

Simulation

Utilizing the National Renewable Energy Laboratory’s Western Wind and Solar Integration Study data set, a test system based on the Balancing Area of Northern California (BANC) was developed. This system was deemed large enough to produce meaningful results and included significant variable generation penetration so as to adequately capture the benefits PSH can provide. The test system was then simulated for two individual weeks. One week was chosen due to its being the system peak period in July. The second week was chosen as a high variable generation output period in April. A brief system overview is provided in Table I.

The day-ahead unit commitment problem was solved every 24 hours for the next 24 hours with hourly time steps.

The real time unit commitment was solved every 15 minutes for the next three hours with 15 minute time steps. The real-time economic dispatch was solved every five minutes for the next 60 minutes with five minute time steps. The automatic generation control was solved every four second interval.

In order to simulate the advanced, adjustable-speed pumped storage hydro plants, the operating characteristics shown in table II were used. The main difference of adjustable-speed and conventional PSH is that the minimum pumping output of conventional PSH is equal to its maximum pumping output of 133 MW. As a result, the adjustable-speed PSH is able to regulate its active power output while pumping while the conventional PSH cannot.

The reserve schedules were determined based on the methodology employed in phase 2 of the Western Wind and Solar Integration Study. The requirements take into account the needs that wind and solar forecast uncertainty have on reserve requirements. Wind generators are assumed to have short term persistence forecasts and solar generators are assumed to use a cloudy, persistence forecast (i.e., assumes the current cloudiness will remain but the daily ramp up and down are factored in). These requirements are then added to the base requirements to obtain the total system reserve requirements.

The types of reserves considered are spinning, non-spinning, regulation up, regulation down, and flexibility reserves. An in depth discussion on these reserves can be found in [13].

The spin and non-spin reserves were taken as 3% of the system load. The regulation reserves were taken as the geometric sum of 1% of load and the additional requirements due to the additional wind and solar generators. The flexibility reserves were taken as the geometric sum of the solar and wind hour-ahead forecast errors covering 70% of the distribution. More details on the methodology used to determine these reserves can be found in [19]. The flexibility reserves were held in the unit commitment problems and dispatched in the economic dispatch problem. This is because the flexibility reserves were viewed as products deployed across dispatch intervals to assist the system operator.

Three different scenarios were simulated. Scenario one is the base case scenario that does not include any pumped storage hydro plants. Scenario two includes a conventional, single-speed pumped storage hydro plant consisting of three units. Scenario three includes an advanced, adjustable-speed pumped storage hydro plant consisting of three units…

click to enlarge

Conclusion

This paper explored these benefits of PSH and quantified the total production cost savings and reliability impacts on CPS2 violations, AACEE, and the standard deviation of the ACE. PSH plants can provide both production cost savings and reliability improvements for systems with significant VG. They provide much more than just energy arbitrage, including numerous ancillary services and, especially adjustable-speed PSH, system reliablitiy improvements. PSH was able to reduce total production costs in all scenarios when compared to the system without PSH. The adjustable-speed PSH was able to improve reliability, especially during high load periods. Both types of PSH were able to reduce the number of CPS2 violations. In a system without VG, the PSH were able to provide more production cost savings rather than reliablity improvements. Adjustable-speed PSH was able to signficantly improve reliablity metrics by following the unfiltered ACE control signal. Further research should be pursued to better model conventional and advanced PSH, and how best to extract and quantify their potential benefits.

Tweet