Understanding options for integrating hybrids into wholesale power markets
Berkeley Lab-EPRI study quantifies the pros and cons of multiple participation models for hybrid power plants
Utility-scale hybrid storage projects are on the rise in the U.S. The growth of these unique resources – which comprise battery energy storage and a separate technology, often solar or wind, behind the same point of interconnection – creates uncertainties for system operators, who grapple with efficient and reliable ways of operating them. An immediate challenge is properly representing hybrid resources in market clearing software, which requires defining a model for their participation in wholesale power markets.
To help system operators better understand and utilize hybrid resources, Berkeley Lab is pleased to announce the release of its Integration of Hybrids into Wholesale Power Markets report. By comparing the implications of different hybrid participation models in a realistic simulated wholesale electricity market, the study offers specific metrics to quantify the advantages and disadvantages of each. Participation models were evaluated across several dimensions: the impact to overall system reliability, economic efficiency, the market’s ability to access hybrids’ full capabilities, and asset profitability.
The study, conducted in partnership with the Electric Power Research Institute (EPRI), evaluated the implications of different levels of hybrid resources on an example system (the New York State Control Area) to quantify both economic and reliability metrics. It focused on hybrid participation in day-ahead energy markets, with a real-time balancing mechanism to ensure metrics realistically captured the impact of forecast errors from load or variable energy resources. Scenarios incorporated the two main modeling options currently being explored for hybrid resource participation in wholesale markets, namely:
- Integrated hybrid resource model (1R), where market participants provide a set of paired price/quantity offers for each market interval and structure those offers to maximize profit while attempting to maintain a feasible schedule.
- Separate co-located resource model (2R), where market participants may or may not provide a single price/quantity offer for each market interval, but will submit other information (e.g., renewable resource forecasts and storage state-of-charge information) so market clearing software can optimize the resource to minimize costs across the system.
Fig. 1: Different configurations of hybrid storage resource participation models.
The study offers conclusions across a number of important areas:
- Economic efficiency. The 2R model generally provides greater cost savings, but differences in efficiency across participation models were not significant for this case study.
- System reliability. No measurable impacts were observed in this case study since sufficient quick-start capability was able to manage infeasible battery state-of-charge levels and variable energy resource forecast error.
- Asset profits and incentives. The 2R model provides greater short-run profits compared to 1R.
- Hybrids’ ability to follow different real-time operational strategies. An inability to follow a day-ahead schedule was observed more frequently with the 1R model.
- Load payments. These depend on hybrids’ cleared energy awards and can differ considerably based on bidding strategies and battery state-of-charge considerations.
- Computational efficiency. The 2R model adds greater computational complexity and solve time as the number of hybrids increases.
- Trading strategies. While the offer strategies were considered state of the art for a study environment, empirical bids from operational hybrids could help substantiate these results as hybrid resources begin to play a larger role in markets.
In general, the study confirms the current advantages of the separate co-located resource model (2R) over the integrated hybrid resource model (1R) under different resource mixes and market practices. Real-world experience appears to validate this conclusion, as most hybrids elect the 2R model and some ISOs/RTOs do not yet offer a 1R model. Whether and for how long this preference will persist is unclear, given potential changes in resource mix, market design, and computational complexity. The introduction of more complex aggregated resources, such as those with more than two technologies, would pose additional challenges.
The full report, along with a briefing slide deck, can be found at https://emp.lbl.gov/publications/integration-hybrids-wholesale-power.
For recent trends on hybrid power plants in the United States, visit https://emp.lbl.gov/hybrid.