Optimizing Electricity Resource Investment with Large Penetrations of Wind and Solar Capacity Considering System Cost and Reliability
Bothwell, Cynthia Diane
MetadataShow full item record
Investors rely on models to give insight into how electricity system components work together to match demand with supply in a reliable manner, those components being: consumer demand, consumer technologies to generate or curtail, and producer technologies. Over the last decade, increasing emphasis has been placed on wind and solar generation sources that are dependent on weather conditions. The contribution of this thesis is to provide the first integrated analysis of the implications for generation adequacy studies, capacity power market design, and long-run market efficiency of alternative methods that account for the unique variability characteristics of renewables, while maintaining comparable treatment both between weather dependent and thermal generators and between renewable resources that provide different system benefits. Using both statistical methods and optimization techniques, I develop model procedures that improve renewable representation to enhance decision making towards efficient solutions. I investigate and recommend methods to determine the renewable generator capacity value for resource adequacy studies and electricity market procurements – each of which significantly differs from current industry practice. I investigate several aspects of renewable generation in electricity markets. First, from the viewpoint of selecting a market capacity mechanism, I explore revenue sufficiency for generation investment combined with large renewable additions and find that although renewable generation plays a role in reducing profits to thermal generators, revenue shortfalls also come from other more significant factors. Generally, capacity mechanisms raise prices for consumers. Secondly, where electricity markets use a secondary market for capacity to maintain adequacy, I show that the correct valuation of renewable generation is the marginal value during critical hours. I further find that the selection of critical hours is dependent on the sample of correlated historical demand and renewable production, and that the increased system variations require more expansive data than used by industry. Finally, I find that each renewable facility can make a unique system contribution and that to most efficiently incentivize renewable additions that meet reliability and renewable tar-gets, each facility should receive a differentiated payment based on its contribution rather than a uniform value broadly determined across a technology type.