Resource adequacy and capacity Procurement: Metrics and decision support analysis

The study of electricity resource adequacy (i.e. quantifying the risk of future resource shortfall) typically uses standard metrics such as Loss of Load Expectation and Expected Energy Unserved. A range of possible risk model outputs are critiqued in this paper as a basis for decision analysis on capacity procurement, starting with the current standard approach (in which risk is commonly monetised as EEU multiplied by VOLL), and moving on to alternatives such as risk averse metrics and wider visualisations of risk profile. There are concerns in principle from a decision analytic perspective with using expected money as a utility function. Also, while there is interest in use of risk-averse metrics such as Conditional Value-at-Risk (CVaR), these might not capture all aspects of risk profile that are of interest. To obtain a comprehensive picture of the risk profile of a system, it is necessary to visualise a range of distributions of outcome measures, rather than using single number metrics. Providing such visualisations is likely not to provide the basis for a transparent risk standard, and options are discussed such as setting a level for a single metric that is occasionally revised according to changing profile of resources. Any such metric-based approach should also be supplemented by assessment of wider aspects of risk profile, for instance through examining distributions of outcome metrics. The inherent uncertainty, due to data limitations and imperfect representation of the system in the model, should also be acknowledged appropriately in estimation of such metrics.