A technical perspective on variability costs: dependence on power variability and cross-correlations

Abstract

With power from integrated vRES added to electricity generated by conventional sources across a commitment interval, unexpected real-time changes in demand and generation may contribute substantial overheads to the optimal cost of energy (CoE). This paper presents analytical formulations and illustrative examples to focus on a significant fraction of the cost overheads, commonly referred to as variability costs. By contrast to studies that report a market perspective to the costs, the paper offers a technical perspective through an augmented electricity dispatch, and further takes into account crosscorrelations between randomly changing power variables. The minimum CoE across the commitment interval is obtained for the zero variability (maximum likelihood, or ML) demand-generation profile, to which variability costs add as an exclusive linear term, while a similar term accounts for forecast errors and other differentials. Modifications to the CoE that allow extension of the concept to primary-grid/microgrid environment are easily accommodated. Generation schedule examples included in the paper illustrate sensitivity of variability costs to demand variability, vRES generation variability, and cross-correlations. Distinction between central and distributed vRES integration are presented in terms of variability costs as well as overall effect on the CoE.