Salt River Project 2016-2017: The Effect of Dynamic Variability in Renewable Generation on Resource Planning and Reserve Activation
It has been observed that increasing levels of renewable integration increases the probability of loss of load events due to variability in solar and wind generation. In this project, we consider the effect of this variability in load on the probability of failure of QoS as measured by worst-case performance of load matching. While loss of load probability has been studied in the literature, the proposed work goes beyond mean-variance models of load coupled with monte-carlo simulation and instead directly models (without approximation) tail events due to confluence of forced outages and Markov models of weather impact on HVAC usage and renewable generation. Specifically, we will apply Markov models of temperature and solar intensity to yield provable bounds on the probabilities of maximum load. Coupled with simplifiedmodels of generator outages, thismodel predicts the probability of loss of load as a function of reserve generating capacity. Furthermore, we will go beyond these predictive models to determine the extent to which optimal use of generating units and batteries can reduce this loss of load probability. That is, we will determine when and how battery storage devices should be used in lieu of strategies such as ramping up additional units to meet reserve margin.
Matthew M. Peet (PI), Arizona State University
Hooman Daghooghi (PhD), Arizona State University
Robert Hess, Power Generation Services (SRP)
Jenika Raub, Power Contracts and Energy Initiatives (SRP)
Richard Anderson, Resource Planning(SRP)