The Effects of Renewable Power and Forecast Error Correlation on Balancing Reserve Needs

Renewables introduce randomness into the operation of power systems through forecast error. This randomness is commonly assumed to increase reserve needs. We explore how, through renewable overbuilding and curtailment, we can reduce the burden on non-renewable resources to provide reserves, since overbuilt and curtailed renewables can provide steadier power than if they run at their maximum power point. In this work, we investigate the interplay between overcapacity of renewables, spatial correlation in forecast error, and consequent reserve requirements using the IEEE 9-bus system as our test case. To obtain reserve capacities, we formulate a dispatch-redispatch paradigm where the system is first dispatched on forecasted power, then redispatched to mitigate forecast error. We conduct Monte Carlo simulation to obtain the quantiles of setpoint deviations under redispatch. Our results show that without significant network constraints, non-renewable reserves approximately vary with the aggregate renewable variability. However, with significant load and network constraints, reserve needs increase significantly and no longer vary with aggregate renewable variability.