Economic Impacts of Wind Covariance Estimation on Power Grid Operations

We study the impact of capturing spatiotemporal correlations between multiple wind supply points on economic dispatch procedures. Using a simple dispatch model, we first show analytically that over/underestimation of correlation leads to positive and negative biases of dispatch cost, respectively. A rigorous, large-scale computational study for the State of Illinois transmission grid with real topology and physical constraints reveals similar conclusions. For this study, we use the RaoBlackwell-Ledoit-Wolf estimator to approximate the wind covariance matrix from a small number of wind samples generated with the numerical weather prediction model WRF and we use the covariance information to generate a large number of wind scenarios. The resulting stochastic dispatch problems are solved by using the interior-point solver PIPS-IPM on the BlueGene/Q (Mira) supercomputer at Argonne National Laboratory. We find that strong and persistent biases result from neglecting correlation information and indicate to the need to design a market that coordinates weather forecasts and uncertainty characterizations.