ABSTRACT
The variability of wind-generated electricity can be reduced by aggregating the outputs of wind generation plants spread over a large geographic area. By utilizing Monte Carlo simulations to investigate upper bounds on the degree of achievable smoothing and clarify how the degree of smoothing depends on the number of wind plants and on the size of the geographic area over which they are spread. This model has two distinct benefits of geographic diversity that have different behaviors: increased tendency of generation level to lie near its mean and decreased tendency for it to lie near its extremes. Gaussian or normal probability distributions give accurate estimates of the first but underestimate the second. The second benefit, which has particular importance for electric grid reliability, has not been widely treated before. The normal approximation, however, does not provide good estimates for the occurrence of near-zero generation levels.
