Sequential Convex Approximation Approach for Chance-Constrained AC Optimal Power Flow Under Arbitrary Random Distribution With Mild Conditions

In this paper, a sequential convex approximation (SCA) approach is proposed for solving the chance-constrained AC optimal power flow (CC-ACOPF) under arbitrary random distribution with mild conditions. Unlike traditional approaches that are only theoretically applicable to transforming chance constraints to some tractable constraints under Gaussian distributions, SCA can complete the same task while not restricted by the forms of random distributions. This makes more sense because the uncertainties of power systems usually follow non-Gaussian distributions. However, the proposed SCA approach is time-consuming, since it needs to sample massive scenarios to simulate the random variables' distributions existing in chance constraints. To this end, a distribution projection (DP) technique is further proposed to directly describe these random distributions in the Bernstein polynomial space. In this way, the repeated calculations in massive sampling scenarios can be omitted, thereby improving the computation efficiency of the SCA approach significantly. Case studies in several test systems validate the effectiveness of the proposed approach.