Stochastic Optimal Dispatch Of Power System Considering Uncertain Line Failures Under Extreme Weather Disaster

Over the past few years several large scale failures of power system due to extreme weather disasters such as natural disasters which not only caused great loss to national economy, also aroused the consideration of power network secure operation with the respect to extreme weather condition. Given that the line failure events caused by extreme weather disaster is a low probability event, Poisson distribution is used to depict the uncertainty of line failure events and chi-square goodness-of-fit test is used to test the feasibility. Considering the complexity of power system network, assuming the multi-line failure events are independent events, the Copula function is used to simulate multi-line failure events and K-S test is adopted to verify the feasibility of the simulation.A stochastic power system optimal dispatch model based on chance constrained programming theory is also proposed in this thesis as well. The Chance constrained programming problems are hard to solve since which involve stochastic parameters in the constraint condition. To solve the chance constrained programming problem effectively and numerically, the sample average approximation is adopted to convert the uncertain stochastic optimal dispatch into non-continuous, non-differentiable optimal problems first, then through the improved differential evolution algorithm to calculate the total cost of shedding loads and generation. Furthermore the results of improved differential algorithm are also taken to make some improvements. Case studies given in this thesis shown that the dispatch model proposed in this thesis can effectively consider the uncertain effects of power components under nature disasters, also provides more reasonable dispatch plan for power system disaster prevention and reduction to ensure the safe& reliable operation.