A forewarning method of cascading failure in power grid based on fuzzy clustering and fuzzy inference

To avoid blackout of power grid, during the research on cascading failure the forewarning of risk brought by it can guide the grid operation personnel to make scientific decision. Firstly, to implement the grading of forewarning of cascading failure risks the fuzzy C-means clustering and its correction are utilized to grade the probability of cascading failures; secondly, taking the indices of lost load, low voltage and overload as input variables, the fuzzy rules and membership function sets of failure severity, which can reflect static security level of power grid, are constructed and applying fuzzy inference the synthetical severity is graded; finally, synthesizing the probability of cascading failures with the severity of failure and according to the principle of the maximum membership, the forewarning levels of cascading failure risk is determined. Taking IEEE-RTS 79 system for example, the risk levels of power grid after the occurrence of cascading failures under different initial disturbances are evaluated, and the impacts of differences of power grid operating conditions on forewarning levels of cascading failures are analyzed, thus the rationality and effectiveness of the proposed algorithm are verified.