Power system load forecasting based upon combination of Markov chain and fuzzy clustering

A new method based on the combination of fuzzy clustering and Markov chain models is presented in this paper. To different types of random phenomena of in time series, several functions are built respectively. State analysis of object is carried out by using Markov chain, while fuzzy clustering is employed to the states of samples to suit the real case, then according to state transfer, the load change is predicted, The new algorithm which is used in load forecasting firstly reaches the global optimum, when the time series have strongly properties of random, the algorithm works well. The simulation results show that the error is below the level of 3.5% in most the case.