Study On The Power System Unit Commitment Problem

Unit commitment (UC) plays an important role in the study of the power system problems and it is the premise of the optimal power flow (OPF), economic dispatch(ED). The first thing UC should deal with is the short time power generation scheme setting. Also this scheme means a lot to the power industry and the whole national economy. On account of the high-dimensional, discrete, nonlinear of the problem, the paper had done some research on the traditional algorithms and the intelligent algorithms. The main content of study covers the following:As for traditional algorithms, we brought the generalized Benders' decomposition algorithm to the UC problem. Based on the idea of dual programming, took the change of the load in consider, we raised a heuristic algorithm. The algorithm decomposed the period of the load, which solve the hard-decomposition problem in UC. We decomposed the problem as more as we could by the generalized Benders' decomposition algorithm and the heuristic algorithm to build a series of independent but small problems which reduced the scale of the sub-problem and enhanced the efficiency of the algorithm in solving the problem.As for intelligent algorithms, we did the future expansion to the ant algorithm and raised Multi Colony Chaotic Ant Optimization Algorithm (MCAO). The algorithm defined different ant colonies, made those ant colonies cooperate, help and communicate with each other. Finally it enhanced the efficiency of the algorithm and improved the quality of the result. In the paper we brought the concept of the window which reduced the scale of the problem, at the same time we brought the algorithm of chaos and 2-opt, which could deal with the shortcoming of the ant algorithm in easily falling into local optimum and improve the quality of the algorithm.