Research On Power Dispatch System Based On Evolutionary Algorithms

With the implementation of reform and the constantly increasing of the reform measures, power plants take increasing emphasis on power generation cost. The question that how to reduce the electricity generation cost of thermal power plants, becomes more and more important. Meanwhile, the environmental pollution is constantly serious in recent years, especially in the last few years, fog and haze happened frequently, the phenomenon that PM2.5exceeds the warning level has happened all over the most China. People’s daily lives and healthy are damaged. So meeting the emissions requirements is necessary. In order to keep the security and stability of power supply, meeting the allotted time of power dispatch is necessary as well. Therefore, the study of these factors and the design of load dispatch systems are very significant.Firstly, a few typical algorithms are selected through traditional optimization algorithms, mathematical programming algorithms, biological swarm intelligent methods, and modern intelligence algorithms. Through comparative analysis of these algorithms, genetic algorithm is selected in this paper.Secondly, there is an analysis of the characteristics of genetic algorithms and the existing deficiencies. The deficiencies of global search capability and convergence performance are improved, by change the rule of selection probabilities, crossover and mutation rate.Then, according to the needs of power generation plants, there is a discussion of the evaluation criteria of load dispatch systems, and established five kinds of dispatch programs.Finally, use case simulations to verify the effectiveness of the improved algorithm, and the superiority. This paper also carried out the analysis of the effect of five plans of load dispatch system. The results can be used as theoretical basis for policy-makers to develop load dispatch plans.