Research On Algorithm Of Reactive Power Optimization Based On Cloud Computing And Improved NSGA-Ⅱ

Reactive power optimization(RPO) of the power system is under the given structure and operation of the current power grid, make the current system achieve the minimum network loss, the most of voltage stability and so forth. RPO involves adjusting some control variables such as the output power of generator nodes, transformer voltage ratio, switching parameters of reactive power compensation device like capacitor and so on. The significance of RPO includes combining power transmission quality with low carbon and economic stability. Therefore, a reasonable multi-objective optimization algorithm for reactive power grid scheduling is essential. However, The rapid expansion of the power system makes the data is too large when dealing with RPO. RPO on a single machine takes long time to calculate and optimize ineffective, Therefore, the research in this article to start to address the problem.Firstly, A RPO model is built with double objective function, including active power loss minimum, maximum static voltage stability margin. NSGA-Ⅱ is used to solve the RPO problem. It has better optimization results and faster convergence speed compared with simple genetic algorithm(SGA), particle swarm optimization(PSO). Demonstrated the superiority of NSGA-Ⅱ algorithm to solve RPO problem.Secondly, An improved eliminate strategy is proposed against the problem of more outstanding individual is easy eliminated when calculating the degree of congestion for the NSGA-Ⅱ algorithm. In addition, the NSGA-Ⅱ algorithm crossover and mutation rate are adaptive improved based on the cloud model. Then got an improved adaptive NSGA-Ⅱ(IA-NSGA-Ⅱ) algorithm. It has better optimization performance compared with NSGA-Ⅱ algorithm when dealing with RPO problem.Finally, the IA-NSGA-Ⅱ algorithm are parallelizing improved based on cloud computing and distributed computing ideas, so that the IA-NSGA-Ⅱ algorithm can run on Hadoop platform environment, and improved the ability to handle large-scale power system RPO. Some examples are tested by the parallelizing IA-NSGA-Ⅱ algorithm. Verified the efficiency and strong optimization capabilities of parallelizing IA-NSGA-Ⅱ algorithm when dealing with RPO problem.