Power System Reactive Power Optimization Based on Direct Neural Dynamic Programming

摘要：

Reactive power optimization in power system is acomplex nonlinear combinatorial optimizationproblem with multiple constrained conditions.However,direct neural dynamic programming(directNDP)approach based on on-line measurements canbe employed in this situation,which is independent ofmodels.In this paper,on the basis of applicableanalysis to reactive power optimization,this algorithmis improved in the expansion of dimension,and then adirect NDP model is established for reactive poweroptimization.It is mainly composed of two neuralnetworks: action network(AN)and critic network (CN),AN is used to control,and CN is used to evaluatecurrent system states and update AN.At last,theimproved algorithm is tested in the IEEE 6-bus systemand compared with the GA optimization algorithm,theresults demonstrate that the new algorithm is afeasible and effective way to solve the reactive poweroptimization problem.

作者: Zhigang Lu Liye ma

作者单位: Department of electrical engineering,Yanshan University,Qinhuangdao 066004,China

会议类型: 国际会议

会议名称: 2008 3rd International Conference on Intelligent System and Knowledge Engineering(第三届智能系统与知识工程国际会议)(ISKE 2008)

会议地点: 厦门

会议语种:英文

页码: 862-866

在线出版日期: 2008-11-17（万方平台首次上网日期，不代表论文的发表时间）