采用多目标网格进化算法并面向对象的舰船电网重构  

Object-oriented reconfiguration of shipboard power network using multi-objective grid evolutionary algorithm

作　　者：蒋燕君[1,2] 姜建国[1] 张宇华[1] 

机构地区：[1]上海交通大学电力传输与功率变换控制教育部重点实验室,上海200240 [2]浙江树人大学信息科技学院,浙江杭州310015

出　　处：《电力自动化设备》2013年第3期26-32,共7页Electric Power Automation Equipment

基　　金：国家高技术研究发展计划(863计划)资助项目(2011AA050403)~~

摘　　要：考虑系统失电负荷量、电网有功损耗、线路负荷分配失衡度和开关操作次数,构造出舰船电网重构模型。以网格为载体,在邻域范围内进行选择、交叉和变异,采用精英策略,提出基于多目标进化并面向对象的舰船电网智能重构方法。该方法在不影响解的全局最优性的基础上,极大缩短了算法执行时间,并将各种Pareto重构算法的共同属性和操作抽象出来形成公共基础平台,改进超体积指标计算方法,实现不同算法性能间的公平比较。算例分析结果表明,在算法运行时间及所获解集的趋近度和分布度方面,所提方法均优于NSGA-Ⅱ和SPEA2。With the consideration of out-of-service loads,network active power loss,line load distribution imbalance and switch operation times,a reconfiguration model of shipboard power network is built. An object-oriented intelligent reconfiguration approach based on the multi-objective grid evolutionary algorithm is proposed for shipboard power network,which takes the grid as a carrier,carries out the selection,crossover and mutation in the neighborhood range and adopts the elitist strategy. Without affecting the global optimality of solution,its execution time is greatly reduced. A public base platform is formed by extracting the common attributes and operations from different Pareto-based reconfiguration algorithms. The calculation method of hypervolume metric is improved to realize the fair comparison among different algorithms. Case analysis shows that,the proposed approach is better than NSGA-Ⅱ and SPEA2 in the tendency and distribution degree of obtained solution set,as well as the calculation time.

关 键 词：网络重构 多目标网格进化算法 PARETO最优 超体积 舰船电网 进化算法 

分 类 号：TM711[电气工程—电力系统及自动化]