基于多智能体伪并行遗传算法的小水电无功优化研究  

Study on Reactive Power Optimization of Small Hydropower Based on the Multi⁃agent Pseudo⁃parallel Genetic Algorithm

作　　者：闫群民 张田[3] 朱娟娟 沈延峰 李杰义[2] 周密 YAN Qunmin;ZHANG Tian;ZHU Juanjuan;SHEN Yanfeng;LI Jieyi;ZHOU Mi(Shaanxi Key Laboratory of Industrial Automation,Shaanxi Hanzhong 723001,China;State Grid Ankang Power Supply Company,Shaanxi Ankang 725000,China;Department of Electrical Engineering,Shaanxi University of Technology,Shaanxi Hanzhong 723001,China)

机构地区：[1]陕西省工业自动化重点实验室,陕西汉中723001 [2]国网安康供电公司,陕西安康725000 [3]陕西理工大学电气工程学院,陕西汉中723001

出　　处：《电力电容器与无功补偿》2021年第1期18-24,共7页Power Capacitor & Reactive Power Compensation

基　　金：陕西省教育厅重点实验研究计划项目(20JS018);国网陕西省电力公司科技项目(5226AK1900CY)。

摘　　要：为了解决小水电接入配电网带来的电压波动、网损增加等问题。构建了以配电网网损最小为目标函数,电压偏移量与无功补偿量为罚函数的无功优化数学模型,提出了基于多智能体和伪并行遗传算法的智能混合型求解算法,该算法通过邻集拓扑和网格环境的交换来完成全体Agent的更新,增加了进化的并行性,在优化遗传算子的基础上引入了自学习算子,该措施在有效强化系统多样性的同时降低系统内部传输的数据量,从而增强了遗传算法进行全局搜索的能力。在Matlab中对IEEE 30节点系统进行仿真和验证,并将所得结果与传统遗传算法、多Agent遗传算法进行比较,结果表明该算法能有效降低系统网损,提高电压水平,而且显著地提高了运算速度,具有更强的全局寻优能力。For solving the problem of voltage fluctuation and the increase of network loss caused by the con⁃nection of small hydropower to the distribution network,the reactive power optimization mathematical model is set up with minimum network loss of distribution network as target function and voltage offset and reactive power compensation as penal function.The intelligent hybrid solution algorithm based on multi⁃agent and pseudo parallel genetic algorithm is proposed,which completes the update of all agents by way of exchange of adjacent set topology and grid environment.The parallelism of evolution is increased.The transmission data amount in the system is reduced in the system at the time of strengthening system diversity by way of introducing self⁃learning operators on the basis of optimizing genetic operator,thus improving the ability of global search of the genetic algorithm.Finally,the IEEE 30 node system is simulated and verified in Matlab and the obtained result is compared with the traditional genetic algorithm and multi⁃agent genetic algo⁃rithm.It is shown by the result that the algorithm can reduce effectively the network loss,improve voltage level and calculation speed,and have stronger global optimization and efficiency.

关 键 词：小水电 无功优化 多智能体 伪并行遗传算法 

分 类 号：TM612[电气工程—电力系统及自动化] TM714.3[自动化与计算机技术—控制理论与控制工程] TP18[自动化与计算机技术—控制科学与工程]