基于群体智能算法的电力台区数据分析技术  

Analysis of Power Station Based on Group Intelligent Algorithm

作　　者：张欣 庄园 宁学玲 治强 ZHANG Xin;ZHUANG Yuan;NING Xue-ling;ZHI Qiang(Training Center of State Grid Jilin Electric Power Co.,Ltd.,Changchun,Jilin 130000,China;State Grid Jilin Comprehensive Energy Service Co.,Ltd.,Changchun,Jilin 130000,China)

机构地区：[1]国网吉林省电力有限公司培训中心,吉林长春130000 [2]国网吉林综合能源服务有限公司,吉林长春130000

出　　处：《计算技术与自动化》2022年第2期87-91,共5页Computing Technology and Automation

摘　　要：电力系统经济负荷运行调节中,存在复杂的非凸、非线性问题,难以实现电力系统运行可靠性和经济性。传统群体智能算法只是单一模仿生物群体行为,不能很好地进行电力系统负荷经济调节。因此提出了一种基于群体智能算法统一框架的粒子群算法,算法根据不同的适应度隶属度函数赋予不同惯性权重,通过外点法构造辅助函数将非可行域的约束以罚函数形式写入目标函数,并建立变罚系数提升目标函数求解过程的搜索范围和收敛速度,同时由搜索策略和变异策略,提升全局搜索能力。通过仿真结果表明:改进PSO算法相较于传统遗传算法、粒子群算法的收敛速度更快,解的离散度更小,收敛精度更高。In the power system economic load operation regulation,there are complex nonconvex and nonlinear problems,which is difficult to realize the operation reliability and economy of the power system.The traditional group intelligent algorithm only imitates the behavior of biological groups and cannot perform the power system load economic regulation well.This paper proposes a particle group algorithm based on the unified framework of group intelligence algorithm,which gives different inertia weights according to different fitness membership function,constructs the auxiliary function to write non-feasible domain to the target function as the penalty function,and establishes the search range and convergence speed of the solution process of variable penalty coefficient,and the variation strategy to improve the global search capability.The simulation results show that the improved PSO algorithm converges faster,has less dispersion and higher convergence accuracy than the traditional genetic algorithm and particle group algorithm.

关 键 词：粒子群算法 群体智能框架 罚系数 经济负荷分配 收敛精度 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]