基于进化性粒子群优化算法的概率潮流计算方法研究  

作　　者：蒋文骏 Jiang Wenjun(Shanghai Wisdom Information Technology Corporation,Shanghai,200052)

机构地区：[1]上海威士顿信息技术股份有限公司,上海200052

出　　处：《科学与信息化》2020年第8期187-189,194,共4页Technology and Information

摘　　要：本文提出了一种处理基于人工智能的潮流计算的新方法.该方法对任意输入变量的分布采用有限混合高斯表达电力系统中重要状态量的非正态发布的不确定性.近年来,清洁能源发电给电力系统带来了更大不确定性.因此,在新环境下表达概率特征电力系统工具得到了进一步的重视.电力系统运筹学研究着重于电力系统的不确定性特征.本文提出了一种基于人工智能算法的概率潮流计算技术,该技术通过元启发式算法的EPSO推理混合高斯函数的最优参数,对输入变量进行有限混合高斯函数表达系统状态变量的任意概率分布特征.EPSO是一种基于人工智能的粒子群优化算法的扩展算法,其利用进化策略调整参数获取模型的最优参数.该方法的优点是可以趋近与输入变量的任意概率分布,使潮流计算的结果更符合现实情况.In this paper,a new method is proposed to deal with artificial intelligence based load flow calculation.The proposed method makes use of the finite mixture Gaussian Approximation for any distribution of input variables.In recent years,the clean energy generation brings about uncertain power system conditions.As a result,stochastic power system tools have been reevaluated under new environment.Power system operations are interested in stochastic characteristics of power system conditions.This paper proposes a new stochastic factory-level load flow technique that considers any distribution of input variables through approximating them with the finite mixture Gaussian functions.EPSO of meta-heuristics is used to estimate the optimal parameters of the Gaussian functions.It is the extension of PSO that tunes up the parameters with the evolutionary strategies.The proposed method has advantage that it handles any distribution of input variables.

关 键 词：元启发式算法 进化性粒子群优化算法 人工智能 概率潮流计算 混合正态分布 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]