一种含牵引负荷的概率潮流计算方法  

A Probabilistic Power Flow Calculation Method with Traction Load

作　　者：邓苏娟 张百甫 王飞 DENG Sujuan;ZHANG Baifu;WANG Fei(Anhui Vocational College of City Management,Hefei 230000,China;China Energy Engineering Group Anhui Electric Power Design Institute Co.,Ltd.,Hefei 230000,China;State Grid Nanchong Supply Company,Nanchong 637000,China)

机构地区：[1]安徽城市管理职业学院,安徽合肥230000 [2]中国能源建设集团安徽省电力设计院有限公司,安徽合肥230000 [3]国网南充供电公司,四川南充637000

出　　处：《电工技术》2023年第15期86-88,101,共4页Electric Engineering

基　　金：安徽城市管理职业学院自然科学项目(编号2021zrkx05)。

摘　　要：含牵引负荷的概率潮流算法一般是将概率统计方法与确定性潮流计算方法相结合,但常见的确定性潮流算法如快速解耦法和牛顿-拉普逊法,在计算过程中存在不收敛、计算精度偏低的问题。为此提出一种含牵引负荷的概率潮流计算方法,采用Copula函数来描述输入变量的相关性,通过拉丁超立方函数对输入变量进行抽样,将深度神经网络法用于确定性潮流的计算。最后在改进的IEEE 30节点系统中验证了所提算法的适用性。Probabilistic power flow algorithms with traction loads generally combine probabilistic and statistical methods with deterministic power flow calculation methods.However,common deterministic power flow algorithms,such as fast decoupling method and Newton-Laptson method,have the problems of non-convergence and low calculation accuracy in the calculation process.Therefore,a probabilistic power flow calculation method with traction load is proposed,and the Copula function is used to describe the correlation of input variables.The input variables are sampled by the Latin hypercube function,and the deep neural network method is used for the calculation of deterministic power flow.Finally,the applicability of the proposed algorithm is verified in the improved IEEE 30-node system.

关 键 词：牵引负荷 COPULA函数 相关性 拉丁超立方函数 深度神经网络 

分 类 号：TM711[电气工程—电力系统及自动化]