基于MIV和GA优化Elman神经网络的春节负荷预测

Spring Festival Load Forecasting Based on MIV and GA Optimized Elman Neural Network

春节作为中国的传统节日,群众集中返乡造成部分配变台区发生重过载现象,严重影响供电可靠性。精准的负荷预测可以帮助公司高效地开展春节保供电工作,确保节日期间居民用电平稳有序。文章对青岛市某一配变春节期间负荷特性进行分析,通过应用平均影响值进行输入变量的筛选,运用遗传算法对Elman神经网络初始阈值和权重进行优化,结合配变的额定参数,预测春节期间是否重载或过载。算理分析表明,该方法预测精度高,在工程应用中具备可行性。

As China's traditional festivals,the masses focus on returning home caused by part of the distribution zone heavyoverload phenomenon during the Sprint Festival,a serious impact on power supply reliability.Accurate load forecasting canhelp companies to carry out the work of the Spring Festival to ensure that the residents during the holiday season is stableand orderly.In this paper,the load characteristics during the Spring Festival in Qingdao are analyzed.The input variablesare selected by applying the Mean Impact Value.The genetic algorithm is used to optimize the initial threshold and weightof the Elman neural network.According to the parameters of transformer,predict if it is overloaded.The arithmetic analysisshows that the method has high accuracy and is feasible in engineering application.