基于极限学习机的居民用电行为分类分析方法

针对大数据背景下智能用户用电行为分类问题,提出一种基于极限学习机(ELM)算法的用户用电行为分类方法。首先,在前期用户行为的特征优选策略的基础上,采用特征优选策略提取负荷曲线的最佳特征集对用户用电数据进行分类分析。然后,将特征优选集作为输入,通过比较不同隐含层激活函数和隐含层节点个数下训练集和测试集的正确率,优选出适用于用户用电行为分析的ELM算法的输入参数。最后,以国内和国外用户用电数据为数据源,进行算例仿真实验,通过与反向传播(BP)神经网络的对比分析表明,所提出的基于ELM算法的用户用电行为分析方法提高了检测的正确率并且降低了算法运行时间,能够更好地掌握用户用电负荷状态,实现配电网的削峰填谷。

Classification Analysis Method for Electricity Consumption Behavior Based on Extreme Learning Machine Algorithm

Aiming at the classification problem of electricity consumption analysis of smart users under the background of big data, a classification method based on extreme learning machine(ELM)algorithm is proposed for electricity consumption behavior analysis. Firstly, based on the previous research of feature preference for electricity consumption behavior of smart users, the feature preference strategy is adopted to extract the best feature sets of the load curve, which helps to classify and analyze the data of electricity consumption for users. Then, the best feature sets are used as the input of ELM network. By comparing the accuracy of the training set and the test set with different hidden layer excitation functions and hidden layer node numbers, input parameters of ELM algorithm are selected, which are suitable for user''s electricity consumption behavior analysis. Finally, the domestic and foreign electricity consumption data is taken as the data source to carry out the simulation experiment. Through the comparison and analysis with back propagation(BP)neural network, the results show that the analysis of electricity consumption behavior based on ELM algorithm improves the detection accuracy and reduces the algorithm operation time, which can better grasp the user load status and realize load balance of distribution network.