Analysis Of Residential Electricity Consumption Behavior Based On Deep Learning

Residential electricity consumption behavior analysis is the premise of "smart electricity consumption".The analysis of residential electricity consumption behavior can guide residents to optimize electricity consumption,stimulate residents’ enthusiasm for demand response,and improve the economy of electricity consumption.At the same time,the analysis of residential electricity consumption behavior can also help the grid to understand the electricity consumption of residents,alleviate the contradiction between supply and demand to a certain extent,and ensure the smooth operation of the grid.This paper focuses on the non-intrusive load decomposition method based on deep learning,and on this basis,realizes the analysis of residential electricity consumption behavior.Firstly,the basic structure and algorithm principle of typical deep neural network algorithms are analyzed,including convolutional neural network,recurrent neural network and its variant network.Secondly,four kinds of load decomposition network models were constructed based on typical deep neural network,and an improved load decomposition model was proposed based on adaptive enhancement algorithm and bidirectional gated recurrent unit neural network algorithm,and the performance of the load decomposition model was optimized by studying the Settings of different super parameters.Then,the REDD open data set was used as experimental data,and two experimental scenarios were designed to simulate and verify the load decomposition model respectively.Five evaluation indexes including accuracy,precision,recall,F1-score and mean absolute error were used to evaluate the performance of the load decomposition model.The simulation results show that the deep learning model performs well in load decomposition and has some generalization.Finally,the non-intrusive load decomposition method is applied to the analysis of residential electricity consumption behavior.According to the total power data monitored at the entrance of the house,the power information of the electrical equipment inside the house is obtained,and the power information can be visualized using tools such as line graphs,bar graphs,and tables,so as to better analyze residents’ electricity consumption behavior.By using the residential power data for example simulation verification,the simulation results prove that the analysis of residential electricity consumption behavior can be well realized on the basis of non-intrusive load decomposition.