Research And Application Of User's Electricity Behavior Analysis Method

With the development of communication technology,computer science and the increasing proportion of consumers' electricity consumption in energy consumption,the analysis of consumers' electricity behavior combined with big data technology is becoming more and more active.Household electricity usage is of great importance to power companies,consumers,and grid systems.This paper is mainly based on distributed computing,load decomposition,sequential pattern mining and other theories to mine consumers' frequent electricity consumption mode.The research in this paper is mainly divided into the following parts: method design based on Hadoop platform,construction of the using sequences about electrical devices and mining of electricity consumption mode.The method based on Hadoop platform is designed to improve the overall efficiency and scalability of the proposed method.The sequences of electrical devices are used to provide samples for sequential pattern mining.Sequential pattern mining is to obtain a frequent using mode of electrical devices about certain types of consumers.The specific research content is as follows.Based on data mining technology,smart meter data,input and output of load decomposition algorithm,input and output of sequence pattern mining algorithm,and characteristics of Hadoop platform,the consumers' electricity behavior analysis method is designed.The input and output of each stage are specified to make it suitable for Hadoop's MapReduce model.Hadoop experimental platform of this paper is built.Then,the load decomposition was studied.Around the goal of accurately constructing the sequence of electrical devices,the current harmonics and power with linear superposition are selected as the characteristics,the normal distribution metric function is introduced to fuse the two to construct the fitness function,and the chicken swarm algorithm is discretized and improved as the load recognition algorithm.The decomposition of electricity consumption events is completed,and the using sequence of electrical devices is successfully constructed.Experiments show that the sequence of electrical devices constructed in this method has good reliability.Subsequently,the principle of several classical algorithms in sequential mode is studied.The advantages and disadvantages of each algorithm are compared and analyzed.The PrefixSpan algorithm with high mining efficiency is parallelized to make it suitable for Hadoop platform.The algorithm is applied to mine using sequence of consumers' electrical devices.The experimental results show that the MR-PrefixSpan algorithm has the highest mining efficiency in the consumers' electricity behavior analysis.The frequent using sequence of the electrical devices is analyzed to obtain the frequent electricity usage mode of the consumers.The practical application scenarios of the method are described.Combined with the time-sharing electricity price,the rationality of the consumers' frequent power consumption mode is evaluated and used as the basis for recommending appliances in smart homes,so as to illustrate the practical significance of the proposed method.Finally,the paper summarizes and analyzes the work of this paper,pointing out the shortcomings of this paper and the research direction of future work.