Non-intrusive Load Identification Method Based On Steady-State Characteristics

Energy consumption in electric equipment is a widely concerned topic,with the rapid development of science and technology,the energy load monitoring and resolving problems of the electrical equipment has become the attention of the researchers.Traditional home energy monitoring programs typically install sensing equipment at the electrical distribution outlet of a single electrical load to monitor the use of electrical equipment and energy consumption.This type of invasive load decomposition solution requires enormous human and material resources.And it also imposes a burden on the user's normal life.Non-intrusive load monitoring collects relevant power signals from sensing devices installed at the home's main power entrance,and processes and extracts the information carried by these signals,and finally systematic modeling to get families within each type of power load and the running state.In contrast,the installation equipment of this monitoring scheme has a higher cost performance,but at the same time it also has many limitations and challenges.How to deal with the collected signals and how to balance the efficiency of identification with the efficiency of the energy load monitoring and decomposition of electrical equipment are the difficulties that researchers urgently need to solve.Based on the self-developed intelligent socket hardware,this paper mainly focuses on the feature selection and load decomposition of two modules in steady state,and carries out experimental verification on the Matlab software simulation platform.The main research contents of this article include:(1)First,through the self-developed intelligent socket hardware device to collect electrical data signals,to study the characteristics of voltage and current data signals of different types of household appliances in steady state and transient state,and to extract the high-performance features of these electrical data signals as to select the feature parameters,a multi-feature modeling method was proposed to make full use of the effective information carried by multiple features for effective identification.At the same time,based on these multi-feature parameters,this paper also establishes a feature data set for multiple electrical appliances identification.(2)In the non-invasive decomposition system,how to determine whether there is an electrical state change is also an important link in the system.In this paper,the status of time series is observed by means of statistical index or statistical method,so as to estimate the position of the change point accurately and effectively.A sliding window is used to determine the state of electrical appliances in the system.(3)In the final stage of the system model,in load identification,this paper adopts an optimization-based scheme that uses the error between the combination of the features of the unknown load and the known load characteristics to find the closest possible match.According to the statistical characteristics of the steady-state operation of the electrical appliances,a quadratic programming is proposed to establish a model,and the model is solved according to the advantages of the particle swarm algorithm.Finally achieve the desired experimental results.