Research On Model-driven Non-intrusive Household Appliance Identification Method

The energy issue is a limitation of human survival and development,electric energy is the most widely used secondary energy.With the development of China's economy and the adjustment of economic structure,the proportion of electricity consumption of urban and rural residents in the whole society continues to grow.It is very important to carry out monitoring of electricity load for urban and rural residents.This thesis takes the household appliances as the main research object,which studies the signal formation mechanism of household appliances,achieves the feature extraction and identification of single electrical appliances and predicts the load of multiple electrical appliances.the specific work of this thesis can be summarized as follows:First,considering the basic principle of work and composition of household appliances,each type of electrical model is equivalent to a combination of different functional units.Secondly,the functional unit in the equivalent model is embodied as the actual component.By using Simulink simulation software,the signal model of various electrical appliances is built.Comparing the simulation port data waveform with the measured data waveform,the mathematical model of the electrical port characteristics of various electrical appliances is obtained,which verifies the accuracy of the model.Secondly,based on the model analysis,the typical steady-state characteristics of various electrical appliances are analyzed.Aiming at the characteristics that the sampling phase of each sample is different when the current and voltage are measured simultaneously in the measured electrical data,the phase alignment algorithm is designed to ensure the uniformity of the data.Aiming at a small amount of transient transition data in steady state data,a data screening method is designed.Thirdly,considering the possible redundancy and conflict between different features,that is,among the many features,the combination of the optimal characteristics of specific electrical appliances cannot be determined.In this thesis,the feature selection algorithm based on genetic algorithm and support vector machine is used to identify each appliance,and the optimal feature combination of the electrical characteristics is selected.Fourthly,in order to solve the identification problem of multi-electrical operation at the same time,based on the superposition principle of multi-electrical current,a multi-electrical current decomposition model is established.According to the measured multi-electrical data and estimated data,the incompatible equations with constraints are obtained and the optimization objective function is established.In this paper,the quantum genetic algorithm is used to solve the optimization objective function,and the identification of multiple electrical components is achieved.