Study On Power Load Disaggregation Method Using Power Signal Characteristics

Non-intrusive power load disaggregation technology is to allocate the total power to each component electrical appliance by only analyzing the measurement data of user exit.This kind of detailed information about indoor appliances is of great significance for power companies to optimize power grid plan,operation and management,power users to save power consumption and energy charge,the whole society to put ecological civilization consciousness into practice.At present,‘disaggregation method' requires less system computing resources than ‘event method',but its disaggregation accuracy is also lower than ‘event method'.This paper revolves around disaggregation method to carry out the corresponding research and obtains the following results:1)Load modeling: a method of background load modeling and removal by using mean shift clustering algorithm is proposed,which can avoid the electric appliance power which forms background load is wrongly assigned to other electrical appliances.Clustering analysis algorithm is adopted to construct steady-state power characteristic template for above mentioned other electrical appliances,which reduces load modeling required labor.2)Real-time disaggregation: reactive power is introduced to establish multifeature real-time disaggregation model for the first time.Composition of power load appliance sparsity and electrical appliance working state transition sparsity are considered in the optimization objective function.The improved discrete monkey algorithm is used to solve the model.These measures improve disaggregation accuracy.A simple working state correction method and a power distribution method are produced,which further optimize the working state identification and power estimation results.Compared with the recent research results on public data sets,it is shown that the method proposed in this paper can effectively improve the accuracy of real-time power load disaggregation.