Non-intrusive EV Charging Load Monitoring Based On Independent Component Analysis Algorithm

Under the premise that the electricity consumption behavior of power generation equipment and energy storage equipment interferes and offsets each other,traditional intrusive monitoring is not conducive to the long-term development of the power system because it is easy to trouble users during maintenance and installation and the installation cost is high..In contrast,the advantages of Non-intrusive Load Monitoring(NILM)are obvious.The collection method is simple.It only needs to collect the mixed signals at the electricity consumption entrance.Load identification can be realized by signal analysis and processing algorithms.Under the premise of not affecting users’ lives and protecting their privacy,costs will also be reduced.In the exploration of solutions to environmental and energy problems,the development of electric vehicles followed by large-scale use and the increase in penetration rate,although alleviating the problem of energy shortage,its own random and disorderly charging behavior also gives stability to the grid Sex caused interference.In order to solve the charging problems of electric vehicles,this paper uses an improved independent component analysis algorithm(ICA)based on NILM to analyze the charging behavior process of household electric vehicles.The following is the main content of this article:(1)Analyze the NILM system framework and the load characteristics of residential electric vehicles.Firstly,a combination method of thresholding and filter is adopted,and a non-intrusive load monitoring method is proposed through data simulation.It is applied to the stable process of electric vehicle charging to monitor the load waveform.This extraction method can only extract the stable charging behavior of electric vehicles.The extraction results of the state charging stage provide a comparative basis for the application of the algorithm proposed in this article to the extraction of load waveforms.(2)Improvements to the ICA algorithm.According to the realization principle and process of the ICA algorithm,build the ICA algorithm model,introduce a modified third-order convergence Newton iterative algorithm on this basis,and analyze and discuss its convergence,thereby improving the update iterative process of the ICA algorithm,and Through the simulation image simulation of the signal generated by the random matrix to compare the effectiveness of the separated observation signal,it is verified that the convergence speed of the ICA algorithm is faster;on this basis,a new denoising algorithm is proposed for the noise problem in the ICA algorithm.When adding a new iterative method to process the noise signal,select an appropriate signal-to-noise ratio to minimize the influence of the noise term on the separation solution of the random signal,and verify the effectiveness and rationality of the improved ICA through simulation experiments.And the convergence of speed.(3)Finally,the improved ICA algorithm is applied to the extraction process of the proposed non-invasive household electric vehicle charging behavior.Including three specific stages of extracting the charging behavior;Different methods for different stages of charging behavior;and 4 methods for estimating the charging power of different types of electric vehicles,including an algorithm that improves the accuracy of the load power for local correction.Finally,through simulation experiments,the non-intrusive load monitoring based on the ICA algorithm and the non-intrusive load threshold extraction were compared and verified,which effectively verified the feasibility of this algorithm and the accuracy of the extracted electric vehicle charging behavior.