Non-intrusive Load Intellisense Technology For Home Energy Router

It is necessary for home energy router to have the function of sensing load information.In order to improve the load intellisense technology for home energy router,this paper uses non-intrusive load monitoring technology to realize the load intellisense function of home energy router,and proposes a non-intrusive load intellisense technology for home energy router.Firstly,the theoretical basis and system modeling of non-intrusive load intellisense are studied.Based on the research of non-intrusive load monitoring technology,the model of non-intrusive load intellisense system for home energy router is built and its workflow is designed.Secondly,the event detection method based on cumulative sum algorithm is studied.On the basis of transforming the event detection problem into the variable point detection problem,the bilateral cumulative sum algorithm with sliding window is adopted for event detection,which can accurately detect the specific time and event type and other important information according to the effective value of the total load current.In the stage of experimental verification and result analysis,the performance of the event detection algorithm is quantitatively evaluated by the evaluation indexes such as omission rate and error rate in the detection problem.Then,the method of load feature extraction and dimension reduction is studied.A feature extraction method combining transient features and steady state features is designed to provide the most identifiable load features for load identification.To solve the problem that the extracted feature dimension is too large,a feature dimension reduction method based on principal component analysis algorithm is designed to realize data visualization while reducing classification tasks.Finally,the load identification method of multi-decision fusion is studied.Under the premise of balancing the accuracy of load identification and the computational cost of algorithm,different identification decisions are taken for different types of loads,mainly including load clustering method and current curve matching method.In the stage of experimental verification and result analysis,evaluation indexes such as confusion matrix,accuracy rate,precision rate,recall rate and F1-score in classification problem are used to quantitatively evaluate the performance of load recognition algorithm.The evaluation results of other similar algorithms are collected and compared with the algorithm presented in this paper.The results prove that the algorithm presented in this paper has a higher load identification accuracy rate than other similar algorithms,and verify the feasibility and effectiveness of the non-intrusive load intellisense technology for home energy router proposed in this paper.The theoretical analysis of this paper is verified by Python language simulation and PLAID data set.