Design Of Power Gateway Based On Embedded Linux And Analysis Of Abnormal Behavior Of Power Consumption

With the "ubiquitous power internet of things" proposed,information and intelligence has become an important direction of the development of smart grid.On the one hand,communication technology is used to realize the extensive communication between low-level electrical equipment and cloud platform.On the other hand,with the gradual maturity of artificial intelligence and data mining technology,the value of power network data needs to be further mined,improve user services in the application layer.Through two aspects,power grid enterprises can realize business transformation and upgrade,and promote the sustainable development of social economy.At present,there are still some electrical equipment can not meet the development of the Internet of things requirements,but the replacement of these electrical equipment is more expensive.In this paper,a power gateway hardware platform is designed to solve the problem that the data of some electrical equipment can not be uploaded to the cloud platform.On hardware,this paper uses FET335XS-? core board to design RS485,USB,Ethernet,wireless remote interface circuit and power supply circuit around it.On software,based on embedded Linux operating platform,on data transmission,the data acquisition of electrical equipment is realized,the Modbus data protocol of the equipment layer is parsed and encapsulated into MQTT protocol frame,which is uploaded to the cloud by remote wireless network.Secondly,embedded Boa Web server and SQlite3 database,the configuration function of power gateway parameters is realized,and the system upgrade service is realized by the operation of U Disk.Finally,the experimental results show that the gateway has strong pertinence and low cost,and meets the basic functional requirements of power occasions.The power gateway collects valuable data on the cloud platform,and its potential economic value can be obtained through data analysis.In this paper,based on the analysis of the abnormal behavior of electricity consumption,the traditional detection method of electricity stealing has the shortcomings of strong subjectivity and high lag,an abnormal behavior detection model based on multi-classifier combination is proposed.Firstly,the data set is analyzed and the abnormal data are synthesized by using the abnormal data function.Then,after extracting the statistical features and reducing the dimensions,the performance indexes of the logistic regression model,SVM model,stochastic forest model and BP neural network model are compared.Finally,we use the multi-classifier combination model,which is composed of three models,such as random forest,Support vector machine and BP neural network,to test the electricity stealing behavior.Experimental results show that the comprehensive performance of multi-classifier combination model is better than that of single machine learning model.The design of the model provides some advice and direction for electric power enterprises to check the electricity stealing,which is beneficial to the stable operation of the power grid.