Research On Electric Bicycle Charging Identification System Based On Smart Meter

With the growing economy of our country,electric power,as a main clean energy,has provided a huge impetus for the construction of green economy in our country and is being used more and more widely.Therefore,how to make good use of electric power reasonably,efficiently and safely has received close attention from the whole society.In the development trend of power grid,non-intrusive load identification and monitoring,as the key core technology in the future intelligent power consumption system,can not only make the power system realize more fine-grained demand-side management,peak shaving and valley filling,providing users with load operating status information and energy-saving and fee-reducing solutions,in addition to monitoring the unsafe usage behavior of certain electric appliances,such as indoor charging of electric bicycles,thereby reducing or avoid the hidden danger of fire,which has significant economic benefits for power companies and various energy service providers in the context of energy digitalization and information transformation.At present,most of the research in the field of non-intrusive load identification at home and abroad is to demonstrate and analyze the validity and reliability of various methods,lacking practical application links and solutions to technical problems encountered in them.At the same time,at this stage,general-purpose smart meters have been popularized among electric power users in my country,but their functions are limited to users’ total energy consumption measurement and control of switching on and off,and the data collection and processing functions of existing general-purpose smart meters have not been fully utilized.This topic uses the general smart meter as the basic hardware platform for data collection,and uses the AdaBoost multi-class learning algorithm based on the SVM weak classifier to design a "smart meter + load identification + electric bicycle charging monitoring" for the fire safety electricity environment.Intrusive load identification system,which can realize online identification of the opening and closing of loads,monitor the charging behavior of electric bicycles,and take measures such as sleep time detection Measures such as automatic poweroff of electric bicycles,automatic power-off after charging timeout,and alarm notification provide a new supervision method for the property,and provide users with electricity bills and energy-saving advice services.The main contents of this thesis include:（1）The research status at home and abroad of two main parts（non-intrusive load identification technology and electric bicycle charging identification）of the research on electric bicycle charging identification system based on smart meter is reviewed.（2）The requirements analysis of the system and the design of the technical architecture composition are composed of the technical architecture of the system obtained from the requirements analysis,and then the key supporting technologies of the system are proposed.（3）Realize the electric bicycle charging identification technology based on smart meter,the main contents include the selection of electric bicycle charger load identification features,data acquisition and transmission scheme design,current waveform preprocessing（denoising）,current harmonic calculation,harmonic difference Coefficient vector feature extraction,event detection method based on the difference between front and rear current harmonics,and AdaBoost multi-classification model based on SVM weak classifier.（4）The solution is put forward to the problems of real-time performance degradation and system capacity problems during the operation of the system.（5）Experiments test the accuracy of the electric bicycle charging identification system,and carry out error analysis and display of some functional interfaces of the system.