Analysis And Research On Electric Bicycle Charging Behavior Based On Data Drive

With the development of China's social and economic level to a new level,the contradiction between people's travel needs,traffic congestion,environmental pollution problems,etc.,these phenomena are becoming more and more acute.In this context,electric bicycle has become an important means of transportation for the masses because of its fast and convenient travel performance and its economical and comfortable cost factor.With the rapid rise of its ownership,the charging of ebikes is receiving more and more attention.The battery,as the power source of e-bikes,is a hotspot of research in the charging behavior of e-bikes.According to the current situation of China's electric bicycle market and the relevant national policy release,the power lithium battery will gradually replace lead-acid batteries as the main body of the electric bicycle market.However,the power battery has been restricting the development of electric bicycles due to long charging time,battery aging,and high temperature rise.This thesis takes the lithium-ion battery as the object of study and models the lithium-ion battery through an equivalent circuit model,which not only predicts the battery's charge state and remaining charging time,but also designs a charging strategy based on a variety of constraints.The main work and results of this paper are as follows.1.Three modeling methods for lithium batteries are analyzed,and the equivalent circuit model is selected as the basis for data analysis.By comparison,the second-order RC model is selected as the final battery model,on which the thermal model and battery aging model are introduced in turn.The electrical parameters of the model are deduced and identified,and finally the experimental conditions are used to verify the model,and the experimental results show that the model can simulate the operating characteristics of the battery more accurately.2.Propose a prediction model for the remaining charging time by predicting the charge state of the battery,and design a prediction model for the remaining charging time based on a double-layer support vector regression machine(DSVR).The parameters were optimized by grid search,particle swarm algorithm and genetic algorithm,respectively,and the results were compared with the single-layer support vector regression machine and BP neural network,respectively,which showed that the prediction error of DSVR was within 70 s and could accurately fit the remaining charging time of lithium batteries.Compared with the single-layer support vector regression machine and BP neural network methods,this method is more suitable for the prediction of the remaining discharge time of lithium batteries.3.A multi-stage constant-current charging strategy based on lithium-ion battery charging time,battery aging and temperature rise is proposed in order to improve the charging speed and balance the battery life.A particle swarm algorithm is used to search for the optimal charging strategy and compare it with the traditional constant-current and constant-voltage charging method,and the results show that the battery only decays by an additional 0.03 and the temperature rise of the battery is within the normal range when the charging time is shortened by nearly 27%.