The Research Into Fast-charging Method Based On Neural Network Control

Increasingly serious environmental pollutions cause the air quality constantly suffer challenge. Asmore and more vehicles, vehicle exhaust emission has become the main source of pollutions. For our livingenvironment, use the environmentally-friendly electric way has become an urgent need. Today, thetraditional charging method that most charging batteries using need a longer time, not only extend thedetection time, but also limit the widespread use of electric cars, the owners’requirement of fast chargingand convenient become more and more high, and many existed ways of fast charging use big chargingcurrent, at the expense of the life of the battery for charging time, but it is not our intention, also did notachieve the purpose of environmental conservation. Therefore, in terms of domestic battery technologydevelopment and application, the research at any time of charging and fast charging method occupies animportant position.By the basic structure and working principle of lead-acid battery in this article, discusses theexisting various traditional and improved battery charging method, compares their advantages anddisadvantages, we can come to the conclusion that each one can’t change charging current in the process ofachieve the ideal curve of Mas. Based on lead-acid battery quick charging characteristics and principles,this paper presents the neural network algorithm which is applied to the process of fast charge, using neuralnetwork self-learning habits and adaptability etc. Suitable for processing the characteristics of the nonlinearsystem to achieve the purpose of the intelligent charging.In this paper, through the analysis of the neural network structure and algorithm, the comparisonof three kinds of controller, the final selection is make the neural network and PID controller incombination with intelligent controller. Through calculation and comparison of all kinds of neural networkalgorithm, based on the goal of this experiment, puts forward to the improved BP algorithm based onconjugate gradient method, and use the battery’s three main parameters as variables of the neural networkinput layer, monitoring the change of the current, voltage, the temperature of the battery at the same time,avoiding the special condition in the process of battery charging for the lead to the detriment of thecharging process. In this experiment, by using the proposed new charging method for MATLAB software simulation, we can get the new and old two current contrast figure, and clearly see that: based on neuralnetwork and PID control of battery fast charge method accuracy is higher than common charging method,closer to Mas charging electric current curve, puts forward the best so as to realize the fast, safe andefficient battery intelligent charging method.