| The biggest problem facing the world’s energy development is the pollution of the atmosphere,the lack of resources and the change of climate.There is a large population in our country,and the living standard of Chinese people is much better than that in earlier years,the car has become a must-have item for every household.As a source of energy for traditional cars,the consumption of gasoline is increasing every year.As a non renewable energy source,the lack of resources is worse,so clean energy is becoming more and more popular.Compared with traditional fuel vehicles,new energy vehicles,mainly electric vehicles,are increasingly supported by the government and welcomed by all.Most of these electric vehicles use battery packs as their energy source.But at present,electric car battery explosion,spontaneous combustion accidents often occur.So the research and development of battery management system has been paid more and more attention in order to make battery pack use long time and safe to use.A large part of the battery management system is the estimation of the state of charge of the battery.The accurate estimation of the charge state of the battery can prolong the life of the battery.And it can improve the safety performance of the battery,and can also make the energy allocation of the electric vehicle reasonable.This topic first talks about some parts of the battery management system.The development of battery management system in some companies and abroad,and development and difficulties in estimating the state of charge of battery.Then compare various performance parameters of several types of batteries,and select the lithium iron phosphate battery that is most suitable for the energy of electric vehicles as the research object.According to the curve of the open circuit voltage and the state of charge of the battery,the concept of voltage platform phase was introduced.Next,several battery models are introduced.The neural network model is used to estimate the state of charge of the battery,and the neural network structure and various parameters are determined.Then,an extended kalman filter,an unscented kalman filter,etc are introduced briefly.Finally,the unscented kalman filter is selected to optimize the battery state of charge estimated by the neural network.,and the SOC process and results of the optimization of the unscented kalman filter are analyzed.This article is based on the measured voltage,current and temperature of the battery,and the artificial neural network model is used as a battery model to predict the state of charge.Because the state of charge estimated by the neural network is larger than the actual deviation,the unscented kalman filter is used to reduce the SOC error estimated by the neural network.By comparing the experimental and simulation data of lithium iron phosphate battery,it is proved that the method can accurately estimate the SOC of the battery.By comparison of multiple groups of real values and simulated value data.This topic finds that the SOC estimated by the neural network has a large error during the voltage plateau period,that is,real values and simulations are worth a large deviation.Then the SOC estimated by the neural network is optimized by using the unscented kalman filter,the deviation between the simulation value and the real value is very small.And it tends to the trend of real value,which greatly improves the estimation accuracy of SOC. |
