Research On Estimation Of State Of Charge And Health Of Lithium-ion Battery Based On Data Driven

Electric vehicles(EV)are an effective way to solve the problems of energy exhaustion and environmental pollution,and their safety performance and cruising range cannot fully meet the need of large-scale development.Lithium-ion power batteries,as the energy supply component that determines the cruising range of EV,account for more than 1/3 of the cost of EV.Therefore,the state has included the power battery special technology research into the "14th Five-Year Plan".Accurate estimation of State Of Charge(SOC)and State Of Health(SOH)can ensure safe and efficient operation of the battery and improve battery life.However,temperature,aging,working conditions and other factors restrict the estimation accuracy of SOC and SOH.Based on this,this article mainly studies the accurate estimation algorithm of SOC and SOH for lithium-ion batteries.First of all,the working principle of the charging and discharging process of lithium-ion batteries is summarized,and the battery voltage,internal resistance,energy,power,capacity,self-discharge rate,depth of discharge and other performance parameters are introduced.The relationship between SOH,summarized the common lithium-ion battery model.Afterwards,a battery test experimental platform was constructed,and the battery pack maximum capacity test,open circuit voltage test and hysteresis test program was designed,and SOC and SOH influence factors were analyzed based on the CALCE and NASA PCOE international databases.Analysis shows that the capacity of high-discharge rate batteries is less than that of low-discharge rate batteries,and the battery capacity is less than normal temperature in low and high temperature environments.The increase in the number of cycles will reduce the maximum available capacity of the battery,and overdischarge will sharply reduce the specific capacity of the battery.After the above analysis,SOC and SOH are closely related to charge and discharge current,charge and discharge voltage,and ambient temperature.These three direct measurement indicators can be used to estimate SOC and SOH of lithium-ion batteries.In the next part,the traditional neural network has no memory function for the SOC of the lithium-ion battery at the previous moment,and cannot accurately estimate the SOC of the lithium-ion battery.In this paper,a Gated Recurrent Rnit(GRU)neural network is used to estimate the SOC of lithium-ion batteries.Compared with the Recurrent Neural Network(RNN)and the Long Short-Term Memory(LSTM)The SOC estimation results at different temperatures have been improved to a certain extent.However,GRU faces a long-term sequence that will dilute the trend information.This article proposes a method for estimating the SOC of lithium-ion batteries using EMD-GRU.Empirical Mode Decomposition(EMD)is used to decompose the current signal into fluctuating signals,periodic signals and trend signals.The five types of voltage and temperature are combined as the input of the GRU network model,which improves the network model’s response to long-term signals.The simulation experiment results show that this method significantly improves the estimation accuracy of SOC and is suitable for different temperature conditions.In the end,the capacity and internal resistance of lithium-ion batteries is not easy to measure online,so indirect health indicators are widely used in lithium-ion battery SOH estimation.During the discharge phase,battery equipment is randomly discharged,which makes it impossible to extract the health indicators.The initial capacity of the constant current charging phase is different,and it is also not suitable for the extraction of health indicators.Only the constant voltage discharge phase starts and ends and the charging voltage is fixed.It is suitable for the extraction of health indicators.This paper analyzes the constant voltage charging stage curve,and extracts three indirect health indicators of constant voltage charging time,constant voltage charging capacity,and constant voltage charging.The correlation analysis verifies that the health indicators can be used for indirect estimation of SOH.On this basis,a method for indirect estimation of the SOH of lithium-ion batteries based on AGA-BP is proposed.The Adaptive Genetic Algorithm(AGA)optimized by arctan function is used to optimize the global optimization of Back Propagation(BP)neural network model parameters,and establish the relationship between health indicators and capacity,and then estimate the SOH of the battery.The simulation results show that this method is suitable for online estimation of lithium-ion batteries and can estimate the capacity more accurately.