The Research Of Online SOH Estimation Of Power Battery In Electric Vehicle

With the rapid outbreak of the energy security and environmental protection problem, and with the universal popularity of environmental protection, the future field of motor vehicle will transform to the "low carbon", "low noise", "low pollution" and "zero discharge" vehicles such as the electric vehicles. Up to now, electric vehicles have become the main development direction of modern cars.Currently, each electric car has a common problem:the life of the power battery is not long. And due to the cost of power batteries is the largest one in an electric vehicle, the lifetime of power batteries for electric cars has a critical influence. The State of Health (SOH), as an important parameter of describing the battery life, means a lot to prolong battery life.By the research of SOH, SOH of a battery has a nonlinear relation between battery parameters such as internal resistance, open circuit voltage, capacity of battery. As the value of SOH could not be measured accurately, and the internal resistance of the battery, the polarization effect of battery and some other parameters such as battery capacity only reflects one aspect of the definition of state of health.The main research content of the thesis is as follows:1) Deep analyzing the working principle and the advantages and disadvantages of lithium-ion batteries, summarizing the battery life degradation principle of lithium-ion batteries by the perspective of internal factors and external factors respectively, to get the relationship between the SOH of battery and several characterization parameters. Respectively analyzing the electrochemical model, empirical model and the equivalent circuit model of lithium-ion batteries, and proposing a way to identify the internal resistance of lithium-ion batteries based on the least squares algorithm.2) Considering the basic principle of neural network algorithm, the online real time estimation of SOH of battery algorithms based on t-s fuzzy neural network and Elman neural network respectively have suggested in this thesis. The simulation model of this two algorithm, respectively, has been established by Matlab. The input of the model is the internal resistance of the battery, the charging current of the battery, the temperature of the battery. And the output of the model is the SOH value. The sample data of the continuous charging and discharging test of the cell cycle have carried on the training to model and verify the accuracy of both methods. The accuracy of both two algorithms can reach 5%, but the algorithm based on Elman neural network is battery than the algorithm based on T-S fuzzy neural network algorithm.3) Designing and developing the on-line real time monitoring system of a powerful battery of an electric vehicle with the development platform of Net 3.5, database management system of SQLSERVER 2008 and development language of C#. The software is built on a three layer architecture models and could display battery real time data and online prediction of the SOH of battery.