Research On Life Prediction Of Power Battery

With the acceleration of urbanization and industrialization throughout the world,the consumption of global energy is rising sharply while emissions of carbon dioxide and pollutants are increasing annually,which worsen the problem of global warming and the environmental issues.New-energy vehicle,which is one of the main means of transportation,has great advantages in environmental protection and resource utilization.It's now developing at a rapid speed and would become the most potential industry in the future.As the critical component of electric vehicles,power battery is the main power source and its own performance can affect the electric vehicles greatly.In recent years,in the pursuit of high power,high energy and high-safety,lithium-ion batteries with many advantages have become the best in battery industry,and have also become the most popular research object.However,after a long period of recycling,the battery will suffer from performance degradation and capacity decline,which would eventually lead the battery life to an end.Therefore,in order to ensure the safety and reliability of battery and avoid unnecessary economic losses and security incidents,the residual capacity and cycle life of the battery require further forecasting research.Aiming at the life prediction of lithium-ion batteries,this paper conducted the following main research work based on the public battery data that provided by NASA PCo E and my personal independent experiment on the cycle performance of Li Fe PO₄batteries:1.In this paper,the structure,characteristics,working principle and basic performance parameters of the lithium-ion batteries were introduced,and the materials of several basic components were analyzed in details,as well as their impact on the performance of lithium-ion batteries,which provided a theoretical basis for the subsequent research.2.In this paper,the decay mechanism of lithium-ion batteries was analyzed,and the factors affecting the decay of battery life were studied and summarized in two parts including internal and external factors.Then,the battery capacity was selected as the electrochemical index to evaluate whether the battery life was at an end,and the battery life was also predicted by this index.3.Based on the problem of poor adaptability of battery experience degradation model to single battery,this study adopted two data-driven approaches to forecast the battery life,one is grey model prediction method generated by grey system theory,the other is ARIMA model prediction method in time series analysis method.Then the experimental data and simulation results were compared to verify the accuracy of the two methods.4.As the number of charge and discharge cycles increases,the capacity of lithium-ion batteries would show a strong nonlinear degradation.In terms of this issue,the using of LSTM algorithm was proposed in the paper to track and predict the battery degradation process.As a result,the LSTM algorithm could express the nonlinear decline trend of the battery and showed a good prediction ability for battery capacity data on life prediction.However,the LSTM algorithm still presented a problem that the early prediction was not accurate enough,then the particle swarm optimization algorithm was also proposed in this study to optimize the LSTM algorithm,which avoided the blindness of artificial parameter selection.Finally,the improved algorithm was used for simulating calculation.Then by compararing the simulation results with the experimental data,the results showed that the improved algorithm can effectively improve the prediction accuracy of battery cycle life.