Modeling And Parameter Estimation Of Retired Lithium-ion Power Battery Based On Capacity?resistance And The State Of Charge

Abstract:Lithium-ion power batteries retired from electric vehicles (or retired batteries for short) still possess high residual capacities as high as80%of the designed capacities. It is of great significance to reuse retired batteries in a suitable way which could reduce the energy cost, relieve the environmental pollution and make waste profitable. In order to make the best use of the residual capacity of the retired batteries, it is necessary to classify the retired batteries, as well as predict the cycle life, estimate the deterioration degree and calculate the state of charge (SOC) of the selected batteries. This is also one of the attractive and hot topics in the field of the recycle and gradient utilization of the retired batteries. In this thesis, the electrochemical test-computer simulation technology were employed to investigate the electrochemical behavior of the retired batteries in different temperature, discharge rate and the depth of discharge (DOD). Then the general models for the life prediction and the deterioration degree estimation of the retired batteries were established. In addition, the SOC estimation model of the retired batteries was developed to ensure the batteries in good working order. Through the in-depth study of the above contents, four main conclusions were obtained as follows:(1) The modern physical and electrochemical test methods were used to examine the safety of the retired batteries. A classification method based on the appearance-capacity selecting method, the voltage-resistance selecting method and the discharge curve selecting method was introduced to select the required retired batteries. Results showed that the batteries with integrated pack, the capacity of1.1?1.0Ah, the resistance of12?12.6m?, the open circuit voltage of3.2998?3.3002V, and the good repeatability of discharge curves at large discharging rates would be considered and reused.(2) The battery life prediction model and match detection method were built, according to the electrochemical testing-computer simulation technology. Electrochemical testing results showed that the number of cycles (N), environmental temperature (T), discharge rate (C) and DOD were the main factors that influence the battery cycle life. With the increase of N, T, C or DOD, the capacity fading of the battery speeded up. When the cycle time was long enough, the influence of DOD could be neglected. Computer simulation analysis indicated that power function model was appropriate to describe the capacity loss for all conditions: Where, cn is the constant.According to power function model, the working condition and the residual life of any batteries could be successfully predicted with a few electrochemical testing data.(3) The battery degradation model was set up by combining the internal resistance measurement with the alternating-current (ac) impedance test method. The results showed that the battery internal resistance can be used to characterize the degree of the battery degradation. With the increase of the internal resistance, the batteries deteriorated seriously. The ohmic resistance and polarization resistance enhanced with the increase of T, C or DOD. T was the most important factor that affected the ohmic resistance and polarization resistance of the retired batteries. C is less important than T, and the least is DOD. The resistance model about the sum of the ohmic resistance and polarization resistance can be expressed by power function model: Where, a, b, c are the constants.(4) Ac impedance measurement, SOC-OCV curve and pulse discharge method were combined and employed to build the SOC estimation model. Ac impedance measurement results showed that both the ohmic resistance and the polarization resistance should be considered when modeling at the full state of charge. While only the polarization resistance should be concerned when modeling at the range of SOC=20?80%. SOC-OCV test results demonstrated that SOC-OCV characteristic of the batteries has nothing to do with T, C, storage time and the charge/discharge status. The value of OCV ascended monotonically with the increasing of SOC, and declined with the decreasing of SOC. multistep pulse discharge test results indicated that the OCV-SOC model could be determined by equivalent circuit model so as to estimate the value of SOC.