| Under the dual constraints of energy shortage and environmental deterioration,in order to achieve the goal of energy conservation and environmental protection,electric vehicles have been unprecedented development.Lithium-ion battery is widely used in electric vehicles due to its advantages of high energy density,long cycle life and environmental friendliness,and the number of installations is also rising rapidly with the potential.Lithium-ion batteries will gradually deteriorate with the increase of service time.When they no longer meet the performance requirements of electric vehicles,they need to be replaced to ensure the safety and reliability of vehicles.With the rapid development of electric vehicles,the number of retired batteries has also grown rapidly,and the recycling of large-scale retired batteries has become a challenging problem.At present,the research on the echelon utilization of retired batteries is still in the initial stage,and the technology development is not mature enough.On the one hand,the state of health(SoH)of retired batteries determines their echelon scenario.Traditional battery SoH estimation methods are mostly aimed at pre-retirement in-vehicle applications.Currently,SoH estimation methods for retired batteries have low accuracy and poor generalization ability.On the other hand,the performance parameters of retired batteries have strong inconsistency,which makes it difficult to directly assemble and utilize,and the current consistency sorting methods for retired batteries are lengthy and inefficient,or have a single index and poor results,which increases the cost of echelon utilization and also restricts the development of the echelon utilization industry.This paper addresses the key issues of retired batteries in echelon utilization,proposes a health state estimation method for retired batteries,and establishes a rapid sorting and consistency comprehensive evaluation model for retired batteries based on a large amount of experimental data,the main contents of which are as follows:Firstly,the aging characteristics of the battery are analyzed.The battery test platform is built and the test scheme is designed with LiNCM and LiFePO4 batteries as the research objects.Based on the high-quality test data,the aging characteristics of the battery monomer and each battery module in the series battery pack are analyzed,and the aging inconsistencies of the batteries were analyzed in three aspects:capacity,internal resistance and open-circuit voltage,which laid a solid foundation for the subsequent research on the echelon utilization of retired batteries.Secondly,a SoH estimation method for retired batteries based on twin support vector(TSVR)regression machine is proposed to address the problems of low accuracy and poor generalization ability of SoH estimation of retired batteries by traditional methods.Eight groups of LiNCM retired battery modules are used as the research object,and 97 retired battery cells are obtained by disassembling and grinding.The capacity distribution of retired batteries under different operating conditions was analyzed by experimental data.The coordinates of the peak point of the IC curve and the time interval of 40mV equal voltage in the peak region are extracted as input features,and the corresponding capacity is taken as output to train the SoH estimation model.The estimation effect of the retired battery SoH estimation method based on TSVR proposed in this paper is verified through experiments,and the maximum relative error of SoH estimation is within 1.3%under different operating conditions,and the average relative error is less than 1%.Based on the SoH estimation results,the preliminary sorting of retired batteries can be carried out for applications in different echelon utilization scenarios.Then,for the variability of the initial state of the mass retired batteries leading to the reduction of sorting efficiency.Using LiNCM retired batteries as the research object,characteristic charging experiments with different initial voltages were designed,and the capacity distribution of 97 retired batteries was analyzed.A fast sorting scheme for retired batteries is demonstrated,and a support vector classification machine(SVM)is used to train a sorting model for retired batteries by extracting the aging characteristics of some characteristic charging curves,and the SVM parameters that may affect the sorting results of the model are optimized by particle swarm algorithm(PSO).It is verified that the optimized fast sorting model of retired batteries has good sorting effect under three different initial voltages,and the average correct rates of 50 times sorting are 86.8%,94.5%and 93.8%,respectively.Finally,in response to the problem that the single evaluation index of different echelon utilization occasion matching groups is too one-sided,the advantages and disadvantages of common comprehensive evaluation methods are analyzed and compared,and a comprehensive evaluation model of retired battery based on CRITIC-Grey Relation Analysis(GRA)is established.The multi-parameter evaluation system is selected and the evaluation results of the model are verified by actual retired battery data.It is found that the comprehensive performance of retired batteries with higher capacity is not necessarily of higher quality in the same capacity category.This demonstrates the necessity of comprehensive evaluation of retired battery performance.After completing the above rapid sorting,a suitable evaluation parameter system is selected for different gradient utilization occasions,and the retired batteries in the same category are comprehensively evaluated for consistency,and the batteries with similar correlation in each category are regrouped.It is verified that the consistency of charge/discharge curves of the regrouped batteries after sorting and evaluation is greatly improved,and the standard deviation of charge/discharge voltage of the three categories of retired batteries is reduced by 2-13 times compared with that of the non-sorted evaluation categories. |
PDF Full Text Request