| Overcharge,as one of the main abuse forms of lithium-ion batteries,has become a research hotspot in recent years.For the research on overcharging,most scholars at home and abroad focus on the stage from continuous overcharge to thermal runaway.Limited by the requirements of national standards,this situation is rare,and the more common one is slight overcharge.However,the damage mechanism of slight overcharge is not clear at present,and the research on fault diagnosis of slight overcharge is relatively less.Therefore,this paper established an electrochemicalthermal coupling model of lithium batteries to study the electrochemical and thermal characteristics from normal charging to slight overcharge,and proposed an overcharge multistage warning strategy based on the threshold method.In addition,it was found through experiments that the battery capacity degradation rate increased during the slight overcharge cycle.Therefore,this paper used LSTM neural network to build a capacity accelerated degradation prediction model.Based on the difference in the capacity degradation rate of a slight overcharged battery and a normal battery during the full life cycle,a battery early warning algorithm for slight overcharge failure was proposed.The specific research contents of this paper are as follows:(1)A three-dimensional electrochemical-thermal coupling model for the period from normal charging to slight overcharge of lithium-ion batteries was established.The model was composed of four sub-models at different stages,and the accuracy of model was verified by comparing with the experimental data of slight overcharge.(2)Based on the validated electrochemical thermal coupling model,the temperature distribution and heat source of the battery were analyzed.Combined with the changes of temperature curve,voltage curve and heat generation,a multi-level early warning scheme for slight overcharge was proposed.The normal charging to slight overcharge was divided into three areas,and the moment when the reaction between deposited lithium and electrolyte began was taken as the key point for early warning.Parametric analysis of the electrochemical-thermal coupling model was carried out to explore the influence of different battery parameters on the safety of the battery during slight overcharge.It was found that appropriately decreasing the positive electrode thickness and the resistance of SEI effectively reduced the maximum temperature of the battery at the end of the slight overcharge and reduced the risk of the battery.(3)Taking the cut-off voltage as a variable,the slight overcharge cycle experiments of lithium battery under different working conditions were carried out.Incremental capacity analysis and electrochemical impedance spectroscopy(EIS)were used to analyze the damage mechanism of Li ion batteries.It was found that the main aging mode of slight overcharged batteries with charge cut-off voltage between 4.3V and 4.4V was loss of lithium inventory.(4)A two-layer LSTM network model was built to realize accurate prediction of battery capacity under slight overcharge cycle conditions.Based on the difference of the capacity degradation rate between the battery under slight overcharge abuse and the normal battery,the root mean square error value between the capacity of each prediction interval and the offline capacity of the corresponding interval of normal battery was calculated.The results of three consecutive comparisons with the selected threshold were used as the basis for fault diagnosis,and the recognition of the battery suffering from micro overcharge abuse was realized in advance. |