| The development of new energy vehicles is an important strategic initiative for China to response the environmental energy crisis and promote its green development,which also becomes a powerful guarantee to achieve "carbon neutrality" and "emission peak".With the strong support of national policies and market,the new energy vehicle industry has a strong development momentum and broad prospects.As the core energy storage component of new energy vehicles,the performance of the lithium-ion battery system directly affects the power and safety of the vehicle.The battery management system(BMS),as the "brain’ of the battery system,plays an important role for the safe and efficient operation of the battery system.However,the current battery management technology is not yet mature,resulting in the decline of battery power,economy,frequent safety accidents and other issues,which has become a keybottleneck restricting the development of new energy vehicles.In fact,the lithium-ion battery is a complex dynamic system with highly nonlinear,multiple time-varying,and multi-state couplings.Hence,the battery has a complex aging mechanism,unmeasurable internal states,and the characteristics affected by multiple factors,which make the refined and intelligent management of the battery system extremely challenging.The problems of battery internal parameter identification and state estimation,on-line diagnosis of aging mechanism,and battery pack consistency diagnosis have not been fundamentally solved.Therefore,the main research and innovation are as follow:Aiming at the problem of low robustness of online identification of battery parameter in complex and harsh vehicle environment,the measurement performance of BMS under wide temperature range and electromagnetic interference is analyzed.The influence mechanism of measurement noise and voltage and current data asynchrony on the identification process is clarified through the parameter identification sensitivity analysis.A robust least squares identification method with adaptive boundary and bias compensation is proposed.The method can effectively suppress the latency error caused by data asynchrony by adding a robust function with an adaptive boundary and combine the bias compensation algorithm to improve the suppression of measurement noise.The experimental results show that under the actual BMS measurement conditions,the proposed method can significantly suppress the latency error and measurement noise compared with the traditional method and has higher identification accuracy and robustness.The proposed method can effectively improve the accuracy and reliability of battery parameter identification under vehicle conditions.Aiming at the problem that it is difficult to accurately extract the deep aging information of the battery under on-board working conditions,the parameter online identification is combined with an open circuit voltage(OCV)reconstruction method and an aging diagnosis method suitable for complex on-board conditions is proposed to achieve estimation of battery and electrode aging under dynamic operating conditions and different temperatures.The method firstly uses the online identification algorithm to extract the OCV of the battery under the working state of the battery.Then,the aging information implicit in the OCV data is extracted by OCV curve reconstruction and the estimation of positive and negative electrode aging and the estimation of battery capacity are realized.Further more,the objective function is improved and the Peukert coefficients are used during OCV reconstruction so as to improve the adaptability of the proposed method to different temperatures and fragmented data.Experiments have verified that the proposed method can accurately estimate the battery capacity and electrode aging parameters under dynamic conditions and different temperatures,and at the same time realize the identification of different aging mechanisms such as lithium ion loss and active material loss inside the battery.Aiming at the problems of difficulty in quantifying the inconsistency of series-connected batteries and low estimation accuracy of the state of charge(SOC)of battery pack,a consistency diagnosis method based on OCV curve transformation is proposed to realize the quantitative diagnosis of the capacity inconsistency and aging inconsistency of battery pack.Meanwhile,an update method of battery pack model parameters and OCV curve is proposed based on the inconsistent diagnosis results,and the SOC estimation of battery pack is realized based on adaptive extend kalman filter(AEKF).Experiment results indicate that the proposed method can accurately diagnose the SOC inconsistency and aging inconsistency and estimate the SOC of battery pack under different inconsistencies.Compared with traditional methods,the proposed method can maintain a high estimation accuracy under severe inconsistency.In view of the lack of single information in the parallel module,it is difficult to carry out consistent diagnosis,the experimental studies are first carried out which shows that the current distribution of the cell inside the parallel battery pack and the terminal voltage of the battery pack are directly related to the battery consistency.To this end,a method for estimating the current distribution of parallel modules based on long short-term memory(LSTM)neural network is proposed.The proposed method uses the LSTM network to learn the inconsistent features of parallel battery packs and estimate the current distribution without adding BMS hardware.It is verified by experiments that the proposed method can accurately estimate the branch current distribution by using the charge/discharge curve of the battery pack under the condition of inconsistent impedance and aging.Based on the estimation results of the proposed method,the consistency of parallel battery packs can be effectively diagnosed,thereby avoiding battery pack damage or safety issues caused by inconsistent faults.In summary,in order to realize the safe and efficient management of lithium-ion battery systems and based on actual needs,this paper has made breakthroughs in battery parameter identification,aging diagnosis and SOC estimation,and consistency management of series/parallel module under complex working conditions.Importantly,a set of practical and efficient theoretical methods is proposed to provide theoretical and methodological support for refined and intelligent management of lithium-ion batteries. |
