Research On The Group Application Technology And The State Parameters Estimation Strategy Of The Lithium-ion Battery

The vigorous development of the new generation green batteries which are represented by Lithium-ion battery,is playing an important role for promoting the researches on the technology about the new energy as well as helping to solve the problem of energy crisis and environmental pollution.The continuous improvement of the related technologies such as group application of battery and battery pack state management would prolong its service life and improve the reliability of the energy storage system.In view of this,the dissertation takes Lithium-ion battery as reaserch object,and pays attention to the two of hot issues about the battery and mainly studies on the battery group design and application technology,the method of battery or battery pack modling,system identification and state of charge estimating.The method of the dissertation could be used to other type batteries.The main research and work of the dissertation could be summarized as follows:(1)The performance of the battery pack would deteriorate over time due to the inconsistency of the cell in the group application.To slove such problem,a complete design solution of the group application of Lithium-ion batteries in the typically harsh environment is proposed and realized,upon studying and analyzing the major factors that influence the performance of the battery pack.The dissertation further researches the charging and discharging management technology of the high power battery pack and then provies a sectionalizing and equalizing charge strategy based on protection of the single cell and its implementation process.With the sectionalizing and equalizing charging for the batteries,the negative effect during charging resulting from the external force such as grid abnormity would be reduced,and the charge and discharge capacity of the batteries and the reliability of the whole power sources system are improved.(2)System parameter identification strategies based on intelligent algorithms including Particle Swarm Optimization(PSO)and its extensions are studied.However,for PSO,the particles are usually unable to find the optimal solution or hovers around the optimal solution but can't locate in it.The Random Disturbance PSO(RDPSO)algorithm is proposed aiming at solving such peoblem and its performance is validated with four common used target optimization functions.The dissertation adopts above mentioned RDPSO to identificate the system parameters of the Equivalent Circuit Model(ECM)and presents a weighted fitness function suitable for the system parameter identification.The experiment results with two different test profiles indicate that the proposed RDPSO algorithm has the feasibility and universal applicability.(3)The path dependence of open circuit voltage(OCV)is a distinctive characteristic of Lithium-ion battery which is termed as OCV hysteresis.Accutate estimation of OCV hysteresis is essential for ECM.The dissertation uses the feedback information,that is the difference between the simulation result of the ECM and the terminal voltage of the battery,to correct the value of the OCV,and the method which dynamically tunes the value of OCV is proposed upon analyzing the movments of the corrected OCV for the four stages of charging,discharging,rest after charging and rest after discharging.The enhanced modle which combines the said OCV tuning to the second order modle could achive higher precision.(4)For the requirements of the system-level simulation and the battery management system,the modeling methods of battery pack to simulate its dynamic circuit characteristics are studied.It has been proved that the inconsistency of the single battery and the non-linear proportional relationship between the single battery and the battery pack are automatically included in the behavior of the batteries.By using such characteristic,the batteries level model can be built directly.Upon studying that how do the temperature and the current rate influence on the available capacity of the battery pack,the compounded battery pack model based on the equivalent circuit model as well as considering the effect of the temperature and the current rate is built and such model would be integrated into the larger system level simulation environment,such as the selection and group design of the battery pack for large scale energy storage systems and so on.(5)The Extended Kalman Filter and Unscented Kalman Filter estimation methods are studied and a novel UKF-EKF joint strategy for the state-of-charge on line estimation is proposed upon analyzing the advantages and disadvantages of the EKF algorithm and UKF algorithm.The experimental results show that the UKF-EKF algorithm could combine with the advantages of two said methods.On the one hand,compares with the EKF algorithm,it improve the robustness,on the other hand,compares with the UKF algorithm,it would reduce the execution time and with the comparable estimation precision,which makes it is more suitable for on-line SOC estimation.