Battery Grouping Technology Research Based On Affinity Propagation Algorithm

With the increasing depletion of fossil fuel resources and people's attention to environmental protection,electric power as a kind of clean energy has been paid more and more attention.The single battery which is with small power and low voltage is unable to meet the needs of the power of high power electric equipment,so it is the main way to solve this problem in the form of a battery pack.However,the inconsistency between single cells has a crucial impact on the performance and life of the battery pack.Based on the research of the current main battery grouping methods,this paper proposed a grouping scheme based on affinity propagation clustering algorithm and battery's dynamic characteristics to improve the performance of the battery pack.This paper introduced a high precision data acquisition device for lithium battery based on STM32F407ZET6.The data acquisition device can sample the voltage,power,temperature and current of the battery,and use the WiFi module to transmit the data to PC in real time.PC software graphical interface can be real-time display data collected by the data collection device,you can also retrieve the historical data of the database,and can draw the curve according to the data.In the experiment,the MATLAB is used to complete the program design of the clustering scheme,and the battery is grouped according to the clustering results.Finally,we chose the multi-parameter grouping method and the method proposed in this paper to compare the experiment.Do the battery cycle performance and storage performance experiments with the battery grouping results.The battery grouped selected by this scheme was used as the experimental group,and the battery grouped based on the capacity and voltage was selected as the comparison group.During the experiment,the battery capacity decay and self-discharge rate of the experimental group and the control group were observed and recorded.By comparing the experimental data of two groups,it can be concluded that the battery storage performance based on the affinity propagation clustering algorithm and the battery dynamic characteristic group is good,the capacity decay rate is slow.