Research On Decision Method Of Lithium Power Battery Grouping Based On Big Data

With the rapid iteration and development of modern science and technology,manufacturing industry and information science have to be widely combined and applied to cope with the complex and changeable manufacturing industry competition and unpredictable market,further improve the level of manufacturing information in China,produce more sophisticated products,and meet the new process standards has become "imminent".In the actual production process,due to the large number of batteries will produce a large amount of process data,in order to solve the fast storage and calculation,the introduction of large data mechanism.Lithium-ion battery pack contains dozens of similar batteries.The disadvantages of single lithium-ion battery,such as low power,low voltage and insufficient power,can not meet the power requirements of electric vehicles and other equipment.At the same time,the inconsistency of the characteristics of single battery will seriously affect the performance and life of the whole battery pack.Therefore,in order to reduce the inconsistency of battery characteristics in the case of a large number of production process data,this paper proposes a reasonable battery allocation decision-making method which can effectively solve the problem of battery inconsistency.The main work and contributions of this paper are as follows:(1)Design and build a large data platform according to the actual situation.According to the actual data volume and application requirements,the overall system architecture is divided into three layers:data access layer,data storage and calculation layer,and large data WEB display layer.The clear framework structure can effectively manage and expand each target.(2)A dynamic Gaussian mixture application model is proposed.Because of the traditional method,the static termination voltage and capacity are used for battery configuration,which does not guarantee the consistency of the internal characteristics of lithium-ion power battery.In this paper,based on the actual production environment of lithium batteries,a method of lithium batteries configuration based on Gaussian mixture model is proposed.Combining the traditional methods with the internal characteristics of batteries,the consistency of batteries can be improved after batteries are assembled,so that the battery life can be improved.Finally,through the actual production standards,the model structure was dynamically adjusted to meet the requirements of lithium battery configuration.(3)An improved K-means clustering algorithm is proposed:due to the limitations of traditional clustering algorithms,and for the lithium battery configuration problem,it can not meet the business requirements of lithium battery configuration consistency.Combining with practice,an improved K-means algorithm is proposed to meet the needs of battery assembly business.The model includes a data preprocessing model and a new battery grouping method based on K-means algorithm.In the battery data preprocessing model,the faulty batteries are removed through the design of process data preprocessing method and actual production standard.In the battery grouping algorithm,the number of batteries in each cluster is equal after the completion of the battery group using this algorithm.This algorithm can not only guarantee the internal characteristics of lithium-ion power batteries,but also ensure that the number of lithium-ion batteries in each cluster is equal after the configuration is completed.