Research On Design And Thermal Comfort Prediction Of Vehicle R&D Data Management System

With the continuous development of the automotive industry,the research and development data of automobiles is rapidly increasing.In response to the problems of low data management and utilization efficiency,high maintenance difficulty,poor data sharing timeliness,and insufficient security and confidentiality in the entire vehicle research and development process of a certain automobile enterprise in Guangxi.Against the backdrop of China’s vigorous development of digital management,in order to achieve more reasonable and effective management of research and development data,based on the production and research and development needs of the enterprise,a set of research and development data management system with the whole vehicle as the main body has been designed and implemented.The main research content is as follows.(1)By analyzing the current situation of enterprise data management,combined with the development needs of automotive companies and the characteristics of vehicle data types,the mvc architecture mode of the Odoo software development platform and the characteristics of the official designated database Postgre SQL were analyzed,laying a theoretical foundation for the design of a vehicle research and development data management system based on the Odoo development tool.(2)A data management system for vehicle research and development was designed based on the Odoo development tool.According to the entire vehicle development process,the system has designed a vehicle management module,weight management module,component management module,general layout parameter management module,and user management module.According to the functions of different modules,the requirements of each module are analyzed and the structures of different modules are designed,and the migration of Odoo is completed on the server.Following the principle of modularity,the system deployment was completed.Save the data in the form of fields to the database through the Object Relational Mapping(ORM)framework,and implement data statistics and analysis in the background,improving the efficiency of data utilization.At the same time,in order to facilitate the design and development of high-performance air conditioning for the entire vehicle,a thermal comfort analysis interface has been designed in the vehicle management module.(3)In response to the problems of mutual iteration,computational complexity,and difficulty in real-time prediction among the parameters of the Predicted Mean Vote(PMV)thermal comfort evaluation index,RBF Neural Network(RBFNN)was used to predict the thermal comfort index.Due to the performance dependence of the RBF neural network thermal comfort prediction model on the selection of parameters,the CS algorithm(Cuckoo search,CS)was used to optimize the center point and width parameters of the basis function of the RBF neural network,and compared with the PSO algorithm(Particle swarm optimization,PSO).The results showed that the CS algorithm had better optimization performance.Finally,the impact of the hidden layer of the RBF neural network on the prediction performance was analyzed.(4)Due to the poor predictive performance of the RBF neural network model optimized by the CS algorithm and the difficulty of model training,an improved CS algorithm optimized Support Vector Regression(SVR)model was used to predict PMV.The adaptive step size strategy is adopted to adjust the step size of Lévy’s global random walk,and the DOA survival strategy is used to replace the CS algorithm’s preference for local random walk behavior to improve the CS algorithm.We compared the predictive performance of single strategy and dual strategy improved CS algorithm in optimizing SVR models.The results show that the CS algorithm optimized SVR model with dual strategy improvement has the best PMV prediction accuracy and prediction effect.