Research On Thermal Comfort System Algorithm Of Automobile Air-conditioning Based On Machine Learning

The automotive air-conditioning system is the core system to ensure the thermal comfort of the automotive.With the growing needs of the people for a better life,automotive air-conditioning system is facing new challenges and needs to achieve more refined,more personalized thermal comfort settings.At the same time,with the industrial upgrading and the national strategy of "Made in China 2025”,intelligent automobile air-conditioning has become one of the research priorities in the automotive field.In this paper,the related research work on the automative air-conditioning thermal comfort system algorithm is carried out.Based on the analysis and modeling of the big data of real intelligent and connected vehicles,a series of new feature engineering methods and thermal comfort system prediction models are proposed.The main research work is as follows:1.The construction and research of the features of automotive air-conditioning thermal comfort system based on feature engineering.The automotive air-conditioning thermal comfort system features are constructed through the theoretical methods of business logic analysis,feature linear combination,feature statistics and so on.It builds the foundation for the automotive air-conditioning thermal comfort model and also provides systematic guidance for the feature engineering of the data-driven automotive air-conditioning system thermal comfort algorithm model.2.Design and research of general model of automotive air-conditioning thermal comfort algorithm.A general automative air-conditioning thermal comfort intelligent algorithm model is built based on the big data training,by logistic regression,decision tree,support vector machine,GBDT,random forest,XGBoost and other algorithms.The model can automatically set the on-off,temperature,air volume and air blowing mode of the air-conditioning system based on the environmental conditions outside and inside the vehicle,which can dramatically reduce the operational complexity of the usage of the automative air-conditioner and improving the user experience of the automative air-conditioner.The weighted kurtosis clustering algorithm is proposed based on this model,and the theory of automative air-conditioning thermal comfort algorithm based on partition planning mode is developed.3.Design and research of personalized model of automotive air conditioning thermal comfort algorithm.The memory system algorithm is proposed.The actual model is generated through the use of each user’s data,the model dynamic start threshold is adjusted timely through the threshold dynamic adjustment algorithm,in order to to achieve both the stability of the system and the satisfaction of the user’s personalized thermal comfort preferences.Based on the big data from automobile air-conditioner users,through a variety of machine learning algorithms,this paper designs and develops a data-driven,more refined,more personalized,more automated,higher performance automotive air-conditioning thermal comfort control algorithm,realizes the cold-start-free general-purpose system algorithm for predicting the thermal comfort of automobile air-conditioning with high precision and the personalized control algorithm for predicting the thermal comfort of automobile air-conditioning with the ability to provide different personalized settings to different users and highly personalized,realizes automatic setting of automotive air-conditioning on-off,the temperature of both main and auxiliary temperature zone,air volume,the internal and external circulation,the air blowing mode.