Research on dynamic prediction method of temperature environment of unmanned aerial vehicle platform based on physical information and neural network

Abstract

During the flight of an unmanned aerial vehicle, the temperature environment inside the cabin is very harsh, resulting in frequent malfunctions of airborne equipment. Therefore, it is necessary to predict the temperature environment experienced by airborne equipment during its life cycle in the early stages of design to support the formulation of design requirements, determine test verification conditions, used and maintained. In order to accurately predict the temperature environment of the unmanned aerial vehicle platform, this paper proposes a prediction method that integrates physical information and neural networks. In view of the thermal environment characteristics of unmanned aerial vehicle platform, the physical mechanism of the formation and transfer of the unmanned aerial vehicle thermal environment was analyzed. Based on the physical mechanism equations and combined with the measured data of unmanned aerial vehicles, an Elman neural network model was established to achieve high flight speed, speed, etc. parameters are characterized as a function of the temperature inside the unmanned aerial vehicle cabin, and finally the model is verified. The results show that the temperature environment prediction model integrates physical information and neural networks can accurately predict the dynamic changes in the temperature inside the unmanned aircraft cabin. The maximum error of the model is 2.76°C, which meets engineering needs.