Analysis and prediction of temperature of an assembly frame for aircraft based on BP neural network

Temperature is one important factor which decides the assembly accuracy and reliability of the aircraft. Especially with the development of large-sized aircraft, the small temperature change in the aircraft assembly will result in a large displacement deviation and non-negligible internal stress. Therefore, it's crucial to characterize and predict the temperature during the assembly process of aircraft. Selecting one type of assembly frame for aircraft wing as the study object, temperature in this structure was measured and recorded for one year. Based on the measured data, a model of an optimized BP neural network is proposed to analyze and predict the temperature distribution. The trained temperature model shows a good result with a relative error of 1% and an absolute error of 2 degrees C. Finally, the displacement of assembly frame is obtained from the temperature distribution.