Modeling of Temperature Control System of Space Experiment High-temperature Furnace Based on XGBoost

With the development of China’s space industry, the construction of China’s space station has been completed in 2022. In the future, China will carry out a series of space material science experiments in space. The high-temperature furnace in the high-temperature material science experimental rack, as the main equipment of the space material science experiment, requires the high-temperature furnace’s temperature to be stable within ± 0.25℃ when conducting the high temperature material science experiment in space. In the face of such high temperature stability requirements, in order to ensure that the scientific experimental system of the high-temperature material science rack can successfully conduct the space material science experiment, it is necessary to first establish the mathematical model of the high-temperature furnace control system. Because the object of high-temperature furnace is a kind of nonlinear and time-delay complex control object, it is difficult to model based on mechanism. To solve this problem, this paper proposes a new solution: based on the experimental input and output data, an intelligent modeling method is adopted to determine an internal equivalent model of the high-temperature furnace control system, which provides a basis for obtaining control parameters that meet the experimental requirements. In this paper, the control system of high-temperature furnace is regarded as a black box model, and four representative sample experimental data are selected. Based on XGBoost method, the mathematical models of temperature zone 2 and temperature zone 3 control system of high-temperature furnace are established respectively. The accuracy of the models can all reach more than 99.98%. Compared with the traditional modeling method, the transfer function is used as the basic model for parameter estimation, and the modeling effect varies according to different samples. In addition, under the best performance of traditional methods, the accuracy of the model based on XGBoost is still improved by 3.8%. The experimental results show that the modeling effect of high-temperature furnace control system based on XGBoost method is good, and the model provides important support for obtaining control parameters to ensure high stability of space experimental temperature.