To improve the simulation resolution and accuracy in thermal analysis of spaceborne electronic devices and the temperature control performance of passive thermal control devices, a system multi-scale model was established to obtain the temperature field and heat flux of electronic devices inside the satellite at different scales as schematic in the below figure. The temperature fluctuation mechanism inside the satellite was analyzed at different physical scales. The thermal analysis resolution of spaceborne electronic equipment was improved, and a method to reduce the power fluctuation of spaceborne equipment was proposed based on the results of system multi-scale thermal analysis.
The results show that the system multi-scale model presents an accuracy deviation below 8% from the actual model. However, the system multi-scale model saves 99.67% of the mesh generation time, which greatly improves the computation efficiency. The system multi-scale model can capture the thermal information of device-level chip microstructures with less computational cost. The system-level model can evaluate the temperature control and insulation performance of passive thermal control materials from a macroscale. The temperature fluctuation amplitude of the platform compartment was 7.95 K, while the temperature fluctuation amplitude of the load compartment was reduced to 2.43 K after the temperature control of the composite phase change insulation material, which was 69.43% lower than that of the platform compartment. Compared with traditional vacuum insulation panels, the composite phase change materials are more superior in controlling the temperature of the chamber and suppressing temperature fluctuations. The temperature fluctuation signal after insulation by the composite phase change insulation materials shows a characteristic of shifting to the high-frequency domain. After selecting the cabins that require key insulation and temperature control through multiple regression analysis, a simplified model at device level was employed to obtain temperature field under different thermal control device layouts as a training dataset. A neural network genetic algorithm was used to predict the optimal installation position of passive thermal control devices at the device scale and a thermal control layout scheme was obtained that reduces the maximum temperature fluctuation of the devices by 2.74 K. If the temperature uniformity coefficient is taken as the optimization goal, the temperature of each device on PCB board can be reduced to 14.39% of the average temperature of all devices through optimization.