Thermal Surrogate Model Based on Temperature Trend Superposition

Thermal analysis is a critical component of spacecraft mission operations, but it often poses significant challenges due to its high computational cost. For instance, in the Mercury exploration mission BepiColombo, the operational planning of the Mercury Magnetospheric Orbiter (MIO), led by JAXA, is constrained by thermal limitations, creating a planning bottleneck. This requires repeated iterations of thermal analysis and temperature validation, but optimizing the plan remains difficult due to computational constraints. To address this issue, a thermal surrogate model that predicts transient temperatures using multiple pre-defined temperature datasets was proposed. The prediction accuracy was evaluated by comparing it with the results from MIO’s detailed thermal mathematical model. The proposed surrogate model reduced analysis time from 100 minutes to less than 30 seconds while maintaining good agreement with the actual transient thermal behavior. This study outlines the methodology for building the surrogate model and presents key findings on its accuracy and computational efficiency.