A System-Level Spacecraft Thermal Model Reduction Method Applicable to Transient Analysis



Journal Title

Journal ISSN

Volume Title


2023 International Conference on Environmental Systems


In spacecraft thermal design, Thermal Mathematical Models (TMMs) play an essential role to evaluate thermal performance in-orbit. However, typical detailed TMMs have a high calculation cost, while many iterations running these TMMs are required in thermal design processes, such as uncertainty analysis, feedback from hardware design, and model correlation using thermal vacuum test data and in-orbit data. Therefore, shortening analysis time for TMMs is essential to reduce time and cost required for thermal design. Creating reduced models is one approach to deal with this problem. Generally, there are two approaches in creating reduced models: model-based and data-driven. Past studies on spacecraft thermal design reduction methods mainly focus on either component-level, model-based reduction or system-level, data-driven reduction. This study presents a model-based reduction method that is applicable to system-level thermal models, based on connected component decomposition algorithms. To make decision of which nodes to be fused in an automatic process, analysis results from the original model, under typical case sets, are used. By this approach, the reduced model preserves physical properties, allowing it to be used in both static and transient cases, and realize application to hardware design, thermal vacuum test correlation, and software-in-the-loop simulations for operation testing and planning. For evaluation of this reduction process, the actual thermal model used for ONGLAISAT, a 6U CubeSat developed by the University of Tokyo, was utilized. Even for CubeSat level models with only less than 1,000 nodes, this reduction process showed a reduction rate of around 0.6 to 0.7, shortening analysis time, while displaying a maximum of just several degrees celsius in temperature difference between the original model and the reduced model. This displayed that an effective system-level reduced model can be created using this method.


Toshihiro Shibukawa, ArkEdge Space Inc., Japan
Shinichi Nakasuka, ArkEdge Space Inc., Japan
ICES207: Thermal and Environmental Control Engineering Analysis and Software
The 52nd International Conference on Environmental Systems was held in Calgary, Canada, on 16 July 2023 through 20 July 2023.


Spacecraft Thermal Design, Thermal Modeling, Model Reduction