Reduced-order modeling for spacecraft thermal-structural applications

Reduced-order modeling has been successfully implemented for a broad range of spacecraft thermal analysis applications. Leveraging the speed of reduced-order models (ROMs), thermal analysis teams have access to rapid optimization, sensitivity studies, and new approaches to model correlation, to name a few. These methods have been successfully applied for such applications as the Mars 2020 Helicopter and Dream Chaser programs. In addition to the speed of the ROMs, an advantage of the developed approach is its relative robustness. Since the method is merely based on sampling and data-fitting, it is not directly tied to a specific analysis tool and can be readily applied to others. With that in mind, the ROM approach was expanded to include structural models leading to combined thermal-structural analyses.

In this paper, we will outline the methods used to implement reduced-order modeling into a structural analysis workspace, specifically Nastran. We will highlight our approach for implementation and discuss how structural ROMs can be robustly built for various Nastran platforms (e.g., MSC Nastran, Autodesk Nastran, and Simcenter Nastran). We will then provide several example cases showing the utility of ROMs for structural analysis. We continue the discussion by examining how structural ROMs can be combined with thermal ROMs for thermal-structural applications. Here we will pair Thermal Desktop and Nastran models to create these multi-disciplinary ROMs. We will conclude the overview by providing several thermal-structural use cases.