Mid-wave infrared (MWIR) and long-wave infrared (LWIR) optical systems often employ cryogenic cooling to suppress background noise by reducing the system's self-emission. In these cooled optical systems, thermal deformation of structural materials, changes in the refractive index and curvature of optical materials, and spatial geometric position vectors constitute the fundamental mapping data connecting the physical entity and its digital twin model. This paper discusses the digital twin modeling of a cooled infrared optical system with an F-number of 3.3, a spectral range of 3.5-5 μm, and six field-of-view stitching configurations. Compared to conventional systems, the digital twin of this system incorporates a mapping process from ambient temperature and pressure to a vacuum cryogenic environment. Experimental results indicate high mapping accuracy between the physical entity and the simulation model at room temperature. However, the testing results of the physical entity in a cryogenic environment show significant deviations from the predictions of the cryogenic simulation model, and the mapping relationships of the six optomechanical systems are scattered with considerable uncertainty. Key factors contributing to the instability of the digital twin model include errors in the original data and inaccuracies in the transformation functions from ambient to cryogenic environments. Finally, this paper discusses feasible methods to improve the mapping accuracy of the digital twin model.
Chou et al. (Mon,) studied this question.