Is there an automatic way of calibrating thermal model towards measured data?

Hello Yes, this can definitely be automated, and Its useful to do so. A few possible options. First of all look at the tutorial "DC Calibration", installed locally. This makes use of an optimiser built into Motor-CAD itself. It allows the user to load in some experimental test results, and the user can set input parameters to vary, along with ranges. The optimiser will aim to minimise the difference between experimental temperature results and the model. For specified temperatures in the machine. We have also looked into combining this approach with surrogate modelling, via Ansys OptiSLang. The advantage here is that a detailed sensitivity analysis is also performed through the surrogate modelling stage. This gives quite a useful insight into the thermal behaviour. However, the simulation time is also reduced, as the fitting stage is performed on a surrogate model. This approach makes it easier to compare the thermal performance of multiple operating points simultaneously. A slide on this was presented at the MDL global user summit, 2021. In "Advanced Design with Motor-CAD: Motor-CAD and optiSLang Best Practices Seminar" (I will try to attach a screenshot if possible). Otherwise a direct optimisation with Motor-CAD and a custom script could be used.