Known issues#

Review known issues for the SimAI Pro application, organized by workflow phase.

Volume training data not supported#

Building a model to predict volume fields is not supported.

Reference Sample#

Each project needs one reference sample. By default, the first valid sample in the selected directory of a project is set as its reference sample.

Domain of analysis#

Domain of analysis can just be set with auto setup and cannot be adjusted manually.

Output variable#

Output variables are the field results that the model is trained on to predict them. They must be present as vector data in all data added to your project. Scalar training output is not supported, but global coefficients evaluated on predicted field results can be configured.

Missing Postprocessing Variables#

Pure postprocessing variables that might be required to compute global coefficients are not supported.

Global Coefficients#

A global coefficient is a scalar value calculated for an entire design based on the predicted field result. It can help to extract further scalar information like maxima or integrated results across the field to support engineering decisions.

Filter functions for the calculation of the global coefficients are not supported.

Prediction#

Using the SimAI Pro application, you only can predict one model at the same time. Each prediction will override the previous one. To preserve predictions, you can save them locally in a directory as .vtp file.