BIMtoSAM.AI: Building Information Models (BIM) to Structural Analysis Models (SAM) Using Artificial Intelligence (latest version 1.0)
BIMtoSAM.AI is a new AI-enabled application that creates FEM models from descriptions of walls. The input is a wall description contained in a building information model (BIM). The output is a structural analysis model (SAM) for input into FEM analysis software.
The pre-trained neural network takes BIM features (e.g. geometry and material properties) as input and gives SAM features as output. The SAM features include mesh size and parameters of constitutive models to generate the FEM model. The neural network was trained on the SimCenter’s wall dataset, which contains experimental and simulation data. This dataset is available on DesignSafe Data Depot. This application and the pretrained neural network are incorporated into the SWIM application.