"Backend Components" are advanced computational tools designed to streamline data organization and extraction, augment CFD inflow conditions, enhance inventory analysis, and perform probabilistic estimations related to losses, injuries, and community resilience metrics. These powerful tools are seamlessly integrated into SimCenter Research Applications, supporting researchers in tackling complex challenges through computational workflows.
BRAILS++ is a new AI-enabled tool to assist regional-scale simulations. BRAILS++ utilizes machine learning (ML) and deep learning (DL) to create enhanced building inventory databases of cities. Examples of its capabilities include: (a) The identification of roof shapes to improve the damage and loss calculations for the hurricane workflow. This implementation used data from open street map and images from Google Maps. (b) The identification of soft-story buildings to improve models in earthquake workflows. This implementation used engineering knowledge and a subset of images from Google Street View to train a neural network to automatically classify the remaining images.
| Release Date |
|---|
| V3.1.3 Aug 2024 |
pelicunis a Python package that provides tools for assessment of damage and losses due to natural hazards. It uses a stochastic damage and loss model that is based on the methodology described in FEMA P58 (FEMA, 2012). While FEMA P58 aims to assess the seismic performance of a building, with pelicun we want to provide a more versatile, hazard-agnostic tool that will eventually provide loss estimates for other types of assets (e.g. bridges, facilities, pipelines) and lifelines. The underlying loss model was designed with these objectives in mind and it will be gradually extended to have such functionality.
| Release Date |
|---|
| V3.3 March 2024 |
Surf is used to study the spatial variability and hidden patterns in large datasets. Initial alpha versions were implemented for spatial uncertainty quantification. SURF creates surrogate models that assist users in making predictions. This library features both classical random field models and machine-learning algorithms as the backends for spatial uncertainty quantification. It has been incorporated in SimCenter efforts such as the hurricane and earthquake testbeds for uncertainty quantification and data enhancement.
| Release Date |
|---|
| V1.0 Feb 2021 |
The Turbulence Inflow Tool (TInF) is designed to collect all required properties and parameters needed for various turbulence inflow models in OpenFOAM, and to augment an existing wind-around-a-building model by adding the necessary sections to respective parameter definition files.
| Release Date |
|---|
| V1.1 July 2020 |