BayCal: Bayesian Calibration for Simulation Models

The BayCal toolkit is a software plugin for the RAVEN framework that allows engineers to calibrate and quantify uncertainties in simulation model parameters based on available experimental data using Bayesian inference. The tool uses machine learning and dimensionality reduction techniques to reduce computational costs and can be applied to various engineering problems. The DOE NE and other industries can benefit from this tool by automatically constructing surrogate models for expensive computational simulations and reducing the number of simulations required for convergence. By coupling with RAVEN, BayCal provides access to complex physical models and extensive scalability. The tool is advantageous over other alternatives due to its automation and cost-effectiveness. Successful implementation of the tool will provide a reliable and powerful tool for performing model calibrations of complex systems, benefiting the nuclear community and the broader technical community. Future collaboration with external companies for developing GUIs and commercialization is foreseeable.

This software is open source and available at no cost. Download now by visiting the product's GitHub page.