Markov Chain Monte Carlo and Gaussian Processes for Calibration of Fuel Performance Codes
Dnr:
NAISS 2023/22-662
Type:
NAISS Small Compute
Principal Investigator:
Gustav Robertson
Affiliation:
Uppsala universitet
Start Date:
2023-07-01
End Date:
2024-07-01
Primary Classification:
10301: Subatomic Physics
Webpage:
Allocation
Abstract
In my project, CaNel - Calibration of Fuel Performance codes, I am looking at advanced statistical methodologies for calibrating fuel performance codes coming from the nuclear industry against measurements. I will be using Markov Chain Monte Carlo and Gaussian Processes (and potentially deep nets). To generate surrogate models for calibration, I need to create training and validation data using parallelization.