Bayesian Uncertainty Quantification and Optimization of Jet Grout Column Diameter Prediction
Publication Date
1-1-2025
Document Type
Conference Proceeding
Publication Title
Geotechnical Special Publication
Volume
2025-March
Issue
GSP 365
DOI
10.1061/9780784485989.039
First Page
384
Last Page
394
Abstract
This paper introduces a Bayesian approach to quantify uncertainties in the parameters of the empirical equation for the prediction of jet grout column diameter, treating them as random variables in a probabilistic calibration framework. The resulting posterior distribution of model parameters integrates uncertainties that stemmed from various sources and considers correlations among model parameters. Sampling from this distribution evaluates parameter variability and confidence of model performance. Initially, deterministic non-linear regression was used to assess the jet grout column diameter based on jet grout type, specific energy, soil type, and SPT-N values. Three Markov Chain Monte Carlo (MCMC) algorithms—Metropolis-Hastings (MH), Hamilton Monte Carlo (HMC), and No-U-Turn-Sampler (NUTS)—were used for posterior sampling. The study concludes with a probabilistically calibrated model predicting jet grout column diameter, demonstrating the capability to quantify the uncertainty associated with the model performance.
Department
Civil and Environmental Engineering
Recommended Citation
Rakam Tamang, Yichuan Zhu, and Joseph Coe. "Bayesian Uncertainty Quantification and Optimization of Jet Grout Column Diameter Prediction" Geotechnical Special Publication (2025): 384-394. https://doi.org/10.1061/9780784485989.039
