Summary of Likelihood-free Inference and Hierarchical Data Assimilation For Geological Carbon Storage, by Wenchao Teng et al.
Likelihood-Free Inference and Hierarchical Data Assimilation for Geological Carbon Storage
by Wenchao Teng, Louis J. Durlofsky
First submitted to arxiv on: 20 Oct 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Dynamical Systems (math.DS); Geophysics (physics.geo-ph)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This study develops a hierarchical data assimilation framework for managing and expanding geological carbon storage operations. The framework treats geological hyperparameters as uncertain variables characterized by hyperprior distributions, addressing uncertainty in practical CO2 storage applications with scarce measurements. A likelihood-free inference algorithm, sequential Monte Carlo-based approximate Bayesian computation (SMC-ABC), is used to estimate posterior samples of hyperparameters given dynamic monitoring well data. An ensemble smoother with multiple data assimilation (ESMDA) procedure then provides posterior realizations of grid-block permeability using a 3D recurrent R-U-Net deep learning-based surrogate model for forward function evaluations. The SMC-ABC-ESMDA procedure is compared to a reference rejection sampling (RS) method, achieving close agreement in all quantities considered and providing a speedup of 1-2 orders of magnitude. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study helps us manage and grow places where we store carbon dioxide underground. It’s important because we need to make sure this process works well. The researchers developed a new way to do this using computer models and math. They treated some things that are hard to know as uncertain, which makes sense when we don’t have enough data. They used a special algorithm to figure out what these unknowns might be, then used another method to make predictions about the underground rock formations. This new approach works well and is much faster than the old way. |
Keywords
» Artificial intelligence » Deep learning » Inference » Likelihood
