Physics-informed gaussian process regression for reproducible and uncertainty-aware CO2 injectivity prediction
Shamsuddeen Adamu, Hitham Alhussian, Said Jadid Abdulkadir, Majdy Mohamed Eltayeb Eltahir, Sallam O F Khairy, Gasim Hayder, Mahdi Ali Lathbl, Hassan Salisu Mohammed · 2026
Read Physics-informed gaussian process regression for reproducible and uncertainty-aware CO2 injectivity prediction by Shamsuddeen Adamu, Hitham Alhussian, Said Jadid Abdulkadir, Majdy Mohamed Eltayeb Eltahir, Sallam O F Khairy, Gasim Hayder, Mahdi Ali Lathbl, Hassan Salisu Mohammed (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0352178.txt.
Description
Shamsuddeen Adamu and seven colleagues present a 2026 machine learning study on storing carbon dioxide underground. Injecting CO2 into rock gets harder over time as the rock clogs. Predicting that decline is vital for safe, lasting storage sites. Most prediction tools give a single number with no sense of how sure they are.
The team built a Gaussian process model with physical laws baked into it. Two rules were added: one about particle behaviour, one about how permeability falls as pores block. Tested on forty-four laboratory cases, the plain model predicted accurately. Its uncertainty estimates were also well calibrated. The physics-guided version broke the built-in physical rules far less often. Operators gain forecasts they can trust, complete with honest margins of error.
Catalog record
| CREATOR | Shamsuddeen Adamu, Hitham Alhussian, Said Jadid Abdulkadir, Majdy Mohamed Eltayeb Eltahir, Sallam O F Khairy, Gasim Hayder, Mahdi Ali Lathbl, Hassan Salisu Mohammed |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 53692 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |