A physics-informed multi-modal transformer framework for predictive maintenance and remaining useful life estimation of electrical submersible pumps

Yasin Khalili, Mohammad Ahmadi, Mostafa Keshavarz Moraveji · 2026

Read A physics-informed multi-modal transformer framework for predictive maintenance and remaining useful life estimation of electrical submersible pumps by Yasin Khalili, Mohammad Ahmadi, Mostafa Keshavarz Moraveji (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/PMC13572462.txt.

Description

Yasin Khalili, with two colleagues, presents a 2026 model that forecasts failure and remaining life for electric pumps deep in oil wells. These pumps lift oil where natural pressure falls short, and they wear fast under heat, grit, and gas. A surprise failure means lost output and a costly repair trip down the well.

The model, HF-ESPNet, reads many data streams at once: electric signals, pressure, heat, vibration, well facts, and repair history. Physics rules on fluid flow, heat, and vibration wear are built into its training aim. It learned from a real data set of more than 58,000 time-series samples from working wells. The model beat standard machine learning and transformer rivals on both tasks. Tests that dropped parts showed vibration features and the physics rules mattered most.

Catalog record

CREATORYasin Khalili, Mohammad Ahmadi, Mostafa Keshavarz Moraveji
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE86911
LICENSEhttps://creativecommons.org/licenses/by/4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/article/MED/PMC13572462

Browse the full library · About the Tower of Babel