A dynamic multi-objective optimization framework for geothermal systems via deep learning surrogate models

doi:10.1016/j.icheatmasstransfer.2026.112491

A dynamic multi-objective optimization framework for geothermal systems via deep learning surrogate models

2026-09-03

0

PRE

AI

DOI：10.1016/j.icheatmasstransfer.2026.112491

Original

Original request for help

Save

Abstract

En 中文

• Finite-element TH simulations generated IPPD datasets for surrogate training. • A PC-LSTM-TCN surrogate was evaluated for dynamic IPPD prediction. • The parallel surrogate achieved incremental gains over tested temporal models. • NSGA-II screened reinjection strategies under pressure, thermal, and profit goals. • Baseline comparison quantified IPPD, thermal mismatch, cost, and profit changes.

Keywords:

Hydrothermal geothermal system

Thermo-hydraulic simulation

Surrogate-assisted optimization

Machine learning

Reinjection strategy

A dynamic multi-objective optimization framework for geothermal systems via deep learning surrogate models

Abstract

Journal

Organization