Document Type
Conference Proceeding
Publication Date
2025
Publication Title
Geothermal Resources Council Transactions
Volume
49
Abstract
This paper presents a data-driven framework for evaluating the effectiveness of bio-polymer gel treatments in fracture-dominated, high-temperature geothermal reservoirs. Laboratory core-flooding experiments were conducted on samples from the Deadwood Formation under reservoir-representative conditions, including temperatures up to 160 °C and salinity levels up to 17% TDS. Key physical, thermal, and flow-related parameters—such as porosity, permeability, fracture width, heat retention, and gel extrusion pressure—were integrated into a Support Vector Machine (SVM) classification model to predict treatment success, defined by a residual resistance factor (RRF) threshold of 1000. To extend the applicability of the model beyond available core material, geologically informed synthetic samples were generated using a temperature–RRF log-linear relationship. The final SVM model achieved an accuracy of 84.6% and an AUC of 0.89, demonstrating strong predictive capability des pite a limited dataset. The proposed approach provides a computationally efficient and physically interpretable screening tool for early-stage assessment of polymer gel treatments in geothermal systems, without reliance on full-scale reservoir simulation.
First Page
2219
Last Page
2236
ISSN
0193-5933
Recommended Citation
Farhad A. Bina. "Support Vector Machine Optimization for Predicting Bio Polymer Gel Effectiveness in Fracture-Dominated High-Temperature Reservoirs" (2025). Petroleum Engineering Student Publications. 16.
https://commons.und.edu/pe-stu/16