Affiliation:
1. Federal University of Technology, Owerri, Imo State, Nigeria
2. University of Oklahoma, Norman, USA
3. University of Nigeria Nsukka, Nigeria
Abstract
Abstract
Thermal conductivity of rocks defined as the ability of rocks to transmit heat, can indicate the potential for geothermal resource in a given location. While direct laboratory core sample analysis and indirect analysis leveraging empirical correlations from electric logs are used to determine thermal conductivity of rocks, they are usually expensive, time consuming and difficult to implement. Hence, in this study, several machine learning methods specifically Gradient Boosting Regressor, Random Forest, K-nearest neighbour, ensemble method (voting regressor), and Artificial Neural Networks were developed for the real-time prediction of thermal conductivity of rocks in geothermal wells. Data being obtained from Utah Forge field project included drilling data, thermal conductivity data and other necessary information from the field. With real-time sensor drilling data such as Rate of penetration (ROP), surface RPM, Flow in, Weight on bit (WOB), and Pump pressure, as input parameters and matrix thermal conductivity (MTC) as output, the models were developed. The results obtained from this study, showed excellent performances for majority of the models. However, it was observed that the ensemble voting regressor, which combined the top three models was able to predict thermal conductivity with above 89% and 80% R2 scores on the train and validation datasets respectively. Thus, this research work describes the feasibility of leveraging several machine learning methods in estimating thermal conductivity of rocks which is cost effective, and practically achievable.
Reference22 articles.
1. Smart Proxy Modeling WVU-LEADS 2 DOE-NETL 3 Leidos Research Support Team Application of Artificial Intelligence & Machine Learning in Computational Fluid Dynamics Smart Proxy Modeling;Aboaba;Application of Artificial Intelligence & Machine Learning in Computational Fluid Dynamics,2020
2. Voting-Averaged Combination Method for Regressor Ensemble
3. Barbier, E.
(2002). Geothermal energy technology and current status: an overview. I. Renewable and Sustainable Energy Reviews (Vol. 6). www.elsevier.com/locate/rser
4. Ekeopara, P., Odo, J., Obah, B., & Nwankwo, V. (2022). Hybridized Probabilistic Machine Learning Ranking System for Lithological Identification in Geothermal Resources. Society of Petroleum Engineers - SPE Nigeria Annual International Conference and Exhibition, NAIC 2022. https://doi.org/10.2118/212015-MS