Geothermal energy production confronts persistent challenges tied to undesirable events occurring during geothermal fluid/steam production, including flow instability, scaling and corrosion issues. Conventional monitoring and preventive methodologies have proven insufficient, necessitating innovative approaches for the prediction and identification of these undesirable events. This paper introduces a novel application of machine learning techniques to address this issue. Specifically, the utilization of supervised classification methods, such as K-Nearest Neighbor (KNN), Decision Tree, Random Forest, Artificial Neural Networks (ANNs), and Support Vector Machines (SVMs), is proposed for the identification and prediction of undesirable events in geothermal fluid/steam production processes. To demonstrate the effectiveness of this approach, a substantial and comprehensive dataset, namely the 3W Petrobras Oil Production Dataset, was leveraged. This dataset encompasses an extensive range of operational data with corresponding undesirable events that occurred, enabling the training and evaluation of machine learning models across diverse real-world scenarios. The results obtained underscore the significant potential of machine learning in the identification and prediction of undesirable events. The core of the analysis centered on the utilization of temperature and pressure data recorded by sensors located both at the bottom of the well and at the wellhead. By employing these vital variables as model inputs, remarkable performance in terms of F1 scores, precision, and recall was achieved. These results highlight the importance of advanced data analytics techniques in geothermal energy production for the identification and prediction of undesirable events. In conclusion, this research contributes to the field of geothermal energy by introducing a data-driven approach for the identification and prediction of undesirable events during geothermal fluid/steam production. The successful application of machine learning algorithms, as demonstrated through the case study employing the 3W Petrobras Dataset, represents a significant advancement in ensuring the sustainability and efficiency of geothermal energy production processes. Furthermore, the insights gained from this study lay the foundation for proactive measures to identify and predict these undesirable events, ultimately enhancing the reliability and economic viability of geothermal power generation.

Press the Back button in your browser, or search again.

Copyright 2024, Stanford Geothermal Program: Readers who download papers from this site should honor the copyright of the original authors and may not copy or distribute the work further without the permission of the original publisher.