Enhancing Geothermal Drilling Efficiency with Stuck Pipe Machine Learning Application

Abstract Intensive geothermal drilling effort is essential to reach the Indonesian government’s target of 3,355 MWe by 2030. To reduce the overall risk of stuck pipe, a significant risk in drilling geothermal wells, innovation such as Stuck Pipe Machine Learning (SPML) was developed to support engineers’ ability to reduce potential problems during drilling and optimize the overall drilling operations. Literature studies were conducted before developing the application, followed by data collection from 80 geothermal wells in Indonesia. Ten drilling parameters were selected, cleaned, normalized, and prepared to model the Artificial Neural Network algorithm, which has the highest accuracy compared to other models. The algorithm was then trained and tested on 12 geothermal offset wells, where 70% of the data was used for training and the rest for testing. Then it was also tested using historical data (blind test) for three geothermal offset wells. After confirming the accuracy, the application was tested on real-time drilling, successfully identifying pre-stuck or anomaly conditions. Although the overall accuracy of the algorithm cannot be determined due to the absence of a stuck pipe, the machine learning model shows its effectiveness in detecting anomalies in the occurrence of stuck pipe incidents in geothermal drilling operations. The SPML application continues to be developed to optimize geothermal drilling in Indonesia.