IMPLEMENTATION OF GRADIENTLESS NEURAL NETWORK FOR GEOTHERMAL PRODUCTION WELL EVALUATION AND EARLY WARNING SYSTEM

• Main Author: Harry Yudha Pratama, Muchamad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/49697
• Institution: Institut Teknologi Bandung

One of the crucial things in managing geothermal production system is monitor the performance of each production well. To evaluate the performance of a production well, engineers takes up to days because of the complicated process. Therefore, engineers often find that the well's performance drops significantly when the well was evaluated, this resulted in the effort to repair the well being very heavy, even impossible to repaired. On the other hand, in terms of technology, there have been many applications of Artificial Intelligence and Machine Learning in helping human work. It would be very good if the Artificial Neural Network (ANN) could be applied in the geothermal production system so that it could help in evaluating the well performance more quickly. The algorithm that is often used in ANN is gradient-descent, even though not all functions have a gradient. In this study, a gradient-less ANN application will be analyzed in evaluating the performance of geothermal wells.