Natural State Modeling for a Geothermal System using Artificial Intelligence

Natural state modelling of a geothermal system is the process of adjusting uncertain reservoir parameters until an acceptable match with the measured field data is achieved. Complexity and insufficient knowledge of reservoir characteristics make this process time-consuming with a high computational cost. This study aims to examine the application of Artificial Intelligence (AI) to improve the time and efforts required for completing a successful natural state modelling. A synthetic 2D model of a vapor-dominated geothermal system is built using TOUGH2-EOS1, and was used as the subject and ground truth of natural state matching. An AI model was created to perform natural state modelling, and play a role as a prototype for full-field reservoir simulation that runs in a few seconds.