Subsurface Natural Fracture Modelling and Prediction on Igneous Rocks of "U" Geothermal Field

Abstrak

United States of America has a promising geothermal energy potential, especially in Roosevelt Hot Springs area in Utah. Geothermal system needs fractures as a considerable aspect in geothermal system evaluation. Fracture formed by the geological condition in the area, so it can affect the characteristic of the fractures. This research aims to analyze the structural geology condition, fracture characteristic, fracture prediction accuracy, and the comparison of the fracture prediction result with the fracture model. To achieve it, there are some data processing steps, such as seismic data interpretation, building seismic attributes, building implicit fracture model, and predicting fracture occurrence using Support Vector Machine (SVM) method which is a machine learning method. The research shows the structural geology condition in the study area consists of east – west trending normal faults and north – south trending reverse faults. The fracture in the study area has a dominant trend of north – south with the intensity ranging from 0 to 3. High fracture intensity zone can be found around faults and curvatures. The fracture prediction using SVM method produces an accuracy value of 73%. Overall, the fracture prediction result is good enough, although there are some zones which have a poor result when it compared to the implicit fracture model.