| Title | The Fluid Temperature Prediction with Hydro-geochemical Indicators Using A Deep Learning Model: A Case Study Western Anatolia (Turkey) |
|---|---|
| Authors | Fusun TUT HAKLIDIR, Mehmet HAKLDIIR |
| Year | 2019 |
| Conference | Stanford Geothermal Workshop |
| Keywords | deep learning, hydrogeochemistry, geothermal, geothermal exploration |
| Abstract | Using geothermal fluids are beneficial however, it requires extensively exploration studies before drilling operations in a region. Geothermal drilling are complex and quite expensive operations that the costs may reach to 2.9 million USD for 1500-2500 m in Western Anatolia (Turkey) conditions. Because of the high operational costs, exploration phase of the geothermal projects is of great importance to reduce project costs. Evaluation of existing earth sciences data, detailed geology studies, mapping and some geochemical studies such as using geothermometers can provide important information to reach geothermal reservoir in a geothermal field. Nowadays, developing technology may give a chance to predict geothermal reservoir temperatures with less cost at geothermal fields. The hydro-geochemistry data quality is quite critical to predict reservoir temperature during the exploration phase. Machine learning is a technique of data analytics, teaching computers to learn from previous experience. Machine learning algorithms use computational methods to learn required information directly from data and these methods adaptively enhance their performance by increasing the number of samples or learning information available. One of the important applications of machine learning is prediction of a result. Deep learning is a subset of machine learning that attempt to learn at various levels, corresponding to various levels of abstraction. It is generally used for abstract useful data information. In this study, we developed a Deep Neural Network (DNN) model, to predict the geothermal fluid temperatures based on hydro-geochemistry data from Western Anatolia (Turkey). This is early stage study for the reservoir prediction and emphasize to data quality also. A comparative study of traditional machine learning algorithms has been performed to benchmark the performance of DNN predictions. |