|
| | 講演要旨(英文) | | An indirect subsurface temperature estimation was made based on resistivity value from magnetotelluric survey through the application of artificial neural network. Back-propagation learning techniques was chosen to train the transfer function between subsurface resistivity and its correspondence temperature values. The training result was applied later to forecast the deeper temperature profiles up to 2 km below the existing boreholes, and was used to laterally predict the temperature below the MT stations nearby. Two different network models were used to calculate the temperature for vertical and lateral estimations, and were tested in two different sample areas. By applying this temperature estimation, the heat transfer mechanism due to the existing geothermal system can be analyzed. |
|
|
|
| |
|