- 56 views
- 80 downloads
Prediction of Horizontal In-situ Stress in Shale Gas Reservoirs Based on Artificial Neural Networks and Conventional Rock Mechanics ——A Case Study on Longmaxi Formation in Southern Sichuan (China)
-
- Author / Creator
- Du, Yifan
-
Shale gas is one of the most important unconventional fossil fuel resources. It is usually developed by horizontal drilling and hydraulic fracturing techniques. The in-situ stress magnitude distribution in a given shale gas field is a significant factor that should be considered by horizontal drilling and hydraulic fracturing. However, the accurate determination of in-situ stresses is normally hindered by the lack of experimental data. Aiming to address this issue, this thesis proposes a method of combining artificial intelligence and conventional rock mechanics to predict the in-situ stress magnitudes in a given shale gas reservoir based on the logging data. Since the experimental in-situ stress data are not sufficiently large to serve as the training dataset, this thesis selects the data in two wells for which the calculated in-situ stress magnitudes are in good agreement with the measured in-situ stress magnitudes as the training samples. Empirical rock-mechanics equations are used to generate more training data based on the data collected from these two wells. Then, a 4-layer artificial neural network model is established to predict the magnitudes of horizontal in-situ stresses in other wells. The results show that the predicted maximum horizontal in-situ stress magnitudes and predicted minimum horizontal in-situ stress magnitudes agree well with the measured data. Finally, a series of 3D maps showing the horizontal in-situ stress distributions in one shale gas reservoir in the Longmaxi formation of Sichuan (China) have been plotted by the newly developed neural network model.
-
- Subjects / Keywords
-
- Graduation date
- Fall 2022
-
- Type of Item
- Thesis
-
- Degree
- Master of Science
-
- License
- This thesis is made available by the University of Alberta Library with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.