Application of Improved BP Neural Network Algorithm in Hydrocarbon Identification of Salinity Mud Invasion Reservoirs
Lixiang Feng, Wang Dan, Rongchao Cheng
Available Online April 2015.
- https://doi.org/10.2991/amcce-15.2015.18How to use a DOI?
- BP Neural Network; Wire Log; Gas and Geochemical Log; Hydrocarbon Identification
- Saline solution mud has been extensively used in many oil and gas fields. Its salinity property leads to the difficulty in identifying hydrocarbon using conventional wire logging information due to the lower reservoir resistivity resulted from saltwater invasion. Gas logging and geochemical logging directly reflect reservoir characteristics and thus can be much less affected by mud invasion. With the aid of FORWARD platform, the wire logging is therefore combined with gas logging and geochemical logging to identify hydrocarbon zones by means of BP neural network. To address the problems of lower convergence speed and easy occurrence of local minimum for BP neural network, a momentum term is added in the BP neural network algorithm, and the prediction accuracy of the method has been significantly improved. The method was successfully used in 48 layers of 5 wells in target oilfield. Testing results show that the coincidence rate of hydrocarbon interpretation is up to 93.7%, which proves the feasibility and efficiency of the method, and shows its application prospect in analogous reservoirs.
- Open Access
- This is an open access article distributed under the CC BY-NC license.
Cite this article
TY - CONF AU - Lixiang Feng AU - Wang Dan AU - Rongchao Cheng PY - 2015/04 DA - 2015/04 TI - Application of Improved BP Neural Network Algorithm in Hydrocarbon Identification of Salinity Mud Invasion Reservoirs BT - 2015 International Conference on Automation, Mechanical Control and Computational Engineering PB - Atlantis Press SN - 1951-6851 UR - https://doi.org/10.2991/amcce-15.2015.18 DO - https://doi.org/10.2991/amcce-15.2015.18 ID - Feng2015/04 ER -