New Artificial Neural Network Model for Predicting the TOC from Well Logs
- Abdullah Sultan (King Fahd University of Petroleum & Minerals)
- Document ID
- Society of Petroleum Engineers
- SPE Middle East Oil and Gas Show and Conference, 18-21 March, Manama, Bahrain
- Publication Date
- Document Type
- Conference Paper
- 2019. Society of Petroleum Engineers
- 1.6 Drilling Operations, 7.6.7 Neural Networks, 7.6 Information Management and Systems, 1.6.9 Coring, Fishing, 7 Management and Information, 5 Reservoir Desciption & Dynamics, 5.6.1 Open hole/cased hole log analysis, 5.6 Formation Evaluation & Management
- Self-Adaptive, Duvernay formation, Artificial Neural Network, Barnett Shale, Total Organic Carbon
- 41 in the last 30 days
- 44 since 2007
- Show more detail
- View rights & permissions
|SPE Member Price:||USD 8.50|
|SPE Non-Member Price:||USD 25.00|
The key factor for characterizing unconventional shale reservoirs is the total organic carbon (TOC). TOC is estimated conventionally by analysis cores samples which requires extensive lab work, thus it is time-consuming and costly. Several empirical models are suggested to estimate the TOC indirectly using conventional well logs. These models assume the TOC and well logs are linearly related, this assumption significantly reduces the TOC estimation accuracy.
In this work, the design parameters of the artificial neural network (ANN) were optimized using self-adaptive differential evolution (SaDE) method to effectively predict the TOC from the conventional well log data. A new correlation for TOC calculation was developed, which is based on the optimized SaDE-ANN model. 460 data points of different well logs from Barnett formation were used to learn and validate the optimized SaDE-ANN model. The predictability of the SaDE-ANN correlation was compared with the available correlations for predicting the TOC using 29 data point from Duvernay formation.
The TOC was estimated using the optimized SaDE-ANN model with an average absolute percentage error (AAPE) and correlation coefficient (R) of 6% and 0.98, respectively. The SaDE-ANN correlation developed for TOC prediction outperformed the recent models suggested by Wang et al. (2016) and Mahmoud et al. (2017). The new empirical equation reduced the AAPE in predicting the TOC by 67% compared to Mahmoud et al. (2017) model in Duvernay formation.
|File Size||1 MB||Number of Pages||13|
Alajmi, M. D., Alarifi, S. A., Mahsoon, A. H., 2015. Improving Multiphase Choke Performance Prediction and Well Production Test Validation Using Artificial Intelligence: A New Milestone. SPE-173394-MS, presented at the SPE Digital Energy Conference and Exhibition, held in The Woodlands, Texas, USA, 2015, 3-5 March.
Passey, Q.R., Bohacs, K., Esch, W.L., Klimentidis, R., Sinha, S., 2010. From oil-prone source rock to gas-producing shale reservoir-geologic and petrophysical characterization of unconventional shale gas reservoirs. In: International oil and gas conference and exhibition in China. Society of Petroleum Engineers.
Schmoker, J.W., 1979. Determination of Organic Content of Appalachian Devonian Shales from Formation-Density Logs. Am. Assoc. Pet. Geol. Bull. 63, 1504–1509. doi:10.1306/2F9185D1-16CE-11D7-8645000102C1865D.
Van S. L., Chon B. H., 2017a. Effective Prediction and Management of a CO2 Flooding Process for Enhancing Oil Recovery using Artificial Neural Networks. ASME. J. Energy Resour. Technol. doi:10.1115/1.4038054.
Wang, P., Chen, Z., Pang, X., Hu, K., Sun, M., Chen, X., 2016. Revised models for determining TOC in shale play: Example from Devonian Duvernay Shale, Western Canada Sedimentary Basin. Mar. Pet. Geol. 70, 304–319. doi: 10.1016/j.marpetgeo.2015.11.023.