Expert System for Acidizing Based on BP Neural Network

Abstract:

The technique of acidizing stimulation is one of the most critical measures in petroleum industry to enhance production. As acidizing technique being an integrated course which combines science, practice and experience in one, it cannot be explained by mathematical technique precisely. For conventional acidizing, the workload is extremely huge and complicated, since it has built an extensive database with the help of a huge amount of the application samples. The Neural Network has the generalization ability, which not only has the most consistency with training samples, but also is a dependable network for predication of test samples, whose data distribution is similar to the previous ones. Expert system for acidizing based on the BP Neural Networks can predict a favorable acidizing fluids system and suitable dosage reasonably, effectively and accurately with a large pool of initial input parameters. Thereby this expert system can assist field application and realize the systematization and intelligence in oil field.

• Periodical: Advanced Materials Research (Volume 548)
• Pages: 438-443
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.548.438
• Online since: July 2012
• Authors: Heng Xue, Ping Li Liu, Nian Yin Li, Zhi Feng Luo, Li Qiang Zhao
• Keywords: Acidizing Fluids, Expert System, Neural Network (NN), Petroleum Engineering, Prediction

References

[1] Information on http: /www. hendrickson-intl. com.

[2] Information on http: /www. arco. com.

[3] Yan Jing: Design Study on Expert System of Sandstone Reservoir Acidification, volume 12 of journal of chongqing university of science and technology(natural sciences edition), chapter, 10, journal of chongqing university pablishers (2010).

[4] A.S. Al-Yami, J. Schubert: Expert System for the Optimal Design and Execution of Successful Completion Practices Using Artificial Bayesian Intelligence. SPE 143826. (2011).

DOI: 10.2118/143826-ms

[5] George J. Moridis: A Self-Teaching Expert System for the Analysis, Design, and Prediction of Gas Production From Unconventional Gas Resources. SPE 149485. (2011).

[6] M. Stundner, Decision Team-Software GmbH: How Data-Driven Modeling Methods like Neural Networks can help to integrate different Types of Data into Reservoir Management. SPE 68163. (2001).

DOI: 10.2118/68163-ms

[7] M.M. Zerafat1: Bayesian Network Analysis as a Tool for Efficient EOR Screening. SPE 143282. (2011).

[8] Yanbin Zang, Ruihe Wang: The method and expert system for risk assessment of drilling in high-sulfur gas field. SPE 134115. (2010).

[9] D. Zhu, N. Radjadhyax: Using Integrated Information to Optimizing Matrix Acidizing. SPE 68930. (2001).

[10] E. De la Vega, G. Sandoval, and M. Garcia: Integrating Data Mining and Expert Knowledge for an Artificial Lift Advisory System. SPE 128636 . (2010).

DOI: 10.2118/128636-ms

[11] Yu SW, Zhu KJ, Diao FQ: A dynamic all parameters adaptive BP neural networks model and its application on oil reservoir prediction, applied mathematics and computation(2008).

DOI: 10.1016/j.amc.2007.04.088

[12] C. Dong, D. Zhu, A.D. Hill: Modeling of the Acidizing Process in Naturally-Fractured Carbonates. SPE 63183. (2000).

DOI: 10.2118/63183-ms

[13] F. Gu, J. Guo, X. Fan: Acidizing Design And Treatment of Horizontal Wells In Naturally Fractured Reservoirs 97-107. (1997).

DOI: 10.2118/97-107

[14] George E. King, Amoco Production Co: Acidizing Concepts - Matrix vs. Fracture Acidizing. SPE 15279. (1986).

[15] Leonard J. Kalfayan, BJ Services: Fracture Acidizing: History, Present State, and Future. SPE 106371. (2007).