PCA-FNN Based Performance Prediction for Water Injection in Oilfields

Abstract:

In order to solve the problem of uncertain cycle of water injection in the oilfield, this paper proposed a numerical method based on PCA-FNN, so that it can forecast the effective cycle of water injection. PCA is used to reduce the dimension of original data, while FNN is applied to train and test the new data. The correctness of PCA-FNN model is verified by the real injection statistics data from 116 wells of an oilfield, the result shows that the average absolute error and relative error of the test are 1.97 months and 10.75% respectively. The testing accuracy has been greatly improved by PCA-FNN model compare with the FNN which has not been processed by PCA and multiple liner regression method. Therefore, PCA-FNN method is reliable to forecast the effectiveness cycle of water injection and it can be used as an decision-making reference method for the engineers.

• Periodical: Advanced Materials Research (Volume 909)
• Pages: 410-417
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.909.410
• Online since: March 2014
• Authors: Ya Juan Tian, Ye Liu, Guo Jian Cheng, Zhe Wang
• Keywords: Cycle Prediction, Effect of Water Injection, FNN, PCA

References

[1] WANG Li-jun, ZHU Ming, SHANG Yan-peng: Numerical simulation of imbalanced injection in block of horizontal well-case of M72 horizontal well block, Aonan oilfield, Petroleum Geology and Recovery Efficiency, vol. 18 (2011), pp.67-69.

[2] WANG Yan, CHANG Yu-lian, JI Feng: Prediction and simulation of flooding system in oil fields, Journal of Daqing Petroleum Institute, vol. 31 (2007), pp.55-58.

[3] ZHOU Sheng-tian: Application of grey comprehensive evaluation method in evaluating oilfield water injection, Petroleum Geology and Engineering, vol. 20 (2006), pp.48-50.

[4] SHI Wei, LIU Jian-hui, QIN Shu-yu: Forecasting the effect of water infusion for coal seam based on BP neural network, Computer Simulation, vol. 26 (2009), pp.172-174.

[5] XIAO Wei, LIU Zhi-bin, GUO Da-li, et al: Dynamic indexes forecasting in injection oilfield recovery system based on neural network, Computer and Communications, vol. 15 (1997), pp.57-60, 78.

[6] LI Heng-wei: Research status review of fuzzy neural networks, Journal of Liaoning Institute of Science and Technology, vol. 12 (2010), pp.15-17.

[7] Marmo R, Amodio S, Tagliaferri R, et al: Textural identification of carbonate rocks by image processing and neural network: Methodology proposal and examples, Computers & geosciences, vol. 31 (2005), pp.649-659.

DOI: 10.1016/j.cageo.2004.11.016

[8] SHI Feng, WANG Xiao-chuan, YU Lei, et al: MATLAB neural network analysis of 30 cases (Beijing University of Aeronautics and Astronautics Press, Beijing 2010).