李智超,赵正文,钟仪华,李建丽,刘道杰
LI Zhi-chao,ZHAO Zheng-wen,ZHONG Yi-hua,LI Jian-li,LIU Dao-jie (Southwest Petroleum University

• 1:西南石油大学

摘要(Abstract)：

在油田开发中,准确的产量预测对开发调整部署和提高措施作业效益有重要作用,它决定了油田生产投资的规模和决策方向。但是,油藏是一个复杂的多变量非线性动力学系统,由于油藏储层的非均质性和决定油田产量因素的不确定性,往往很难对油田产量进行准确的预测。小波神经网络是小波分析与前馈神经网络的融合,具有比BP网络更好的收敛性,同时具有处理复杂性、时变性和防震性的功能,可以对地质条件比较复杂、影响因素不确定的油田进行产量预测。通过对油田产量预测的实例计算表明,该方法具有很强的理论指导和较好的实际应用效果。
In the course of oilfield development,accurate.production prediction has always played important role in practicing development adjustment and enhancing the effects of stimulation operation,and moreover it determines the investment size of oilfield production and decision orientation.But oil reservoir is a complex,multiple variables and nonlinear dynamic system and furthermore because of the heterogeneity of oil reservoirs and the uncertainty of the factors that determine the oilfield production rate,the oilfield production is always very difficult to predict accurately.Wavelet neural network(WNN) is the integration of both wavelet analysis and feedforward neural network.This network has the following advantages:much better convergence than BP network;possesses the function of dealing with complicated, time-variant and vibration-proof problems and moreover predicts production rate for the oilfield with more complicated geological conditions and uncertain influencing factors.Through the actual calculation of production prediction,the method is proven to own much more powerful theoretical guide and much better application effects.

关键词(KeyWords)： 小波分析;神经网络;小波神经网络(WNN);油田产量预测;应用
wavelet analysis;neural network;wavelet neural network(WNN);oilfield production prediction;application

Abstract:

Keywords:

基金项目(Foundation): 西南石油大学研究生院创新基金项目资助(课题编号为:cxjj27084)

作者(Author): 李智超,赵正文,钟仪华,李建丽,刘道杰
LI Zhi-chao,ZHAO Zheng-wen,ZHONG Yi-hua,LI Jian-li,LIU Dao-jie (Southwest Petroleum University

参考文献(References)：

• [1]王洪元,史国栋.人工神经网络技术及其应用[M].北京:中国石化出版社,2002.
• [2]朱亚东.预测油气田产量和可采储量的F模型[J].石油勘探与开发,2000,27(1):54-56.
• [3]刘雨华,叶小岭.小波神经网络在综合评价中的应用[J].统计与决策,2005,195(8):15-16.
• [4]李留仁,焦李成.基于人工神经网络的油田产量多因素非线性时变预测[J],西安石油学院学报(自然科学版),2002,17(4): 42-44.
• [5]张广杰,刘明新,武若霞.神经网络在油田动态预测方面的应用[J],石油学报,1997,18(4):70-75.
• [6]吴新根,葛家理.应用人工神经网络预测油田产量[J].石油勘探与开发,1994,21(3):75-78.
• [7]周廷刚.基于遗传神经网络模型的大气环境质量评价方法[J].四川环境,2003,22(3):73-74.
• [8]樊纪香,张宏,李辉,等.BP网络和多元线性回归在产量预测中的应用[J].计算机工程与应用,2007,43(23):203- 204.
• [9]Ren Shou-xin,Gao Ling.Wavelet packet transform and artificial neural network applied to simultaneous kinetic multicomponent determination [J].Analytical and Bioanalytical Chemistry,2004,378(5): 1392-1398.
• [10]Yu Shi-wei,Zhu Ke-jun,Diao Feng-qin.A dynamic all parameters adaptive BP neural networks model and its application on oil reservoir prediction[J].Applied Mathematics and Computation,2008,195 (1):66-75.
• [11]Philip F,Coveney P V,Hughes T L.Predicting quality and performance of oilfield cements with artificial neural networks and FTIR spectroscopy [R].SPE 30079,1995.