数据挖掘方法在石油勘探开发中的应用研究
|
摘要
随着石油勘探开发的不断深入,要想从海量的地震数据中创造新的效益,有必要将数据挖掘方法应用于石油勘探和开发中,以获取高性能的地质、油藏、储层及流体性质评价的预测模型。该方法由特征选择、模型参数优化、性能评估等三大循环组成,核心技术是将遗传算法用于特征选择和参数优化,通过重复交叉验证得到泛化准确率的无偏估计以及从多种学习方法中优选出最终模型。本文以克拉玛依油田砾岩油藏水淹层评价为例,研究了6种特征子集方案和决策树、神经网络、支持向量机、贝叶斯网络及组合学习等5种方法,综合考虑预测模型的准确率和生成规则的可操作性,并选择决策树模型作为砾岩油藏水淹级别评价的最终预测模型。与传统的地球物理勘探方法相比较,采用该数据挖掘方法的优势在于:可以充分利用多专业数据;获得丰富的预测模型;探查和发现规律;提高预测准确度,因而能更好地为油气勘探开发服务。
With oil E&P(Exploration and Production) developing continually,in order to make new profit from massive oil data,it is necessary to apply the data mining method in oil exploration and development so that a high performance prediction model for geology,reservoir and fluid property evaluation was built up.The method consists of 3 loops which are feature selection,model parameter optimization and performance evaluation,the key technology is to apply generic algorithm in feature selection and model parameter optimization,unbiased estimation for generalized accuracy was obtained through repeating cross validation and the final model was optimized from the several methods.By taking water-flooded formation evaluation for conglomerate reservoir in Karamay oilfield as the example,6 feature subset schemes and 5 learning methods which are decision tree,neural networks,support vector machine,Bayesian network and array learning were optimized,by integratedly considering the accuracy of the prediction model and the operability of the generation rules,decision tree model was selected as final prediction model for conglomerate reservoir water-flooded grade evaluation.Compared with other traditional geophysical methods,the advantages for application of the data mining method are as below:the data in multi-disciplines could be utilized,complete prediction model was built,the regularities were searched and found out,prediction accuracy was raised and oil&gas E&P could be served better.
With oil E&P(Exploration and Production) developing continually,in order to make new profit from massive oil data,it is necessary to apply the data mining method in oil exploration and development so that a high performance prediction model for geology,reservoir and fluid property evaluation was built up.The method consists of 3 loops which are feature selection,model parameter optimization and performance evaluation,the key technology is to apply generic algorithm in feature selection and model parameter optimization,unbiased estimation for generalized accuracy was obtained through repeating cross validation and the final model was optimized from the several methods.By taking water-flooded formation evaluation for conglomerate reservoir in Karamay oilfield as the example,6 feature subset schemes and 5 learning methods which are decision tree,neural networks,support vector machine,Bayesian network and array learning were optimized,by integratedly considering the accuracy of the prediction model and the operability of the generation rules,decision tree model was selected as final prediction model for conglomerate reservoir water-flooded grade evaluation.Compared with other traditional geophysical methods,the advantages for application of the data mining method are as below:the data in multi-disciplines could be utilized,complete prediction model was built,the regularities were searched and found out,prediction accuracy was raised and oil&gas E&P could be served better.
引文
[1]石广仁等.多地质因素的勘探目标优选——人工神经网络法与多元回归分析法比较研究.石油学报,2002,23(5):19～22
[2]Aminzadeh F.Applications of AI and soft computing for challenging problems in the oil industry.Journal of Petroleum Science and Engineering,2005,47(1,2):5～14
[3]Ali monti C,Falcone G.Knowledge discovery in data-bases and multiphase flow metering:the integration of statistics,data mining,neural networks,fuzzy log-ic,and Ad hoc flow measurements towards well moni-toring and diagnosis.SPE Annual Technical Confer-ence and Exhibition,2002,Texas
[4]Mohaghegh S D.A new methodology for the identifi-cation of best practices in the oil and gas industry,u-sing intelligent systems.Journal of Petroleum Sci-ence and Engineering,2005,49(3,4):239～260
[5]Nikravesh M.Soft computing-based computational intelligent for reservoir characterization.Expert Sys-tems With Applications,2004,26(1):19～38
[6]Fayyad U,Gregory Piatetsky-Shapiro,Smyth P.Knowledge discovery and data mining:towards a uni-fying framework//Portland,Oregon:AAAI Press,1996:82～88
[7]Tan P,Steinbach Mand Kumar V.Introduction to Data Mining.Addison Wesley,2005
[8]Han J and Kamber M.Data Mining:Concepts and Techniques(The Morgan Kaufmann Series in Data Management Systems).Second Edition,Morgan Kauf mann,2006
[9]Witten I H and Frank E.Data Mining:Practical Machine Learning Tools and Techniques.Second E-dition,Morgan Kauf mann,2005
[10]Quinlan J.Induction of decisiontrees.Machine Learn-ing,1986,1(1):81～106
[11]Quinlan J.C4.5:Programs for Machine Learning.Morgan Kauf mann,1993
[12]Brei man L.Classification and Regression Trees.Chapman&Hall/CRC,1998
[13]Bishop C M.Neural Net works for Pattern Recogni-tion.Oxford University Press,USA,1995
[14]Cortes C and Vapnik V.Support-vector networks.Machine Learning,1995,20(3):273～297
[15]Platt J C.Fast training of support vector machines u-sing sequential mini mal opti mization.Advances in kernel methods:support vector learningtable ofcon-tents.Cambridge,MIT Press,1998,185～208
[16]Heckerman D.Bayesian networks for data mining.Data Mining and Knowledge Discovery,1997,1(1):79～119
[17]Langley P,Iba W,Thompson K.An analysis of Bayesian classifiers.Proceedings of the Tenth National Confer-ence on Artificial Intelligence,1992,223～228
[18]Friedman N,Geiger D,Goldszmidt M.Bayesian net-work classifiers.Machine Learning,1997,29(2):131～163
[19]Dietterich T G.Ensemble methods in machine learn-ing.Lecture Notes in Computer Science,2000,1857,1～15
[20]Freund Y,Schapire R E.A decision-theoretic gener-alization of on-line learning and an application to boosting.Journal of Computer and System Sci-ences,1997,55(1):119～139
[21]Goldberg D E.Genetic Algorithms in Search,Opti-mization and Machine Learning.Addison-Wesley Longman Publishing Co Inc Boston,MA,USA,1989
[22]Mierswa I,Wurst M,Klinkenberg R,Scholz M,Euler T.YALE:rapid prototypingfor complex data mining tasks.Proceedings ofthe12th ACMSIGKDDinter-national conference on knowledge discovery and data mining,2006,935～940
[2]Aminzadeh F.Applications of AI and soft computing for challenging problems in the oil industry.Journal of Petroleum Science and Engineering,2005,47(1,2):5～14
[3]Ali monti C,Falcone G.Knowledge discovery in data-bases and multiphase flow metering:the integration of statistics,data mining,neural networks,fuzzy log-ic,and Ad hoc flow measurements towards well moni-toring and diagnosis.SPE Annual Technical Confer-ence and Exhibition,2002,Texas
[4]Mohaghegh S D.A new methodology for the identifi-cation of best practices in the oil and gas industry,u-sing intelligent systems.Journal of Petroleum Sci-ence and Engineering,2005,49(3,4):239～260
[5]Nikravesh M.Soft computing-based computational intelligent for reservoir characterization.Expert Sys-tems With Applications,2004,26(1):19～38
[6]Fayyad U,Gregory Piatetsky-Shapiro,Smyth P.Knowledge discovery and data mining:towards a uni-fying framework//Portland,Oregon:AAAI Press,1996:82～88
[7]Tan P,Steinbach Mand Kumar V.Introduction to Data Mining.Addison Wesley,2005
[8]Han J and Kamber M.Data Mining:Concepts and Techniques(The Morgan Kaufmann Series in Data Management Systems).Second Edition,Morgan Kauf mann,2006
[9]Witten I H and Frank E.Data Mining:Practical Machine Learning Tools and Techniques.Second E-dition,Morgan Kauf mann,2005
[10]Quinlan J.Induction of decisiontrees.Machine Learn-ing,1986,1(1):81～106
[11]Quinlan J.C4.5:Programs for Machine Learning.Morgan Kauf mann,1993
[12]Brei man L.Classification and Regression Trees.Chapman&Hall/CRC,1998
[13]Bishop C M.Neural Net works for Pattern Recogni-tion.Oxford University Press,USA,1995
[14]Cortes C and Vapnik V.Support-vector networks.Machine Learning,1995,20(3):273～297
[15]Platt J C.Fast training of support vector machines u-sing sequential mini mal opti mization.Advances in kernel methods:support vector learningtable ofcon-tents.Cambridge,MIT Press,1998,185～208
[16]Heckerman D.Bayesian networks for data mining.Data Mining and Knowledge Discovery,1997,1(1):79～119
[17]Langley P,Iba W,Thompson K.An analysis of Bayesian classifiers.Proceedings of the Tenth National Confer-ence on Artificial Intelligence,1992,223～228
[18]Friedman N,Geiger D,Goldszmidt M.Bayesian net-work classifiers.Machine Learning,1997,29(2):131～163
[19]Dietterich T G.Ensemble methods in machine learn-ing.Lecture Notes in Computer Science,2000,1857,1～15
[20]Freund Y,Schapire R E.A decision-theoretic gener-alization of on-line learning and an application to boosting.Journal of Computer and System Sci-ences,1997,55(1):119～139
[21]Goldberg D E.Genetic Algorithms in Search,Opti-mization and Machine Learning.Addison-Wesley Longman Publishing Co Inc Boston,MA,USA,1989
[22]Mierswa I,Wurst M,Klinkenberg R,Scholz M,Euler T.YALE:rapid prototypingfor complex data mining tasks.Proceedings ofthe12th ACMSIGKDDinter-national conference on knowledge discovery and data mining,2006,935～940
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心