Prediction method of petroleum exploration risks based on Bayesian network:a case study of the Jurassic Sangonghe Formation in the hinterland of Junggar Basin

doi:10.3969/j.issn.1672-7703.2023.01.010

Abstract

Abstract: Effective prediction of petroleum exploration risks is of great significance for optimizing exploration deployment and improving drilling success rate and economic benefits. Based on the analysis of the progress of risk prediction technology, a new prediction method of petroleum exploration risks is proposed based on Bayesian network. The transformation process from petroleum geological problems to probability prediction model is discussed, and the algorithm and prediction steps of average one-dependence estimators (AODE) training model are constructed. The quantitative evaluation of geological conditions of hydrocarbon accumulation in the Jurassic Sangonghe Formation in the hinterland of Junggar Basin is conducted, four main geological parameters of hydrocarbon supply, reservoir, trap, cap rock and preservation are determined, and a data set composed of parameters from 203 exploration wells is established. The results of five-fold cross validation show that: (1) The discrimination accuracy of the training set of four Bayesian networks (Naive Bayesian classifier, tree-augmented Bayesian classifier, AODE and k-dependence Bayesian classifier) is greater than 85%, indicating that the classification of parameters in the training set is effective; (2) The discrimination accuracy of the test set is greater than 82%, indicating a high success rate of prediction. AODE has the best prediction results, with an accuracy of 85.22%, therefore, it is used to predict the risk of petroleum exploration in the study area. The prediction results of AODE model not only agreed well with the exploration results within the reserve area, but also supported to predict favorable areas of three types of oil and gas resources outside the reserve area.

CLC Number:

TE19

Guo Qiulin, Ren Hongjia, Yu Jingdu, Liu Jifeng, Chen Ningsheng. Prediction method of petroleum exploration risks based on Bayesian network:a case study of the Jurassic Sangonghe Formation in the hinterland of Junggar Basin[J]. China Petroleum Exploration, 2023, 28(1): 108-119.

References

[1] White D A. Oil and gas play maps in exploration and assessment[J]. AAPG Bulletin, 1988,72(8):944-949.
[2] White D A. Geologic risking guide for prospects and plays[J].AAPG Bulletin, 1993,77(12):2048-2061.
[3] 付小东，张金区，王云鹏，等. 基于网格与GIS 的油气资源空间分布与定量评价系统及其应用[J]. 地质科技情报，2006,25(5):69-74.
Fu Xiaodong, Zhang Jinqu, Wang Yunpeng, et al . Oil and gas resources spatial distribution and quantitative evaluation system based on the grids and GIS and its application[J]. Geological Science and Technology Information, 2006,25(5):69-74.
[4] 胡素云，郭秋麟，谌卓恒，等. 油气空间分布预测方法[J]. 石油勘探与开发，2007,34(1):113-117.
Hu Suyun, Guo Qiulin, Chen Zhuoheng, et al . A method of predicting petroleum resource spatial distribution and its application[J]. Petroleum Exploration and Development, 2007,34(1):113-117.
[5] 郭秋麟，谢红兵，米石云，等. 油气资源分布的分形特征及应用[J].石油学报，2009,30(3):379-385.
Guo Qiulin, Xie Hongbing, Mi Shiyun, et al . Fractal model for petroleum resource distribution and its application[J]. Acta Petrolei Sinica, 2009,30(3):379-385.
[6] Olea R A, Cook T A, Coleman J L. A methodology for the assessment of unconventional (continuous) resources with an application to the greater natural buttes gas field, Utah[J].Natural Resources Research, 2010,19(4):237-251.
[7] 李军，杨晓娟，张晓龙，等. 基于三维马尔可夫链模型的岩性随机模拟[J]. 石油学报，2012,33(5):846-853.
Li Jun, Yang Xiaojuan, Zhang Xiaolong, et al . Lithologic stochastic simulation based on the three-dimensional Markov chain model[J]. Acta Petrolei Sinica, 2012,33(5):846-853.
[8] Chen Z H, Osadetz K G. An assessment of tight oil resource potential in Upper Cretaceous Cardium Formation, western Canada sedimentary basin[J]. Petroleum Exploration and Development, 2013,40(3):320-328.
[9] 庞敏，吴欣桐，梁琳. 基于贝叶斯概率的页岩气开发投资决策研究[J].统计与决策，2014(19):63-65.
Pang Min, Wu Xintong, Liang Lin. Research on investment decision-making in shale gas development based on Bayesian probability[J]. Statistics and Decision-Making, 2014(19):63-65.
[10] 王泽龙，刘先贵，李志勇，等. 基于贝叶斯理论的四维地震与生产动态数据的油藏历史拟合新方法[J]. 石油科技论坛，2022,41(2):9-22.
Wang Zelong, Liu Xiangui, Li Zhiyong, et al . A novel Bayesian-based reservoir history matching method for 4D seismic and production data[J]. Petroleum Science and Technology Forum, 2022,41(2):9-22.
[11] 王宜林，赵文智，杜金虎，等. 中国石油风险勘探工程决策管理与实施成效[J]. 中国石油勘探，2021,26(1):1-16.
Wang Yilin, Zhao Wenzhi, Du Jinhu, et al . CNPC: decision management and implementation results of venture exploration projects[J]. China Petroleum Exploration, 2021,26(1):1-16.
[12] 杨丽娜，解国军. 概率论和数理统计在油气资源评价中的应用[J]. 内蒙古石油化工，2016,42(4):121-124.
Yang Lina, Xie Guojun. Application of probability theory and mathematical statistics in oil and gas resource assessment[J].Inner Mongolia Petrochemical Industry, 2016,42(4):121-124.
[13] Peel F J, Brooks J R V. A practical guide to the use of success versus failure statistics in the estimation of prospect risk[J]. AAPG Bulletin, 2016,100(2):137-150.
[14] Fanailoo P, Clemens J. Application of Bayesian networks for estimating risk in shale operations[C]//Unconventional Resources Technology Conference, Society of Exploration Geophysicists,American Association of Petroleum Geologists, Society of Petroleum Engineers, 2016:1776-1782.
[15] 盛秀杰，金之钧，肖晔. 区带勘探中的油气资源评价方法[J]. 石油与天然气地质，2017,38(5):983-992.
Sheng Xiujie, Jin Zhijun, Xiao Ye. Petroleum resources assessment methodology in play exploration stages[J]. Oil & Gas Geology, 2017,38(5):983-992.
[16] Lilleborge M, Hauge R, Eidsvik J. Information gathering in Bayesian networks applied to petroleum prospecting[J].Mathematical Geosciences, 2016,48(3):233-257.
[17] Martinelli G, Eidsvik J, Hauge R, et al . Bayesian networks for prospect analysis in the north Sea[J]. AAPG Bulletin, 2011,95(8):1423-1442.
[18] Van Wees J D, Mijnlieff H, Lutgert J, et al . A Bayesian belief network approach for assessing the impact of exploration prospect interdependency: an application to predict gas discoveries in the Netherlands[J]. AAPG Bulletin, 2008,92(10):1315-1336.
[19] de Lucena B R D, Lustosa L J. Bayesian approach to the assessment of geological risk in oil and gas exploration[J].International Journal for Innovation Education and Research,2020,8(7):203-210.
[20] Ren H J, Wang X C, Guo Q L, et al . Spatial prediction of oil and gas distribution using tree augmented Bayesian network[J].Computers & Geosciences, 2020,142:104518.
[21] Bielza C, Larranaga P. Discrete Bayesian network classifiers: a survey[J]. ACM Computing Surveys (CSUR), 2014,47(1):1-43.
[22] Wang X, Ren H, Guo X. A novel discrete firefly algorithm for Bayesian network structure learning[J]. Knowledge-Based Systems, 2022,242:108426.
[23] Friedman N, Geiger D, Goldszmidt M. Bayesian network classifiers[J]. Machine Learning, 1997,29 (2):131-163.
[24] Sahami M. Learning limited dependence Bayesian classifiers[C].Menlo Park: AAAI Press, 1996:335-338.
[25] Webb G I, Boughton J R, Wang Z. Not so naive Bayes:aggregating one-dependence estimators[J]. Machine Learning,2005,58(1):5-24.
[26] 王小军，宋永，郑孟林，等. 准噶尔盆地复合含油气系统与复式聚集成藏[J]. 中国石油勘探，2021,26(4):29-43.
Wang Xiaojun, Song Yong, Zheng Menglin, et al . Composite petroleum system and multi-stage hydrocarbon accumulation in Junggar Basin[J]. China Petroleum Exploration, 2021,26(4):29-43.