神经网络在天然气水合物勘探中的应用研究
|
摘要
近年来,我国在南海进行了多学科多手段综合海洋地质调查研究,揭示了天然气水合物存在的证据,并于2007年5月实施深海钻探并首次获取了天然气水合物实物样品,研究显示我国南海天然气水合物储量达194亿立方米之多。
随着天然气水合物勘探程度的不断深入,获得的天然气水合物储层资料、岩心样品与测井数据的丰富,如何有效利用已取得的数据对天然气水合物的赋存前景进行预测,是广大海洋科技工作者急需解决的难题。
海洋天然气水合物勘探方法主要有地球化学与地球物理两大类。但无论那类方法,所涉及的水合物评价的方法、模型与公式都非常复杂,相关的参数也存在不确定性,并存在误差传递导致结果不可信等问题。而人工神经网络在模式分类、数据预测、函数拟合等方面都有强大的优势,并在石油勘探领域应用已有了许多成功运用的案例。
本论文尝试运用人工神经网络技术,在天然气水合物勘探领域选取相关的应用切入点,通过训练建立神经网络模型,利用其强大的非线性映射能力,揭示天然气水合物勘探评价中涉及的多个属性之间的非线性关系,为寻找、评价天然气水合物提供一定的方法与技术支持。
Comprehensive research of the marine geology of South China Sea has been carried out in recent years. The past studies confirmed the existence of Natural Gas Hydrate in this area. Especial in 2007 5th, scientists had conducted dedicated deep drilling expeditions and got the samples of the Natural Gas Hydrate for the fist time. Also some literates show that the reserves maybe up to 19.4 billion m3.
As the exploration of Natural Gas Hydrate gets deeper and deeper, we got so much data and files as well as the samples of Gas Hydrate. How to use these data to valuate the reserves of the natural gas hydrate becomes an urgent problem that all scientists met.
The exploration of natural gas hydrate contains two main methods including geochemistry and geophysics. However both of them have some limits such as 1) the complication of the evaluation method, model and formula, 2) the uncertainty of the parameters, 3) the lack of credibility in the result due to the transfer error etc. Nevertheless artificial neural network can figure out all these problems. It also has some special advantages such as pattern classification, data forecast, fitting function and so on.
This study took the application of artificial neural network technology as a breakthrough point, and established a neural network model. Through that we can find the nonlinear relationship in various attributes concern about the valuation of the natural gas hydrate. This paper is aiming at providing some method and technology supports in searching and evaluating the natural gas hydrate.
随着天然气水合物勘探程度的不断深入,获得的天然气水合物储层资料、岩心样品与测井数据的丰富,如何有效利用已取得的数据对天然气水合物的赋存前景进行预测,是广大海洋科技工作者急需解决的难题。
海洋天然气水合物勘探方法主要有地球化学与地球物理两大类。但无论那类方法,所涉及的水合物评价的方法、模型与公式都非常复杂,相关的参数也存在不确定性,并存在误差传递导致结果不可信等问题。而人工神经网络在模式分类、数据预测、函数拟合等方面都有强大的优势,并在石油勘探领域应用已有了许多成功运用的案例。
本论文尝试运用人工神经网络技术,在天然气水合物勘探领域选取相关的应用切入点,通过训练建立神经网络模型,利用其强大的非线性映射能力,揭示天然气水合物勘探评价中涉及的多个属性之间的非线性关系,为寻找、评价天然气水合物提供一定的方法与技术支持。
Comprehensive research of the marine geology of South China Sea has been carried out in recent years. The past studies confirmed the existence of Natural Gas Hydrate in this area. Especial in 2007 5th, scientists had conducted dedicated deep drilling expeditions and got the samples of the Natural Gas Hydrate for the fist time. Also some literates show that the reserves maybe up to 19.4 billion m3.
As the exploration of Natural Gas Hydrate gets deeper and deeper, we got so much data and files as well as the samples of Gas Hydrate. How to use these data to valuate the reserves of the natural gas hydrate becomes an urgent problem that all scientists met.
The exploration of natural gas hydrate contains two main methods including geochemistry and geophysics. However both of them have some limits such as 1) the complication of the evaluation method, model and formula, 2) the uncertainty of the parameters, 3) the lack of credibility in the result due to the transfer error etc. Nevertheless artificial neural network can figure out all these problems. It also has some special advantages such as pattern classification, data forecast, fitting function and so on.
This study took the application of artificial neural network technology as a breakthrough point, and established a neural network model. Through that we can find the nonlinear relationship in various attributes concern about the valuation of the natural gas hydrate. This paper is aiming at providing some method and technology supports in searching and evaluating the natural gas hydrate.
引文
[1]陈剑平.基于MATLAB的神经网络模式识别技术在油气化探中的研究及应用[D], 2008
[2]陈敏,曹志敏,龚建明等.海底天然气水合物地球化学勘探新技术[J].矿物岩石, 2004, 4(24): 102-107
[3]程思海,陈道华,张欣.海底天然气水合物地球化学探测技术[J].海洋地质动态, 2004
[4]邓希光,付少英,黄永祥,等.南海北部东沙群岛HD196站位地球化学特征及其对水合物的指示[J ] .现代地质, 2006 , 20 (1): 92-102
[5]董长虹. MATLAB神经网络与应用[M]. 2版.北京:国防工业出版社, 2007
[6]杜云艳,温伟,曹锋.空间数据挖掘的地理案例推理方法及试验[J].地理研究, 2009
[7]樊栓狮,刘锋,陈多福.海洋天然气水合物的形成机理探讨[J ].天然气地球科学, 2004, 15 (5) : 524-530
[8]方银霞,金翔龙,杨树锋.海底天然气水合物的研究进展[J].海洋科学, 2000, 24(4): 18-21
[9]付少英.海底沉积物中不同形式烃类气体的地球化学意义[J].地学前缘, 2005, 12(3): 254-257
[10]葛哲学,孙志强.神经网络理论与MATLAB R2007实现[M].北京:电子工业出版社, 2007
[11]黄永样, Suess E,吴能友.南海北部陆坡甲烷和天然气水合物地质[M].地质出版社, 2008
[12]金庆焕,张光学,杨木壮,等.天然气水合物资源概论[M].北京:科学出版社, 2006
[13]李少波,林维正,顾轶东.神经网络在海洋天然气水合物检测中的应用[J].海洋技术, 2004
[14]林和平,张秉正,乔幸娟.回归分析人工神经网络[N].吉林大学学报, 2010, 3
[15]刘坚,陆红峰,廖志良,等.东沙海域浅层沉积物硫化物分布特征及其与天然气水合物的关系[J ].地学前缘, 2005 , 12(3): 258-262
[16]刘小平,杨晓兰.海底天然气水合物地球化学方法勘探进展[J].天然气地球科学,2007, 4
[17]陆敬安,梁金强,罗文造.径向基函数神经网络在计算天然气水合物饱和度中的应用研究[J].海洋通报, 2009
[18]牛滨华,孙春岩,苏新,等.勘查地球化学方法适用于勘查天然气水合物的依据[J].现代地质, 2005, 19(1): 61-66
[19]彭远进,刘建仪,张烈辉.天然气水合物生成条件的神经网络方法预测[J].天然气工业, 2006
[20]舒宁,马洪超,孙和利.模式识别的理论与方法[M].武汉:武汉大学出版社, 2004
[21]苏新,陈芳,于兴河,等.南海陆坡中新世以来沉积物特性与气体水合物分布初探[J ] .现代地质, 2005, 19(1): 1-13
[22]孙帆,施学勤.基于MATLAB的BP神经网络设计[J].计算机与数字工程, 2007
[23]王宏斌,张光学,梁劲,等.南海北部陆坡构造坡折带中的天然气水合物[J ].沉积学报, 2008 , 26 (2): 283-293
[24]王钦军,薛林福.数据挖掘技术及其在地学中的应用[J].世界地质, 2000, 9
[25]韦钦胜,胡仰栋,安维中.基于BP神经网络的天然气水合物相平衡计算及预测[J].海洋技术, 2007
[26]吴能友,苏新,宋海斌,等.南海北部陆坡天然气水合物成藏机理研究:意义、现状与问题[J ] .海洋地质, 2007, (3): 1-11
[27]吴能友,张海啟,杨胜雄,等.南海神狐海域天然气水合物成藏系统初探[J ].天然气工业, 2007, 27 (9): 1-6
[28]夏登文,石绥祥,于戈等.海洋数据仓库及数据挖掘技术方法研究[J].海洋通报, 2005, 6
[29]杨木壮,梁金强,郭衣群.天然气水合物调查方法与技术[J].海洋地质动态, 2001, 17 (7): 14-19
[30]杨涛薛,杨竞红,蒋少涌.南海北部地区海洋沉积物中孔隙水的氢、氧同位素组成特征[J].地球学报, 2003, 24 (6): 511-514
[31]阴江宁,肖克炎,李楠. BP神经网络在化探数据分类中的应用[J].地质通报, 2011, 9
[32]张光学,陈邦彦.南海甲烷水合物资源研究与找矿前景[J ].海洋地质, 2000, 12 (3): 1-9
[33]张美.天然气水合物的研究现状和前景[J ].中山大学研究生学刊:自然科学与医学版, 2006, 27 (2)
[34]张涛. BP神经网络在测井解释中的应用研究[D], 2010
[35]赵青芳.天然气水合物的地球化学识别标志及探测技术[j].海洋地质动态, 2006, 22(12): 24-27
[36]朱大奇.人工神经网络研究现状及其展望[N].江南大学学报(自然科学版), 2004
[37]朱凯,王正林.精通MATLAB神经网络[M].电子工业出版社, 2010
[38]祝有海,刘亚玲,张永勤.祁连山多年冻土区天然气水合物的形成条件[J].地质通报, 2006, 25(1/2): 58-63
[39] Borowski W S, Paull C K, Ussler WⅢ. Blobal and local variations of inter stitial sulfate gradients in deep-water, continental margin sediments; sensitivity to underlying methane and gas hydrates [J]. Mar . Geol .1999, 159 (124): 131-154
[40] Chow J, Lee J S, Sun R, et al. Characteristics of the bottom simulating reflectors near mud diapirs: offshore southwestern Taiwan[J ]. Geo-Marine Letters, 2000, 20: 3-9
[41] Egeberg P K, Dickens G R. Thermodynamic and pore water halgen constraints on gas hydrate distribution at ODP Site 997 (Blake Ridge) [J]. Chem Geo l, 1999, 153: 53-79.
[42] Gerald R D.大型天然气水合物储层中甲烷的实地直接测量[J ].张延敏译.天然气地球科学, 1998, 9 (3/4) : 77-79
[43] Kastner M, Sample J C, Whiticar M J, et al. Geochemical evidence for fluid flow and diagenesis at the cascadia convergentmargin [J]. ProcODP Sci Results, 1995, 146: 375-384
[44] Kvenvolden K A, Kastner M. Gas hydrate of the Peruvianouter continental margin [J]. Proceeding ODP ScientificResults, 1990, 112: 517-526
[45] Kvenvolden K.A, Lorenson T.D. A Global Inventory of Natural Gas Hydrate Occurrence. http://walrus.wr.usgs.gov/globalhydrate/, 2008
[46] MacDonald, G. The future of methane as an energy resource. [J].Annual Review of Energy. 1990. 15: 53-83
[47] Max M.O.&Lowrie A. Oceanic methane hydrate: A“frontier”gas resource. [J]. Petroleum Geology.1996. 19(1): 41-56
[48] Torres M E, Mcmanus J. Fluid and chemical fluxes in and out of sediments hosting methane hydrate deposition Hydrate Ridge,OR, I: Hydro logical provinces[J]. Earth and Planetary Science Letters, 2002, 201: 525-540
[49] Yakushev V S, Chuvilin E M. Natural gas and gas hydrate accumulations within permafrost in Russia[J]. Cold Regions Science and Technology ,2000 ,31: 189-197
[2]陈敏,曹志敏,龚建明等.海底天然气水合物地球化学勘探新技术[J].矿物岩石, 2004, 4(24): 102-107
[3]程思海,陈道华,张欣.海底天然气水合物地球化学探测技术[J].海洋地质动态, 2004
[4]邓希光,付少英,黄永祥,等.南海北部东沙群岛HD196站位地球化学特征及其对水合物的指示[J ] .现代地质, 2006 , 20 (1): 92-102
[5]董长虹. MATLAB神经网络与应用[M]. 2版.北京:国防工业出版社, 2007
[6]杜云艳,温伟,曹锋.空间数据挖掘的地理案例推理方法及试验[J].地理研究, 2009
[7]樊栓狮,刘锋,陈多福.海洋天然气水合物的形成机理探讨[J ].天然气地球科学, 2004, 15 (5) : 524-530
[8]方银霞,金翔龙,杨树锋.海底天然气水合物的研究进展[J].海洋科学, 2000, 24(4): 18-21
[9]付少英.海底沉积物中不同形式烃类气体的地球化学意义[J].地学前缘, 2005, 12(3): 254-257
[10]葛哲学,孙志强.神经网络理论与MATLAB R2007实现[M].北京:电子工业出版社, 2007
[11]黄永样, Suess E,吴能友.南海北部陆坡甲烷和天然气水合物地质[M].地质出版社, 2008
[12]金庆焕,张光学,杨木壮,等.天然气水合物资源概论[M].北京:科学出版社, 2006
[13]李少波,林维正,顾轶东.神经网络在海洋天然气水合物检测中的应用[J].海洋技术, 2004
[14]林和平,张秉正,乔幸娟.回归分析人工神经网络[N].吉林大学学报, 2010, 3
[15]刘坚,陆红峰,廖志良,等.东沙海域浅层沉积物硫化物分布特征及其与天然气水合物的关系[J ].地学前缘, 2005 , 12(3): 258-262
[16]刘小平,杨晓兰.海底天然气水合物地球化学方法勘探进展[J].天然气地球科学,2007, 4
[17]陆敬安,梁金强,罗文造.径向基函数神经网络在计算天然气水合物饱和度中的应用研究[J].海洋通报, 2009
[18]牛滨华,孙春岩,苏新,等.勘查地球化学方法适用于勘查天然气水合物的依据[J].现代地质, 2005, 19(1): 61-66
[19]彭远进,刘建仪,张烈辉.天然气水合物生成条件的神经网络方法预测[J].天然气工业, 2006
[20]舒宁,马洪超,孙和利.模式识别的理论与方法[M].武汉:武汉大学出版社, 2004
[21]苏新,陈芳,于兴河,等.南海陆坡中新世以来沉积物特性与气体水合物分布初探[J ] .现代地质, 2005, 19(1): 1-13
[22]孙帆,施学勤.基于MATLAB的BP神经网络设计[J].计算机与数字工程, 2007
[23]王宏斌,张光学,梁劲,等.南海北部陆坡构造坡折带中的天然气水合物[J ].沉积学报, 2008 , 26 (2): 283-293
[24]王钦军,薛林福.数据挖掘技术及其在地学中的应用[J].世界地质, 2000, 9
[25]韦钦胜,胡仰栋,安维中.基于BP神经网络的天然气水合物相平衡计算及预测[J].海洋技术, 2007
[26]吴能友,苏新,宋海斌,等.南海北部陆坡天然气水合物成藏机理研究:意义、现状与问题[J ] .海洋地质, 2007, (3): 1-11
[27]吴能友,张海啟,杨胜雄,等.南海神狐海域天然气水合物成藏系统初探[J ].天然气工业, 2007, 27 (9): 1-6
[28]夏登文,石绥祥,于戈等.海洋数据仓库及数据挖掘技术方法研究[J].海洋通报, 2005, 6
[29]杨木壮,梁金强,郭衣群.天然气水合物调查方法与技术[J].海洋地质动态, 2001, 17 (7): 14-19
[30]杨涛薛,杨竞红,蒋少涌.南海北部地区海洋沉积物中孔隙水的氢、氧同位素组成特征[J].地球学报, 2003, 24 (6): 511-514
[31]阴江宁,肖克炎,李楠. BP神经网络在化探数据分类中的应用[J].地质通报, 2011, 9
[32]张光学,陈邦彦.南海甲烷水合物资源研究与找矿前景[J ].海洋地质, 2000, 12 (3): 1-9
[33]张美.天然气水合物的研究现状和前景[J ].中山大学研究生学刊:自然科学与医学版, 2006, 27 (2)
[34]张涛. BP神经网络在测井解释中的应用研究[D], 2010
[35]赵青芳.天然气水合物的地球化学识别标志及探测技术[j].海洋地质动态, 2006, 22(12): 24-27
[36]朱大奇.人工神经网络研究现状及其展望[N].江南大学学报(自然科学版), 2004
[37]朱凯,王正林.精通MATLAB神经网络[M].电子工业出版社, 2010
[38]祝有海,刘亚玲,张永勤.祁连山多年冻土区天然气水合物的形成条件[J].地质通报, 2006, 25(1/2): 58-63
[39] Borowski W S, Paull C K, Ussler WⅢ. Blobal and local variations of inter stitial sulfate gradients in deep-water, continental margin sediments; sensitivity to underlying methane and gas hydrates [J]. Mar . Geol .1999, 159 (124): 131-154
[40] Chow J, Lee J S, Sun R, et al. Characteristics of the bottom simulating reflectors near mud diapirs: offshore southwestern Taiwan[J ]. Geo-Marine Letters, 2000, 20: 3-9
[41] Egeberg P K, Dickens G R. Thermodynamic and pore water halgen constraints on gas hydrate distribution at ODP Site 997 (Blake Ridge) [J]. Chem Geo l, 1999, 153: 53-79.
[42] Gerald R D.大型天然气水合物储层中甲烷的实地直接测量[J ].张延敏译.天然气地球科学, 1998, 9 (3/4) : 77-79
[43] Kastner M, Sample J C, Whiticar M J, et al. Geochemical evidence for fluid flow and diagenesis at the cascadia convergentmargin [J]. ProcODP Sci Results, 1995, 146: 375-384
[44] Kvenvolden K A, Kastner M. Gas hydrate of the Peruvianouter continental margin [J]. Proceeding ODP ScientificResults, 1990, 112: 517-526
[45] Kvenvolden K.A, Lorenson T.D. A Global Inventory of Natural Gas Hydrate Occurrence. http://walrus.wr.usgs.gov/globalhydrate/, 2008
[46] MacDonald, G. The future of methane as an energy resource. [J].Annual Review of Energy. 1990. 15: 53-83
[47] Max M.O.&Lowrie A. Oceanic methane hydrate: A“frontier”gas resource. [J]. Petroleum Geology.1996. 19(1): 41-56
[48] Torres M E, Mcmanus J. Fluid and chemical fluxes in and out of sediments hosting methane hydrate deposition Hydrate Ridge,OR, I: Hydro logical provinces[J]. Earth and Planetary Science Letters, 2002, 201: 525-540
[49] Yakushev V S, Chuvilin E M. Natural gas and gas hydrate accumulations within permafrost in Russia[J]. Cold Regions Science and Technology ,2000 ,31: 189-197