BP人工神经网络在青藏铁路南段地壳稳定性定量评价中的应用
|
摘要
将BP人工神经网络方法引入区域构造活动性、区域地壳稳定性研究领域,对青藏铁路南段沿线的构造活动性进行定量分析。选用断层运动速率、地震震级、温泉温度及剪切应变4个关键影响因子作为BP人工神经网络的输入向量,构造活动强度(α)作为输出向量,以α为定量判据,将全区划分为相对稳定区(α<0.22)、较不稳定区(α≈0.22~0.38)、不稳定区(α≈0.38~0.69)、极不稳定区或强烈构造活动区(α≥0.69)。在青藏铁路南段沿线划分出格仁错、崩错、当雄-羊八井、错那湖、唐古拉山口南、聂荣东北、聂荣西北、雅鲁藏布江断裂沿线、萨迦等不稳定区,在不稳定区内部进一步划分出申扎、蓬错、尼木、桑雄、羊八井5个极不稳定区。
BP artificial neural network is introduced to study tectonic activities and crust stability and applied into the quantitative analysis on tectonic activities along the Golmud-Lhasa Railway across south Tibetan Plateau.For key factors,slip rate of active fault,magnitude of earthquake,temperature of hotspring and tectonic strain for any unit in south Tibetan Plateau,are selected as the input vector in BP artificial neural network,and strength of tectonic activity(α) as the output vector.According to the values of α,south Tibetan Plateau can be classified into four regions:tectonically stable area with α<0.22,tectonically relatively unstable region with α≈0.22—0.38,tectonically active region with α≈0.38—0.69 and tectonically quite active region with α≥0.69.Typical unstable regions include Gerencuo,Bengcuo,Dangxiong-Yangbangjing,Cuonan lake,southeast Tangula Mts.,Sajia unstable regions and those along Yangluzangbu River.While five extremely unstable regions,Shenzha,Pengcuo lake,Nimu,Sangxiong,Yangbajing,are identified along the Golmud-Lhasa Railway across south Tibetan Plateau.
BP artificial neural network is introduced to study tectonic activities and crust stability and applied into the quantitative analysis on tectonic activities along the Golmud-Lhasa Railway across south Tibetan Plateau.For key factors,slip rate of active fault,magnitude of earthquake,temperature of hotspring and tectonic strain for any unit in south Tibetan Plateau,are selected as the input vector in BP artificial neural network,and strength of tectonic activity(α) as the output vector.According to the values of α,south Tibetan Plateau can be classified into four regions:tectonically stable area with α<0.22,tectonically relatively unstable region with α≈0.22—0.38,tectonically active region with α≈0.38—0.69 and tectonically quite active region with α≥0.69.Typical unstable regions include Gerencuo,Bengcuo,Dangxiong-Yangbangjing,Cuonan lake,southeast Tangula Mts.,Sajia unstable regions and those along Yangluzangbu River.While five extremely unstable regions,Shenzha,Pengcuo lake,Nimu,Sangxiong,Yangbajing,are identified along the Golmud-Lhasa Railway across south Tibetan Plateau.
引文
[1]Haykin S.神经网络原理[M].叶世伟,史忠植,译.北京:机械工业出版社,2004.
[2]Hagan M T,Demuth H B,Beale M H.神经网络设计[M].戴葵,译.北京:机械工业出版社,2002.
[3]陈守余,周梅春.人工神经网络模拟实现与应用[M].北京:中国地质大学出版社,2000.
[4]吴珍汉,胡道功,吴中海,等.青藏高原中段活动断层及诱发地质灾害[M].北京:地质出版社,2005:318-357.
[5]吴珍汉,周春景,王薇,等.青藏铁路沿线构造活动性评价与工程稳定性区划[J].地质通报,2005,24(5):401-410.
[6]胡海涛.“安全岛”——相对稳定地块概念的应用——以广东核电站场址选择为例[C]∥第四届国际工程地质大会论文集.北京:科学出版社,1984:38-46.
[7]李兴唐,许兵,黄典成,等.区域地壳稳定性研究的理论与方法[M].北京:地质出版社,1987:133-152.
[8]陈庆宣,孙叶,王治顺.运用地质力学方法研究区域地壳稳定性[M]∥地质力学文集:第9集.北京:地质出版社,1989:1-6.
[9]殷跃平.区域地壳稳定性研究的专家知识结构模型[J].水文地质工程地质,1990,(2):8-15.
[10]易明初.中国区域地壳稳定性图[M].北京:地质出版社,1997.
[11]胡道功,谭成轩,吴树仁.长江三峡及邻区地震活动的灰色预测[J].地震地质,1996,18(4):436-442.
[12]孙叶,谭成轩,李开善,等.区域地壳稳定性定量化评价[M].北京:地质出版社,1998:173-203.
[13]石广仁.地学中的计算机应用新技术[M].北京:石油工业出版社,1999.
[14]Kohonen T.An introduction to neural computing[J].NeuralNetworks,1998,(1):3-16.
[15]Nielsen R H.Theory of the Backpropagation Neural Network[C]∥Proceedings of the International Joint Conference onNeural Network.[S.l.]:[s.n.],1989:592-611.
[16]Robert H N.Theory of the Backpropagation Neural Network[C]∥Proceedings of the International Joint Conference onNeural Network.[S.l.]:[s.n.],1989:583-604.
[17]Guang L,Cheng S,Zhou R.Artificial neural network powersystem stabilizer trained with an improved BP algorithm[J].IEE Proceedings,1996,143(2):135.
[18]Wu Zhonghai,Zhao Xitao,Wu Zhenhan,et al.Quaternary ge-ology and faulting in the Damxung-Yangbajain basin[J].ActaGeological Sinica,2004,78(1):273-282.
[19]马宗晋.活动构造基础与工程地震[M].北京:地震出版社,1992:180-182.
[20]佟伟.西藏温泉志[M].北京:科学出版社,2000.
[21]王连捷,吴珍汉,王薇.青藏高原中段现今构造应力场与速度场的数值模拟[M]∥地质力学专业委员会和第四纪地质专业委员会.青藏高原地质过程与环境灾害效应文集.北京:地质出版社,2005:168-176.
[22]张培震,王琪,马宗晋.青藏高原现今构造特征与GPS速度场[J].地学前缘,2002,9(2):442-450.
[2]Hagan M T,Demuth H B,Beale M H.神经网络设计[M].戴葵,译.北京:机械工业出版社,2002.
[3]陈守余,周梅春.人工神经网络模拟实现与应用[M].北京:中国地质大学出版社,2000.
[4]吴珍汉,胡道功,吴中海,等.青藏高原中段活动断层及诱发地质灾害[M].北京:地质出版社,2005:318-357.
[5]吴珍汉,周春景,王薇,等.青藏铁路沿线构造活动性评价与工程稳定性区划[J].地质通报,2005,24(5):401-410.
[6]胡海涛.“安全岛”——相对稳定地块概念的应用——以广东核电站场址选择为例[C]∥第四届国际工程地质大会论文集.北京:科学出版社,1984:38-46.
[7]李兴唐,许兵,黄典成,等.区域地壳稳定性研究的理论与方法[M].北京:地质出版社,1987:133-152.
[8]陈庆宣,孙叶,王治顺.运用地质力学方法研究区域地壳稳定性[M]∥地质力学文集:第9集.北京:地质出版社,1989:1-6.
[9]殷跃平.区域地壳稳定性研究的专家知识结构模型[J].水文地质工程地质,1990,(2):8-15.
[10]易明初.中国区域地壳稳定性图[M].北京:地质出版社,1997.
[11]胡道功,谭成轩,吴树仁.长江三峡及邻区地震活动的灰色预测[J].地震地质,1996,18(4):436-442.
[12]孙叶,谭成轩,李开善,等.区域地壳稳定性定量化评价[M].北京:地质出版社,1998:173-203.
[13]石广仁.地学中的计算机应用新技术[M].北京:石油工业出版社,1999.
[14]Kohonen T.An introduction to neural computing[J].NeuralNetworks,1998,(1):3-16.
[15]Nielsen R H.Theory of the Backpropagation Neural Network[C]∥Proceedings of the International Joint Conference onNeural Network.[S.l.]:[s.n.],1989:592-611.
[16]Robert H N.Theory of the Backpropagation Neural Network[C]∥Proceedings of the International Joint Conference onNeural Network.[S.l.]:[s.n.],1989:583-604.
[17]Guang L,Cheng S,Zhou R.Artificial neural network powersystem stabilizer trained with an improved BP algorithm[J].IEE Proceedings,1996,143(2):135.
[18]Wu Zhonghai,Zhao Xitao,Wu Zhenhan,et al.Quaternary ge-ology and faulting in the Damxung-Yangbajain basin[J].ActaGeological Sinica,2004,78(1):273-282.
[19]马宗晋.活动构造基础与工程地震[M].北京:地震出版社,1992:180-182.
[20]佟伟.西藏温泉志[M].北京:科学出版社,2000.
[21]王连捷,吴珍汉,王薇.青藏高原中段现今构造应力场与速度场的数值模拟[M]∥地质力学专业委员会和第四纪地质专业委员会.青藏高原地质过程与环境灾害效应文集.北京:地质出版社,2005:168-176.
[22]张培震,王琪,马宗晋.青藏高原现今构造特征与GPS速度场[J].地学前缘,2002,9(2):442-450.
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