紫金山金铜矿初始地应力场反演分析
|
摘要
初始地应力是影响工程边坡、硐室和地基岩体稳定性的重要因素之一,也是工程设计需要的基本指标之一。根据紫金山矿区地应力的实测资料,利用支持向量机和模拟退火算法对初始地应力场进行反演分析,一方面用支持向量机代替有限元计算,提高了计算分析速度;另一方面用模拟退火算法代替传统的优化算法,避免优化过程中目标函数陷入局部极小值而无法继续寻优的状态,提高了反演的效率精度。通过实测点的计算应力值与现场实测值的比较,两者在量值和方向上接近,表明该方法较合理地反映了初始地应力场的分布规律,能够满足矿区设计和研究的需要。
Initial geostress is one of the key factors influencing stability of slope,underground cavity and foundation;and it is one of the elementary indexes required in engineering design.According to the measured geostress data of Zijinshan gold and copper mine,initial geostress field is derived from back analysis based on the support vector machine and simulated annealing.Firstly the support vector machine is used to substitute the time-consuming finite element analysis;secondly simulated annealing algorithm is used for the optimization of objective function for fear that the search often falls in one of the minimums and cannot go any further when the conventional mathematical optimization methods are employed.Through the comparison between computed and measured geostress values of measuring points,it is shown that the two are similar in values and directions,so that the results by the back analysis results are very close to the real geostress field;and it can satisfy design requirement.
Initial geostress is one of the key factors influencing stability of slope,underground cavity and foundation;and it is one of the elementary indexes required in engineering design.According to the measured geostress data of Zijinshan gold and copper mine,initial geostress field is derived from back analysis based on the support vector machine and simulated annealing.Firstly the support vector machine is used to substitute the time-consuming finite element analysis;secondly simulated annealing algorithm is used for the optimization of objective function for fear that the search often falls in one of the minimums and cannot go any further when the conventional mathematical optimization methods are employed.Through the comparison between computed and measured geostress values of measuring points,it is shown that the two are similar in values and directions,so that the results by the back analysis results are very close to the real geostress field;and it can satisfy design requirement.
引文
[1]柴贺军,刘浩吾,王明华.大型电站坝区应力场三维弹塑性有限元模拟与拟合[J].岩石力学与工程学报,2002,21(9):1314-1318.CHAI He-jun,LIU Hao-wu,WANG Ming-hua.Stress field simulation and fitting by 3D elasticity FEM for large hydropower project[J].Chinese Journal of Rock Mechanics and Engineering,2002,21(9):1314-1318.
[2]付成华,汪卫明,陈胜宏.溪洛渡水电站坝区初始地应力场反演分析研究[J].岩石力学与工程学报,2006,25(11):2305-2312.FU Cheng-hua,WANG Wei-ming,CHEN Sheng-hong.Back analysis study on initial geostress field of dam site for Xiluodu hydropower project[J].Chinese Journal of Rock Mechanics and Engineering,2006,25(11):2305-2312.
[3]易达,徐明毅,陈胜宏,等.人工神经网络在岩体初始应力场反演中的应用[J].岩土力学,2004,25(6):943-946.YI Da,XU Ming-yi,CHEN Sheng-hong,et al.Application of artificial neural network to back analysis of initial stress field of rock masses[J].Rock and Soil Mechanics,2004,25(6):943-946.
[4]许传华,任青文,郑治,等.索风营水电站地下洞室岩体力学参数的位移反分析[J].岩土工程学报,2006,28(11):1981-1985.XU Chuan-hua,REN Qing-wen,ZHENG Zhi,et al.Displacement back analysis of rock mechanic parameters of underground grotto of Suofengying Hydraulic Power Plant[J].Chinese Journal of Geotechnical Engineering,2006,28(11):1981-1985.
[5]赵洪波,冯夏庭.位移反分析的进化支持向量机研究[J].岩石力学与工程学报,2003,22(10):1618-1622.ZHAO Hong-bo,FENG Xia-ting.Study on genetic-support vector machine in displacement back analysis[J].Chinese Journal of Rock Mechanics and Engineering,2003,22(10):1618-1622.
[6]刘开云,乔春生,腾文彦.边坡位移非线性时间序列采用支持向量机算法的智能建模与预测研究[J].岩土工程学报,2004,26(1):57-61.LIU Kai-yun,QIAO Chun-sheng,TENG Wen-yan.Research on non-linear time sequence intelligent model construction and prediction of slope displacement by using support vector machine algorithm[J].Chinese Jounal of Geotechnical Engineering,2004,26(1):57-61.
[7]方开泰,马长兴.正交与均匀试验设计[M].北京:科学出版社,2001.
[8]徐果明.反演理论及其应用[M].北京:地震出版社,2003.
[9]杨林德.岩土工程问题的反演理论与工程实践[M].北京:科学出版社,1996.
[2]付成华,汪卫明,陈胜宏.溪洛渡水电站坝区初始地应力场反演分析研究[J].岩石力学与工程学报,2006,25(11):2305-2312.FU Cheng-hua,WANG Wei-ming,CHEN Sheng-hong.Back analysis study on initial geostress field of dam site for Xiluodu hydropower project[J].Chinese Journal of Rock Mechanics and Engineering,2006,25(11):2305-2312.
[3]易达,徐明毅,陈胜宏,等.人工神经网络在岩体初始应力场反演中的应用[J].岩土力学,2004,25(6):943-946.YI Da,XU Ming-yi,CHEN Sheng-hong,et al.Application of artificial neural network to back analysis of initial stress field of rock masses[J].Rock and Soil Mechanics,2004,25(6):943-946.
[4]许传华,任青文,郑治,等.索风营水电站地下洞室岩体力学参数的位移反分析[J].岩土工程学报,2006,28(11):1981-1985.XU Chuan-hua,REN Qing-wen,ZHENG Zhi,et al.Displacement back analysis of rock mechanic parameters of underground grotto of Suofengying Hydraulic Power Plant[J].Chinese Journal of Geotechnical Engineering,2006,28(11):1981-1985.
[5]赵洪波,冯夏庭.位移反分析的进化支持向量机研究[J].岩石力学与工程学报,2003,22(10):1618-1622.ZHAO Hong-bo,FENG Xia-ting.Study on genetic-support vector machine in displacement back analysis[J].Chinese Journal of Rock Mechanics and Engineering,2003,22(10):1618-1622.
[6]刘开云,乔春生,腾文彦.边坡位移非线性时间序列采用支持向量机算法的智能建模与预测研究[J].岩土工程学报,2004,26(1):57-61.LIU Kai-yun,QIAO Chun-sheng,TENG Wen-yan.Research on non-linear time sequence intelligent model construction and prediction of slope displacement by using support vector machine algorithm[J].Chinese Jounal of Geotechnical Engineering,2004,26(1):57-61.
[7]方开泰,马长兴.正交与均匀试验设计[M].北京:科学出版社,2001.
[8]徐果明.反演理论及其应用[M].北京:地震出版社,2003.
[9]杨林德.岩土工程问题的反演理论与工程实践[M].北京:科学出版社,1996.
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