A preliminary analysis of the time series of displacements at fiducial stations obtained from continuous GPS observations during the period of Sept. 1998 to Oct. 2001 in the Crustal Movement Observation Network of Chi...A preliminary analysis of the time series of displacements at fiducial stations obtained from continuous GPS observations during the period of Sept. 1998 to Oct. 2001 in the Crustal Movement Observation Network of China (CMONOC) is made. The selection of datum for producing displacement time series suitable for earthquake prediction is discussed. Time series of horizontal crustal displacements are obtained by using a datum of a stable group of 9 stations with very small relative horizontal displacements in eastern China as reference. Time series of vertical crustal displacements are obtained by using a stable group of 7 stations scattered in different regions with relatively small relative vertical displacements as reference. During the period of 2000 to 2001, anomalous horizontal and vertical displacements occurred twice at the fiducial stations in western China. These anomalies may be related to seismic activities of magnitudes about 6 in the Yunnan region on the North South seismic belt.展开更多
结合局部均值分解LMD(Local mean decomposition)算法和BP神经网络算法,提出一种全新的局部均值分解——BP神经网络位移时序预测模型。通过把实际监测的位移值作为训练样本,利用局部均值分解算法对其进行高度的自适应分解,得到多个生产...结合局部均值分解LMD(Local mean decomposition)算法和BP神经网络算法,提出一种全新的局部均值分解——BP神经网络位移时序预测模型。通过把实际监测的位移值作为训练样本,利用局部均值分解算法对其进行高度的自适应分解,得到多个生产函数PF(Product function)分量;而后通过BP神经网络模型对每一个PF分量进行预测,再把各个PF分量预测值进行重构累加,即可得到位移的预测值。通过BP神经网络对相关参数进行优化,达到了对于预测精度的改善。将该模型应用到永久船闸高边坡的三个监测点上进行位移时序预测中,结果表明,预测精度较高,具有一定的科学依据,在边坡体位移时序预测领域中具有极大的潜在价值。展开更多
针对滑坡位移难以预测、影响因素难以选择等问题,提出一种结合了二次移动平均(DMA)法、变分模态分解(VMD)、改进灰狼优化(IGWO)算法与支持向量回归(SVR)的模型进行滑坡位移预测。首先,利用DMA提取滑坡位移趋势项和周期项,采用多项式拟...针对滑坡位移难以预测、影响因素难以选择等问题,提出一种结合了二次移动平均(DMA)法、变分模态分解(VMD)、改进灰狼优化(IGWO)算法与支持向量回归(SVR)的模型进行滑坡位移预测。首先,利用DMA提取滑坡位移趋势项和周期项,采用多项式拟合对趋势项进行预测;其次,对滑坡周期项的影响因素进行分类,采用VMD对原始影响因子序列进行分解获得最优序列;再次,提出一种结合SVR与基于改进Circle多策略的灰狼优化算法CTGWO-SVR(Circle Tactics Grey Wolf Optimizer with SVR)对滑坡周期项进行预测;最后采用时间序列加法模型求出累计位移预测序列,并采用灰色预测的后验证差校验和小概率误差对模型进行评价。实验结果表明,与GA-SVR和GWO-SVR模型相比,CTGWO-SVR的预测精度更高,拟合度达到0.979,均方根误差分别减小了51.47%与59.25%,预测精度等级为一级,可满足滑坡预测的实时性和准确性要求。展开更多
文摘A preliminary analysis of the time series of displacements at fiducial stations obtained from continuous GPS observations during the period of Sept. 1998 to Oct. 2001 in the Crustal Movement Observation Network of China (CMONOC) is made. The selection of datum for producing displacement time series suitable for earthquake prediction is discussed. Time series of horizontal crustal displacements are obtained by using a datum of a stable group of 9 stations with very small relative horizontal displacements in eastern China as reference. Time series of vertical crustal displacements are obtained by using a stable group of 7 stations scattered in different regions with relatively small relative vertical displacements as reference. During the period of 2000 to 2001, anomalous horizontal and vertical displacements occurred twice at the fiducial stations in western China. These anomalies may be related to seismic activities of magnitudes about 6 in the Yunnan region on the North South seismic belt.
文摘结合局部均值分解LMD(Local mean decomposition)算法和BP神经网络算法,提出一种全新的局部均值分解——BP神经网络位移时序预测模型。通过把实际监测的位移值作为训练样本,利用局部均值分解算法对其进行高度的自适应分解,得到多个生产函数PF(Product function)分量;而后通过BP神经网络模型对每一个PF分量进行预测,再把各个PF分量预测值进行重构累加,即可得到位移的预测值。通过BP神经网络对相关参数进行优化,达到了对于预测精度的改善。将该模型应用到永久船闸高边坡的三个监测点上进行位移时序预测中,结果表明,预测精度较高,具有一定的科学依据,在边坡体位移时序预测领域中具有极大的潜在价值。
文摘针对滑坡位移难以预测、影响因素难以选择等问题,提出一种结合了二次移动平均(DMA)法、变分模态分解(VMD)、改进灰狼优化(IGWO)算法与支持向量回归(SVR)的模型进行滑坡位移预测。首先,利用DMA提取滑坡位移趋势项和周期项,采用多项式拟合对趋势项进行预测;其次,对滑坡周期项的影响因素进行分类,采用VMD对原始影响因子序列进行分解获得最优序列;再次,提出一种结合SVR与基于改进Circle多策略的灰狼优化算法CTGWO-SVR(Circle Tactics Grey Wolf Optimizer with SVR)对滑坡周期项进行预测;最后采用时间序列加法模型求出累计位移预测序列,并采用灰色预测的后验证差校验和小概率误差对模型进行评价。实验结果表明,与GA-SVR和GWO-SVR模型相比,CTGWO-SVR的预测精度更高,拟合度达到0.979,均方根误差分别减小了51.47%与59.25%,预测精度等级为一级,可满足滑坡预测的实时性和准确性要求。
