边坡位移 LMD-BP神经网络模型研究

摘要

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。%We present a novel LMD-BP neural network displacement time series prediction model in combination with the algorithms of local mean decomposition ( LMD ) and BP neural network .By selecting actual monitoring displacement data as the training sample and conducting highly adaptive decomposition on it using LMD algorithm , several product function (PF) components are obtained.After that, every PF component is predicted through BP neural network model , and then each prediction value is reconstructed and accumulated , the prediction value of displacement can be derived .BP neural network is used to optimise the correlated parameters , thus the improvement in prediction accuracy is reached .The model is put into application in displacement time series prediction carried out on three monitoring points at the high slope of permanent lock , result shows that the prediction accuracy is high , scientifically valid and has great potential value in the field of slope body displacement time series prediction .