基于BP神经网络的公路高边坡力学参数反演与位移预测
Backward Propagation Neural Network-Based Mechanical Parameter Inversion&Displacement Prediction of Highway High Slopes
黄凯1
作者信息
- 1. 贵阳市城市建设投资集团有限公司,贵州 贵阳 550014
- 折叠
摘要
研究以某公路高边坡为对象,并针对该地区的地质情况选取Duncan-Chang模型建立数值模拟.通过设计正交试验,结合BP神经网络算法对边坡位移监测数据的反分析,求得 4 个重要的边坡力学参数.数值模拟计算所得的结果与实测变形监测数据接近.研究结果表明,通过几种方法的综合运用,能够在较小误差范围内对边坡的位移进行预测,该结论对边坡的防护和预警技术的运用具有重要的参考价值.
Abstract
This study takes a highway high slope as the object,and selects the Duncan-Chang model to establish numerical simulation based on the geological conditions of the region.Through the design of orthogonal experiments,combined with the back analysis of slope displacement monitoring data using BP neural network algorithm,four important slope mechanical parameters are obtained.The results of numerical simulation calculation are close to the measured deformation monitoring data.The research results show that the displacement of the slope can be predicted within a small error range through the comprehensive application of several methods.This conclusion has important reference value for the application of slope protection and early warning technology.
关键词
变形监测/神经网络/数值模拟/反分析/边坡Key words
deformation monitoring/neural network/numerical simulation/back analysis/slope引用本文复制引用
出版年
2024
NETL
NSTL
万方数据