基于BP神经网络的边坡可靠度分析
Slope Reliability Analysis Based on BP Neural Network
陈敏1
作者信息
- 1. 湖南联智科技股份有限公司,湖南 长沙 410203
- 折叠
摘要
为高效准确计算边坡稳定可靠度指标,提出了一种基于BP神经网络响应面的边坡可靠度计算方法.基于有限元强度折减法和结构可靠度理论建立了边坡结构可靠度功能函数,通过BP神经网络对边坡结构响应特性进行学习拟合,采用蒙特卡洛模拟方法在边坡临界状态进行重要抽样,计算边坡稳定可靠度指标,以洪辰边坡工程为例验证了该方法的可行性.结果表明:BP神经网络可以精确拟合边坡结构响应特性;基于BP神经网络得到的边坡可靠度指标为 1.255,与有限元计算得到的相对误差仅为-0.51%;提出的方法计算得到的可靠度指标精度较高,结果略偏保守,边坡结构整体处于稳定状态.
Abstract
A slope reliability calculation method based on BP neural network response surface is proposed to efficiently and accurately calculate the reliability index of slope stability.A functional function of slope structural reliability was established based on the finite element strength reduction method and structural reliability theory.The response characteristics of the slope structure were learned and fitted using a BP neural network.The Monte Carlo simulation method was used to conduct important sampling at the critical state of the slope,and the stability reliability index of the slope was calculated.The feasibility of this method was verified using the Hongchen slope project as an example.The results indicate that the BP neural network can accurately fit the response characteristics of slope struc-tures;The slope reliability index obtained based on BP neural network is1.255,with a relative error of only-0.51%compared to the finite element calculation;The reliability index calculated by the proposed method has high accuracy,and the results are slightly con-servative.The overall slope structure is in a stable state.
关键词
边坡工程/可靠度/稳定性系数/BP神经网络/蒙特卡洛法Key words
slope engineering/reliability/stability coefficient/BP neural network/Monte Carlo method引用本文复制引用
出版年
2024
NETL
NSTL
万方数据