基于灰色神经网络的边坡稳定性预报方法研究

本文在前人研究成果的基础上,研究了BP神经网络模型和灰色系统理论的原理,依据其适用条件及优缺点,对同一已知边坡取不同样本区间建立GM(1,1)模型,得到不同的预测结果。将多个灰色预测的结果作为输入变量使用BP人工神经网络进行组合,输出组合预测结果。提出了基于灰色神经网络范例推理的边坡稳定性评价方法,针对边坡的稳定性影响因素的复杂多变性和相当强的不确定性,建立了边坡范例检索模型,通过对边坡稳定性因素的灰色模型预处理,以及边坡范例的神经网络学习,最终实现了当前边坡的稳定性评价。

Study on Forecasting of Slope Stability Based on Grey Neutral Network

Based on the studies before, BP (Back Propagation) Neural Network and Grey Theory are researched in this paper. With the using conditions, as well as the advantages and disadvantages, GM(1,1) model is established under different sample intervals of the same known slope, and different forecasting results are obtained. Using several Grey results as input data, combine them with BP Neural Network, and then the combination results are gained. The evaluation of slope stability forecasting is proposed based on example-inference of Grey Neural Network. According to the complicated flexible influences and strong uncertainty of slope stability, the slope example index model is set up. Finally, the current slope stability evaluation can be realized, through Grey model preprocessing of the slope stability factors, and Neural Network study of slope examples.