基于改进模糊聚类迭代和神经网络的边坡失稳预测方法

发布时间：2018-03-05 20:34

本文选题：滑坡预测　切入点：神经网络预测　出处：《江西理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：滑坡灾害是全球三大自然灾害之一,边坡失稳所造成的滑坡灾害已f严重影响了人类日常的生活与生产,滑坡预报预测是有效降低滑坡带来损失的途径,目前,边坡稳定性预测方法的研究还处于不断的发展和完善中。本文针对现阶段滑坡预报预测方法中对边坡模糊性处理的不足,引入模糊数学的相关理论,提出了一种基于模糊聚类迭代理论的滑坡预测预报方法,主要研究内容及成果包括以下几个方面:(1)为了提高预测精度:首先设计了边坡数据收集及预处理方案,在进行滑坡预报预测前,设计了三维边坡数据收集方案,并引入小波降噪理论,对监测数据进行降噪处理。其次,基于原始的模糊聚类迭代模型,引入灰色关联分析理论,构建了改进的模糊聚类迭代模型,避免了确定各影响因素权重时的主观性。(2)基于改进的模糊聚类迭代模型,构建了由各影响边坡稳定因素评价边坡稳定性状态的方法,针对模糊聚类迭代模型无法预测边坡未来状态的特点,引入BP神经网络的相关理论,构建了预测边坡稳定性的改进模糊聚类迭代和神经网络复合模型。(3)基于改进模糊聚类迭代和神经网络复合模型,实现了边坡失稳的预测,为了便于计算与分析,运用MATLAB对预测的实现过程进行了程序化设计。(4)将本文建立的边坡失稳预测方法和以位移量作为预测对象的预测方法运用于同一工程实例中,通过对比两种方法的预测结果,证实了该方法的工程实用性。
[Abstract]:Landslide disaster is one of the three natural disasters in the world. Landslide disaster caused by slope instability has seriously affected human daily life and production. Landslide prediction and prediction is an effective way to reduce landslide losses. The research of slope stability prediction method is still in the process of continuous development and perfection. In this paper, the fuzzy mathematics theory is introduced to solve the problem of slope fuzziness treatment in the present landslide prediction method. A landslide prediction method based on fuzzy clustering iteration theory is proposed. The main research contents and results include the following aspects: in order to improve the prediction accuracy: firstly, a slope data collection and pretreatment scheme is designed. Before the landslide prediction and prediction, the 3D slope data collection scheme is designed, and the wavelet denoising theory is introduced to deal with the monitoring data. Secondly, based on the original fuzzy clustering iteration model, the grey correlation analysis theory is introduced. An improved fuzzy clustering iteration model is constructed to avoid subjectivity in determining the weight of each factor. Based on the improved fuzzy clustering iteration model, a method for evaluating slope stability state by influencing slope stability factors is constructed. In view of the fact that fuzzy clustering iteration model can not predict the future state of slope, the related theory of BP neural network is introduced. The improved fuzzy clustering iteration and neural network compound model for predicting slope stability are constructed. Based on the improved fuzzy clustering iteration and neural network composite model, the slope instability prediction is realized. In this paper, the program design of the process of prediction is carried out by using MATLAB. The prediction method of slope instability established in this paper and the prediction method with displacement as the prediction object are applied to the same engineering example. The prediction results of the two methods are compared by comparing the results of the two methods. The engineering practicability of this method is proved.
【学位授予单位】：江西理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU43

【参考文献】