神经网络在泥石流活跃程度评判中的应用

Application of the neural network in the debris flows activity evaluating

泥石流活跃程度的评判结果对保护当地人民生命财产安全和经济建设的发展及地质灾害的防治工程布置有着很大的影响。然而以往的评判方法多以定性评判为主。由于每个人的知识水平、工作经验及评判问题的思维方式的差异,从而使评判结果或多或少存在一定的误差。论文旨在寻求一种新的方法来实现对泥石流活跃程度的定量分析,以便尽可能的减少人为误差。人工神经网络是一种具有学习、记忆、计算、仿真等功能的网络结构。BP网络是目前工程上运用最为广泛的一种误差反传的人工神经网络。它可以模拟任意复杂的非线形映射关系。应用神经网络对泥石流活跃程度进行定量分析评判,可以在一定程度上减少定性评判中的人为因素影响,提高评判的准确性。论文简要介绍了BP神经网络的基本原理、训练过程,以及如何利用MATLAB软件中的神经网络工具箱来创建、训练和应用评判泥石流活跃程度BP网络。在BP网络模型建立时采用了对研究区泥石流活跃程度影响最主要的8个参数作为输入层,并选取了研究区的20个样本对网络进行训练。最后用训练好的网络对研究区的10条支沟分别进行计算。计算结果与实际情况相符,说明利用BP神经网络来评判泥石流活跃程度具有很好的实用价值。

The evaluating result of the debris flows activity is very important for protecting local people's life and property.It also has great influence on the development of economy and the control of geological hazard.However,the present major methods of evaluating are qualitative analysis,it often makes evaluating results some errors because of the discrepancy of everybody's working experience,knowledge and the ways of assessing problems.This paper aims at seeking a kind of new method to analyze quantitatively the activity of debris flows in order to reduce artificial errors as far as possible.;Artificial neural network characterized as a network possessing functions of learning,memory,calculation,emulation,etc..BP network is a kind of artificial neural network,which has been applied in a lot of engineering fields at present.The quantitative analysis for activity of debris flow with neural network can decrease artificial influence and improve the accuracy of evaluating.Based on introducing briefly training process and the basic principle of BP neural network,this paper analyzes how to use the toolbox of the MATLAB software to create and train the BP network and apply it for evaluating the debris flows activity.When BP network model is established,8 parameters affecting the debris flows activity of the studying area are chose as the net input layer,and 20 samples are used to train the network.Ten ditches of debris flows area are respectively calculated with the trained network.The results accord with the actual condition.So it has profound practical value to use BP neural network to evaluating debris flows activity.