暴雨致灾机理分析及预测模型构建  

Mechanism analysis and prediction model construction of rainstorm disaster

作　　者：王卓颖 南华[1] WANG Zhuoying;NAN Hua(School of Energy Science and Engineering,Henan Polytechnic University,Jiaozuo Henan 454000,China)

机构地区：[1]河南理工大学能源科学与工程学院,河南焦作454000

出　　处：《工业安全与环保》2023年第11期43-49,共7页Industrial Safety and Environmental Protection

基　　金：国家自然科学基金(51974106)。

摘　　要：台阶边坡、路堑边坡等地常选用乔木作为护坡植物,而降雨易影响乔木的稳定性,进而影响其护坡效果。提前预测降雨对乔木稳定性的影响,则可及时预判边坡失稳、泥石流等灾害的发生。以非饱和土体力学特征和乔木根系固土机理为基本理论依据,分析失稳原因并建立相应理论模型;基于多种降雨方案,利用数值模拟得到不同降雨情况对乔木稳定性影响的相关数据;然后以模拟结果作为训练样本,利用BP神经网络构建预测模型,并制定相应的分类预警评判指标。研究表明:此预测模型不仅精度高,相比于数值模拟计算速度可提高数万倍,此成果可为灾害防治及预警工作提供参考。Step slopes,road graben slopes and other places often choose trees as slope protection plants,and rainfall is easy to affect the stability of trees,which in turn affects the effect of its slope protection.Advance prediction of ra-infall on the stability of trees,can be timely prediction of slope instability,debris flow and other disasters.On the basis of unsaturated soil mechanics and arbor root soil fixation mechanism,analyze the causes of instability and establish the corresponding theoretical model in this study.On the basis of various rainfall schemes,its influence on the stability of arbores was studied by numerical simulation.Then,using the simulated results as training samples,the prediction model is constructed by using BP neural network,and the corresponding classification and early warning evaluation index are formulated.The research shows that the prediction model has not only high precision,but also tens of thousands of times faster than numerical simulation.This result can provide reference for disaster prevention and early warning.

关 键 词：降雨入渗 非饱和土力学 安全稳定性 BP神经网络 预测模型 

分 类 号：P426.616[天文地球—大气科学及气象学]