摘要
益阳市是湖南省地质灾害易发区之一,其中滑坡地质灾害占比74.1%。滑坡的敏感性分析通常采用专家知识评价、编录数据评价、滑坡过程物理评价等方法,其评价结果受评价人员水平的影响或评价过程复杂,适用性有限。基于BP神经网络算法,将益阳市338处滑坡以及1∶10000地质灾害详查得到的307处未滑坡的斜坡调查数据,提取到1个数据库中,利用BP神经网络方法,经过信号正向传播和反向传播的误差逐步修正,研究益阳市滑坡的敏感性影响因子。结果表明,计算模型和滑坡影响因子的重要性从高到低依次为斜坡类型、坡度、坡高、坡顶高程、地层岩性、斜坡结构类型,为滑坡野外识别和防治提供了数据支撑。通过实例分析验证了计算模型的正确性,该计算模型可应用于大批量斜坡稳定性分析。
Yiyang City is one of the areas prone to geological disasters in Hunan Province with landslide disasters accounting for 74.1%.The sensitivity analysis of landslide is usually carried out by means of expert knowledge evaluation,compilation data evaluation and physical evaluation of landslide processes.The evaluation results of the above methods are affected by the evaluation personnel quality,and the complicated evaluation process,and hence with limited applicability.Based on BP neural network algorithm,a database was set up with inputs from 338 landslides and 307 non-landslide slopes obtained from 1∶10000 geological disaster survey in Yiyang City.Studies of the influence factors of landslide sensitivity were carried out on those data by BP neural network method and gradually corrected by errors of signal forward and reverse propagation.The results show that the importance of calculation model and landslide sensitive factors from high to low is slope type,slope,slope height,slope top elevation,stratigraphic lithology,and slope structure type,which provides data support for landslide field identification and prevention.The calculation model is verified by an example analysis,and the results are proved to be accurate.This calculation model can be applied to a large number of slope stability analysis.
作者
黄小龙
熊远红
李红零
HUANG Xiaolong;XIONG Yuanhong;LI Hongling(Urban Geological Survey and Monitoring Institute of Hunan,Changsha 410000,Hunan,China)
出处
《城市地质》
2023年第4期8-14,共7页
Urban Geology
关键词
BP神经网络
滑坡
敏感性
影响因子
稳定性预测
BP neural network
landslide
sensitivity
influence factor
stability prediction
