| 郭强,谌宏伟,张馨妍,尹芳,李波,陈雪新.基于贝叶斯理论的慈利县降雨型滑坡预警模拟研究[J].矿产勘查,2025,16(11):2515-2526 |
| 基于贝叶斯理论的慈利县降雨型滑坡预警模拟研究 |
| Study on early-warning model of rainfall-induced landslide based on Bayesian theory in Cili County |
| 投稿时间：2024-03-14 |
| DOI：10.20008/j.kckc.202511013 |
| 中文关键词: 贝叶斯理论 预警模型 第四系 降雨型滑坡 慈利县 |
| 英文关键词: Bayesian theory early-warning model quaternary precipitation-induced landslide Cili County |
| 基金项目:本文受湖南省地质院科研课题(HNGSTP202106)与湖南省水利科技重大项目(XSKJ2019081-09)联合资助。 |
|
| 摘要点击次数: 486 |
| 全文下载次数: 125 |
| 中文摘要: |
| 湖南省慈利县降雨充沛，由降雨诱发的山地第四系松散层滑坡是区内主要的滑坡类型。开展该类滑坡的预警研究，对于及时采取应对措施，避免生命财产遭受损失具有重要意义。在滑坡影响因素分析的基础上，构建基于贝叶斯理论的滑坡预警概率模型，并选取典型滑坡开展了预警模拟验证。预警模型以坡度、细粒土含量、降雨及土壤水含量和孔隙水压力为输入变量建立贝叶斯网络，以“降雨—变形”为滑坡预警输出模式。典型滑坡监测数据的关系分析显示，滑坡倾角和位移的变化与降雨强度均呈较好的相关性，较好地反映了降雨对滑坡的诱发作用。预警模型模拟验证结果表明，滑坡倾角和位移发生变化的概率随降雨强度的增大而增大，从小雨、中雨、大雨到暴雨，滑坡 X、Y、Z三方向倾角发生变化的概率从 13.0%增大到约 20.7%、21.9%和 31.7%，滑坡位移发生变化的概率从 45.0%增大到 47.6%、48.2%和 54.0%，均位于黄色预警区间，与实际监测数据反映的滑坡状况吻合。综合来看，以坡度、细粒土含量、降雨及土壤水含量和孔隙水压力作为预警模型输入变量能充分体现降雨对滑坡的影响，预测模型反映了这些影响因素及其相互作用关系，较为合理，可以作为研究区及其他相似区域滑坡预警参考。 |
| 英文摘要: |
| There is abundant rainfall in Cili County in Hunan Province, and the quaternary unconsolidatedlandslide induced by rainfall is the main type in the study area. It is of great significance to carry out early-warningresearch on this kind of landslide for timely adopting countermeasures and avoiding loss of life and property. Ac.cording to the influence factor analysis, the landslide early-warning probability model based on the Bayesian theorywas built, and the model was verified using a classical landslide. Slope, content of the fine-grained soil, precipita.tion, soil water content and pore water pressure were used as the variables to build the Bayesian network, and the re.lationship between precipitation and deformation was determined as the output of the early-warning model. The cor.relation analysis of the monitoring data of the classical landslide showed that the variation of the dip angle and thedisplacement correlated with the rainfall better, which indicated the inducing effect of rainfall to the landslide. Theearly-warning model modelling showed that the probability of the variation of the dip angel and the displacement in.creased with the precipitation intensity increasing, and with precipitation increasing from light to moderate, heavyand torrential, the probability of the dip angle variation of X, Y, Z direction increased from13.0% to 20.7%, 21.9%and 31.7%, respectively, and that of the displacement variation increased from 45.0% to 47.6%, 48.2% and 54.0%,respectively. The probabilities all fall into the yellow alert range, which accord with the landslide situation showedfrom the monitoring data. In conclusion, using slope, content of the fine-grained soil, precipitation, soil water con.tent and pore water pressure as the input variables of the early-warning model can fully reflect the influence of pre.cipitation on landslides. The early-warning model embodies the influence factors and their interaction, and is rela.tively reasonable, and can be adopted as the reference model in the landslide early-warning in the study area andthe similar regions. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|