基于监测信息多尺度变量提取的滑坡灾害态势评估分析

打开文本图片集

Landslide Disaster Situation Assessment based on Multi-scale Variable Extraction from Monitoring Information

，王 猛，（，）

关键词：滑坡；变分模态分解；变形预测；相对误差；灾害态势

中图分类号：TU196 文献标志码：A 文章编号：1005-8249（2025）01-0150-06

DOI：10.19860/j.cnki.issn1005- 8249.2025.01.028

WANG Rongrong，WANG Meng， WEI Zhongkai(Henan Fourth Geological Exploration Institute Co.，Ltd.， Zhengzhou 45OOO1， China）

Abstract：Toaccuratelyassessthesituationoflandslidedisasters，basedonlandslidedeformationdata，deformation characteristicanalysis iscarredoutfirst，andthen multi-scale variableextractionof deformationdataiscarredoutusing variationalmodedecompositiontoobtainseveralmodalcomponents；Atthesametime，alandslidedeformationpredictiomodel is constructed through extreme learning machinesandimproved sparrowsearch algorithms，and thepredictionresultsare used to achievelandslidedisastersituationasessmentandanalysis.Theanalysisresultsshowthatduringthemonitoring period，the cumulative deformation of each monitoring point ranges from 84.16 to $212.48~\mathrm{\mm}$ ，indicating that the overall cumulative deformation ofthelandslideisrelatively large，andthedeformationratealso hasalarge fluctuationrange.Therefore，the deformationcharacteristicsofthelandslidearesignificant；Atthesametime，variationalmodedecompositioncanffectiely extractmulti-scalevariablesfromlandslidedeformationdata，andthepredictionresultsshowthatbothmulti-scalevariable extractionandparameteroptimizationcanefectivelyimprovethepredictionaccuracy.Therelativeerorof theobtained prediction results mainly ranges from 1.98% to 2.14% ，which has a better prediction effect. According to the principle of disadvantage，thedisaster situationoflandslideswillfurther worsen，anddisasterpreventionandremediationshouldbecarred out as soon as possible.

Key words：landslide；variational mode decomposition；deformation prediction；relative error；disaster situation

0 引言

灾害态势评估是近年才兴起的热点课题，其主要用于评估灾害后续可能的状态，具有重要的工程意义，潜在研究价值较大；一般来说，滑坡变形是其稳定性的直观体现，基于滑坡变形数据基础上的灾害态势评估是可行的[1-2]

在以往滑坡变形预测方法中，江松[认为变形预测是实现滑坡灾害态势评估的有效方法，针对具体滑坡进一步完善滑坡灾害态势评估方法的构建是必要的；李仁江等4构建了考虑库水位变动引发滑坡变形滞后的预测模型；雷海智等5探究了长短期记忆人工神经网络在滑坡变形预测中的适用性，取得了良好的预测效果；袁维等利用灰色理论、随机森林算法等构建了滑坡变形预测模型，并通过预测结果实现了滑坡预警。（剩余11120字）