| The loess landslide developed in central and eastern Gansu Province and the fault zone accumulation layer landslide in southern Gansu Province have caused a large number of casualties and property losses.In order to deepen the understanding of the structural characteristics of these two types of typical landslides in Gansu Province and reduce the loss of landslide disasters,how to quickly obtain the stratigraphic structure and sliding surface characteristics has become an important research topic.Traditional detection methods can not meet the requirements of convenient operation,large detection depth and high precision,so the research of new detection means and methods is particularly important.Micro-motion detection method is a new passive seismic detection method,which has been widely used in urban space detection,karst cave detection,geothermal resource detection and other fields.However,there are few researches on landslide structure detection,especially for the large-scale fault accumulation layer landslide widely developed in the Belong River basin.In this paper,the loess landslide in Heifangtai area,Yongjing County,Linxia,and the fault zone accumulation layer landslide in Bailong River basin,which are the two most typical landslides in Gansu Province,are selected as the test landslide.The frets data of these two types of landslides are obtained respectively,and the spatial autocorrelation method is used for inversion to obtain the velocity structure of the landslide area.The sliding surface,stratigraphic structure and groundwater depth were extracted and compared with the borehole data and the results of high-density electrical method.At the same time,we discuss the influence of different acquisition parameters on the data quality in the process of field micro data acquisition.Finally,the application of genetic algorithm in the inversion of fretting dispersion curve is discussed.This study can provide reference for rapid and accurate acquisition of structural characteristics of typical landslides in Gansu Province.The main research achievements are as follows:(1)This paper studies the selection of collection conditions when obtaining landslide fretting data,and the influence of different collection conditions on the fretting data,and gives the optimal data collection scheme of loess landslide and the landslide of accumulation layer in fracture zone.The results show that the optimal collection radius of loess landslide and fault zone accumulation landslide is 30 ～ 100 meters.The best array type is triple circle,followed by L shape and one font.The collection time of loess landslide is 50-60 min,and that of fault zone accumulation layer landslide is 60-90 min.It is advisable to select 2Hz geophone for data collection of loess landslide and fault zone accumulation landslide.For areas with large relief(> 5m),the radius of the array should be reduced.(2)In this paper,the data of 160 micro-motion stations in two profiles in Heifangtai loess landslide area are analyzed,and the shear wave velocity structure of each point is obtained by inversion.Combined with single point data,the fretting profiles of two black square platforms and the fretting velocity structures of 16 single points were obtained.By analyzing the fretting data,the stratum structure at the measuring point is obtained.The fretting results show that in the Heifangtai area,the thickness below the surface is about 0 ～ 18.1m,and the average thickness is 17.9m.The apparent shear wave velocity is 0 ～ 250m/s,and the layer is mainly silt.The subsurface thickness is about 18.1 ～ 33.5m,with an average thickness of 12.6m.The lithology is saturated silt.The lithology is silty clay with an average thickness of 24.5m and a thickness of 33.5m～ 60.3m below the surface.The thickness below the surface is about 60.3 ～ 62.4m,with an average thickness of 1.7m,and the lithology is pebble bed.Below a depth of 65 m,there is breezed-unweathered mudstone bedrock.The mean error of comparison with borehole data and electrical data is 1.5m.(3)The micro-motion detection technology was used to obtain the micro-motion data of the landslide in the fault zone of the Belongjiang River basin,including the tooth haukou landslide,the spillway slope landslide,the Zhongpai landslide and the Lijianbeishan landslide.The micro-motion data of the longitudinal section of the tooth Haukou landslide,the micro-motion cross section data of the upper,middle and lower part of the spillway slope landslide,and the longitudinal and transverse micro-motion data of the Lijianbeishan landslide were obtained.At last,eight microdynamic profiles were synthesized.Based on the microdynamic profile data,we extracted the strata structure of the three landslides and analyzed the position of the sliding surface.There are two sliding surfaces,the new sliding surface and the old sliding surface,the depth of the new sliding surface is about 8 ～ 10 m and the depth of the old sliding surface is about 18 ～ 20 m.The depth of the sliding surface is 35 ～ 42 m.The sliding surface of Lijie Beishan landslide is 10 ～ 15 m.After comparing with borehole data and existing geological data,the error is less than 2m.(4)This paper selected a known borehole in the Heifangtai area and the gouge landslide respectively,and established the stratigraphic model parameters of each borehole.The generalized reflection-transmission coefficient method was used to obtain the theoretical dispersion curve of each borehole.After comparing with the stratigraphic structure inversion of genetic algorithm,the error was 2.82% and 4.41%,respectively.The feasibility and accuracy of genetic algorithm for inversion of fretting dispersion curve are verified. |
