CN-122020053-A - Railway tunnel portal slope instability risk assessment method and system based on image recognition

CN122020053ACN 122020053 ACN122020053 ACN 122020053ACN-122020053-A

Abstract

The invention provides a railway tunnel portal slope instability risk assessment method and system based on image recognition, and relates to the technical field of computers, wherein the method comprises the steps of acquiring video image data of a target railway tunnel portal slope area and synchronously acquiring waveform data; the method comprises the steps of constructing a slope surface deformation field according to video image data to obtain a dynamic differential displacement field, carrying out tunnel axial dynamic load propagation simulation according to waveform data, carrying out inversion to obtain a dynamic stress tensor field, carrying out slope-tunnel structure power interaction analysis according to the dynamic differential displacement field and the dynamic stress tensor field to obtain a time-varying stability coefficient sequence, carrying out instability precursor identification according to the time-varying stability coefficient sequence to obtain a slope instability mode, carrying out operation risk quantification according to the slope instability mode, and outputting a risk assessment result taking unscheduled vehicle interruption time as a measurement. The invention realizes accurate and dynamic assessment from a mechanical mechanism to an operation result of the tunnel portal slope instability risk under the cyclic dynamic loading of the train.

Inventors

CHEN XIAOXIONG
ZHANG DONGLIN
GUO JIANQIANG
ZHAO YINING
WANG CHAOYANG
DING DAWEI
YI MINGWEI
GUO GE
HAO BAOFENG
YIN CHANGMING
LIU SHUMING
LENG ZHICHAO
TAN FEI

Assignees

中铁上海工程局集团第七工程有限公司
山东潍烟高速铁路有限公司

Dates

Publication Date: 20260512
Application Date: 20260202

Claims (10)

1. The method for evaluating the risk of the slope instability of the railway tunnel portal based on the image recognition is characterized by comprising the following steps of: acquiring video image data of a side slope area of a target railway tunnel portal, wherein the video image data comprises visual information of the side slope surface, a tunnel portal lining structure and a joint area, and synchronously acquiring waveform data of a train passing through, wherein the waveform data are acquired by sensors arranged on sleeper on the inner side of the tunnel portal; constructing a slope surface deformation field according to the video image data, and performing constraint expansion and curvature analysis on space-time characteristic manifolds of the image by introducing geometric constraints of tunnel portal topography as manifold boundary conditions to obtain a dynamic differential displacement field of the slope surface; Carrying out tunnel axial dynamic load propagation simulation according to the waveform data, and inverting to obtain a dynamic stress tensor field in the surrounding rock of the tunnel portal by analyzing the dispersion relation of the vibration waveform in the steel rail-sleeper-railway ballast-lining medium and the waveguide effect of the tunnel tubular structure; Performing dynamic interaction analysis of a side slope-tunnel structure according to the dynamic differential displacement field and the dynamic stress tensor field to obtain a time-varying stability coefficient sequence, wherein the time-varying stability coefficient sequence is a ratio sequence of anti-slip virtual work and load virtual work of a side slope potential sliding body to change with time; Carrying out instability precursor identification according to the time-varying stability coefficient sequence, and obtaining a slope instability mode by monitoring the residual accumulation effect and the attenuation characteristic of the train passing clearance period and judging whether to trigger a convergence threshold defined by the maximum allowable deformation of the tunnel lining structure; And carrying out operation risk quantification according to the slope instability mode, and outputting a risk assessment result taking the off-schedule vehicle interruption time as a measure.
2. The method for evaluating the risk of slope instability of a railway tunnel portal based on image recognition according to claim 1, wherein the construction of a slope surface deformation field is performed according to the video image data, and the constrained expansion and curvature analysis are performed on the space-time characteristic manifold of the image by introducing geometric constraint of the tunnel portal topography as a manifold boundary condition, so as to obtain a dynamic differential displacement field of the slope surface, comprising: Carrying out space-time manifold construction processing according to the video image data, and obtaining a slope space-time characteristic manifold subjected to geometric constraint by identifying the edge contour of a tunnel portal structure and defining the boundary constraint as non-deformable manifold; performing constraint manifold expansion processing according to the side slope space-time characteristic manifold, and mapping a high-dimensional manifold to a measurable low-dimensional space by performing Riemann geometric expansion on the space-time characteristic manifold under a boundary constraint condition to obtain a side slope surface deformation field; And carrying out differential displacement solving treatment according to the slope surface deformation field, and obtaining a dynamic differential displacement field of the slope surface by calculating a tangent space covariant derivative of the surface deformation field relative to a tunnel lining fixed reference point.
3. The method for evaluating the risk of slope instability of a railway tunnel portal based on image recognition according to claim 1, wherein the method for simulating the propagation of dynamic load in the axial direction of the tunnel according to the waveform data, and inverting to obtain a dynamic stress tensor field in the surrounding rock of the tunnel portal by analyzing the dispersion relation of vibration waveforms in a rail-sleeper-ballast-lining medium and the waveguide effect of a tunnel tubular structure comprises the following steps: According to the waveform data, carrying out train dynamic load component separation treatment, and separating out direct vibration components transmitted by a rail-lining path and vibration components scattered by railway ballast-surrounding rock by analyzing the dispersion characteristics of waveforms transmitted in a rail-sleeper-railway ballast-lining medium, so as to obtain a separated train induced vibration field; Carrying out tunnel axial waveguide propagation simulation according to the train induced vibration field, and analyzing the modal propagation and attenuation rules of the train induced vibration field under waveguide constraint by modeling a tunnel lining structure as a tubular waveguide to obtain the modal dynamic stress distribution of the tunnel axial direction; And taking the modal dynamic stress distribution as a boundary condition, solving a wave equation by combining the stress release effect of the free surface of the tunnel portal, and inverting to obtain a dynamic stress tensor field in the surrounding rock of the tunnel portal.
4. The method for evaluating the slope instability risk of a railway tunnel portal based on image recognition according to claim 1, wherein the dynamic interaction analysis of the slope-tunnel structure is performed according to the dynamic differential displacement field and the dynamic stress tensor field to obtain a time-varying stability coefficient sequence, wherein the time-varying stability coefficient sequence is a ratio sequence of the anti-slip virtual work of a potential sliding body of the slope to the change of the load virtual work with time, and the method comprises the following steps: performing potential slip surface mode construction treatment according to the dynamic differential displacement field, and determining a slip surface shape penetrating through a tunnel portal by screening a continuous area with a non-zero covariant derivative in the dynamic differential displacement field and combining geometric boundary constraint of a tunnel lining to obtain a virtual displacement mode of a potential slip surface of a side slope; according to the virtual displacement mode of the potential sliding surface of the side slope and the dynamic stress tensor field, carrying out dynamic virtual work calculation, and respectively calculating anti-sliding virtual work resisting sliding and load virtual work driving sliding by convolving the effective stress perpendicular to the sliding surface to obtain a time sequence data set of the anti-sliding virtual work and the load virtual work; and performing time-varying stability analysis processing according to the time sequence data set, and obtaining a time-varying stability coefficient sequence representing the instantaneous safety state of the slope by dividing the anti-skid virtual work at the same moment by the load virtual work and arranging the ratio along a time axis.
5. The method for evaluating the risk of slope instability of a railway tunnel portal based on image recognition according to claim 1, wherein the step of performing the precursor recognition of instability according to the time-varying stability coefficient sequence, and determining whether to trigger a convergence threshold defined by the maximum allowable deformation of the tunnel lining structure by monitoring the residual accumulation effect and the attenuation characteristic of the passing clearance period of the train, to obtain a slope instability mode, comprises: performing gap period residual effect extraction processing according to the time-varying stability coefficient sequence, and performing band-pass filtering and integral operation on the sequence in the intermittent period of train load action to obtain a residual stability coefficient attenuation curve representing the fatigue and damage states of the structure; Performing accumulated damage and recovery analysis processing according to the residual stability coefficient attenuation curve, and calculating accumulated increment of residual effect in the adjacent train load period by fitting an exponential attenuation rule of the attenuation curve to obtain time-varying accumulated damage degree of the slope-lining coupling system; And carrying out destabilization mode discrimination processing according to the time-varying accumulated damage degree, and obtaining a slope destabilization mode by mapping the time-varying accumulated damage degree to a failure envelope surface taking the maximum allowable deformation of a tunnel lining structure as a boundary, judging whether to trigger the envelope surface and determining a leading mechanism of destabilization.
6. The method for evaluating the risk of slope instability of a railway tunnel portal based on image recognition according to claim 1, wherein the step of quantifying the operation risk according to the slope instability mode and outputting a risk evaluation result measured by the off-schedule vehicle interruption time comprises the steps of: performing instability development process simulation processing according to the side slope instability mode, and predicting critical time required by development to influence the safety of the tunnel lining structure under a dry precondition by combining the space directivity and the development rate of the side slope instability mode to obtain an instability development time line; according to the instability development time line, carrying out maintenance intervention simulation processing of the skylight period, and simulating the operation flow and the required time for reinforcing or supporting the potential sliding body in the skylight period by combining the length of the established skylight maintenance time and the construction process limit to obtain the maintenance intervention time; And according to the instability development time line and the time required by maintenance intervention, carrying out running interruption risk calculation, judging whether the time required by maintenance intervention exceeds the set skylight maintenance time, calculating the total delay time of the unscheduled running caused by line occupation of construction, and outputting a risk assessment result taking the unscheduled running interruption time as a measurement.
7. The image recognition-based railway tunnel portal slope instability risk assessment method according to claim 6, wherein the step of performing an instability development process simulation process according to the slope instability mode, predicting critical time required for development to influence safety of a tunnel lining structure under a dry precondition by combining spatial directivity and development rate of the slope instability mode, and obtaining an instability development timeline comprises: Performing parameter extraction processing according to the slope instability mode, and calculating an included angle between the slope instability mode and the axial direction of the tunnel and an effective slip crack surface dip angle by analyzing the space directivity of the slip crack surface in the instability mode to obtain a geometrical and mechanical parameter set of a potential slip body; Performing instability evolution law quantification treatment according to the geometric and mechanical parameter sets, and fitting a nonlinear evolution function describing the relation between the deformation quantity of the sliding body and time by coupling the development rate with the parameter sets to obtain a progressive deformation index of slope instability; and carrying out critical time prediction processing according to the progressive deformation index, substituting the progressive deformation index into a differential equation taking the maximum allowable deformation of the tunnel lining structure as a boundary condition for calculation, and predicting the time required for developing from the current state to the critical deformation amount to obtain a instability development time line.
8. The method for evaluating the risk of the slope instability of the railway tunnel portal based on the image recognition is characterized by comprising the following steps of: the acquisition module is used for acquiring video image data of a side slope area of a target railway tunnel portal, wherein the video image data comprises visual information of the side slope, a tunnel portal lining structure and a joint area, and waveform data of a train passing through, which is acquired by a sensor arranged on a sleeper at the inner side of the tunnel portal, is synchronously acquired; the construction module is used for constructing a slope surface deformation field according to the video image data, and performing constraint expansion and curvature analysis on space-time characteristic manifolds of the image by introducing geometric constraints of tunnel portal topography as manifold boundary conditions to obtain a dynamic differential displacement field of the slope surface; The simulation module is used for carrying out tunnel axial dynamic load propagation simulation according to the waveform data, and obtaining a dynamic stress tensor field in the surrounding rock of the tunnel portal through analyzing the dispersion relation of the vibration waveform in the steel rail-sleeper-railway ballast-lining medium and the waveguide effect of the tunnel tubular structure; The analysis module is used for carrying out dynamic interaction analysis of the side slope-tunnel structure according to the dynamic differential displacement field and the dynamic stress tensor field to obtain a time-varying stability coefficient sequence, wherein the time-varying stability coefficient sequence is a ratio sequence of the anti-slip virtual work of the side slope potential sliding body and the load virtual work changing along with time; the recognition module is used for recognizing the destabilizing precursor according to the time-varying stability coefficient sequence, and obtaining a slope destabilizing mode by monitoring the residual accumulation effect and the attenuation characteristic of the train passing clearance period and judging whether to trigger a convergence threshold defined by the maximum allowable deformation of the tunnel lining structure; And the evaluation module is used for quantifying the operation risk according to the slope instability mode and outputting a risk evaluation result taking the off-schedule vehicle interruption time as a measure.
9. The image recognition-based railway tunnel portal slope instability risk assessment system of claim 8, wherein the building module comprises: the first construction unit is used for carrying out space-time manifold construction processing according to the video image data, and obtaining a slope space-time characteristic manifold subjected to geometric constraint by identifying the edge contour of the tunnel portal structure and defining the boundary constraint as a non-deformable manifold; the second construction unit is used for carrying out constraint manifold expansion processing according to the side slope space-time characteristic manifold, and mapping the high-dimensional manifold to a measurable low-dimensional space by carrying out Riemann geometric expansion on the space-time characteristic manifold under the boundary constraint condition to obtain a side slope surface deformation field; and the third construction unit is used for carrying out differential displacement solving processing according to the slope surface deformation field, and obtaining a dynamic differential displacement field of the slope surface by calculating the tangent space covariant derivative of the surface deformation field relative to the tunnel lining fixed reference point.
10. The image recognition-based railway tunnel portal slope instability risk assessment system of claim 8, wherein the simulation module comprises: The first simulation unit is used for carrying out separation treatment on dynamic load components of the train according to the waveform data, and separating out direct vibration components transmitted by a rail-lining path and vibration components scattered by railway ballast-surrounding rock by analyzing the dispersion characteristics of the waveform transmitted in the rail-sleeper-railway ballast-lining medium, so as to obtain a separated train induced vibration field; The second simulation unit is used for carrying out tunnel axial waveguide propagation simulation according to the train induced vibration field, and analyzing the modal propagation and attenuation rule of the train induced vibration field under waveguide constraint by modeling a tunnel lining structure as a tubular waveguide to obtain the modal dynamic stress distribution of the tunnel axial direction; And the third simulation unit is used for taking the modal dynamic stress distribution as a boundary condition, solving a wave equation by combining the stress release effect of the free surface of the tunnel portal, and inverting to obtain a dynamic stress tensor field in the surrounding rock of the tunnel portal.

Description

Railway tunnel portal slope instability risk assessment method and system based on image recognition Technical Field The invention relates to the technical field of computers, in particular to a railway tunnel portal slope instability risk assessment method and system based on image recognition. Background In the field of railway transportation safety, the monitoring and risk assessment of the stability of a tunnel portal slope are key links for guaranteeing the operation safety of a line all the time, particularly in mountain railways with complex terrains, the tunnel portal is used as a junction transition area of the slope and an artificial structure, the stability of the tunnel portal is comprehensively influenced by the cyclic dynamic load of a train, geological conditions and climate factors, and obvious dynamic time variability and structural coupling are presented. The traditional evaluation method relies on periodic manual inspection or isolated sensor data, is difficult to effectively capture continuous evolution rules of deformation of the surface of a side slope and the internal stress state of a tunnel structure in time and space, is more incapable of quantifying potential threats of residual deformation accumulation to long-term stability in a passing clearance period of a train, and is difficult to provide sufficient basis for preventive maintenance decision due to the fact that the prior art tries to judge the stability by simplifying a mechanical model or a single data source, but is often incapable of fully considering dynamic interaction of the side slope-tunnel system and lacking accurate description of dynamic evolution process of a destabilizing precursor, so that risk evaluation results are limited in timeliness and accuracy. Based on the defects of the prior art, a method and a system for evaluating the instability risk of a railway tunnel portal slope based on image recognition are needed. Disclosure of Invention The invention aims to provide a railway tunnel portal slope instability risk assessment method and system based on image recognition so as to solve the problems. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: In a first aspect, the application provides a railway tunnel portal slope instability risk assessment method based on image recognition, which comprises the following steps: acquiring video image data of a side slope area of a target railway tunnel portal, wherein the video image data comprises visual information of the side slope surface, a tunnel portal lining structure and a joint area, and synchronously acquiring waveform data of a train passing through, wherein the waveform data are acquired by sensors arranged on sleeper on the inner side of the tunnel portal; constructing a slope surface deformation field according to the video image data, and performing constraint expansion and curvature analysis on space-time characteristic manifolds of the image by introducing geometric constraints of tunnel portal topography as manifold boundary conditions to obtain a dynamic differential displacement field of the slope surface; Carrying out tunnel axial dynamic load propagation simulation according to the waveform data, and inverting to obtain a dynamic stress tensor field in the surrounding rock of the tunnel portal by analyzing the dispersion relation of the vibration waveform in the steel rail-sleeper-railway ballast-lining medium and the waveguide effect of the tunnel tubular structure; Performing dynamic interaction analysis of a side slope-tunnel structure according to the dynamic differential displacement field and the dynamic stress tensor field to obtain a time-varying stability coefficient sequence, wherein the time-varying stability coefficient sequence is a ratio sequence of anti-slip virtual work and load virtual work of a side slope potential sliding body to change with time; Carrying out instability precursor identification according to the time-varying stability coefficient sequence, and obtaining a slope instability mode by monitoring the residual accumulation effect and the attenuation characteristic of the train passing clearance period and judging whether to trigger a convergence threshold defined by the maximum allowable deformation of the tunnel lining structure; And carrying out operation risk quantification according to the slope instability mode, and outputting a risk assessment result taking the off-schedule vehicle interruption time as a measure. In a second aspect, the application further provides a railway tunnel portal slope instability risk assessment system based on image recognition, which comprises the following steps: the acquisition module is used for acquiring video image data of a side slope area of a target railway tunnel portal, wherein the video image data comprises visual information of the side slope, a tunnel portal lining structure and a joint area, and waveform data of a train