A high-speed railway tunnel operation environment safety intelligent monitoring method and system

Abstract

This application provides an intelligent monitoring method and system for the operational environment safety of high-speed railway tunnels, relating to the field of intelligent monitoring technology. It acquires surrounding rock deformation data, target sound data, and tunnel environmental data within the high-speed railway tunnel; generates a stress field based on the surrounding rock deformation data; identifies normal sound signals, abnormal sound signals, and their corresponding anomaly types based on the target sound data and stress field, and determines the sound source location corresponding to each abnormal sound signal; performs cluster analysis on the tunnel environmental data, stress field, abnormal sound signals, their corresponding anomaly types, and sound source locations at the time the abnormal sound signals are acquired, obtaining clustering results; generates hazard warning information based on the clustering results; and uses a digital twin model to extrapolate the stress field, sound source location, and tunnel environmental state based on the hazard warning information, obtaining the operational environment monitoring results of the high-speed railway tunnel, thus achieving proactive predictive monitoring of the structural and environmental states of the high-speed railway tunnel.

Description

Technical Field

[0001] This application relates to the field of intelligent monitoring and intelligent control technology, and in particular to an intelligent monitoring method and system for the safety of the high-speed railway tunnel operation environment. Background Technology

[0002] During long-term operation, high-speed railway tunnels are affected by high-frequency vibrations from trains and aging of surrounding rock. The lining may gradually develop cracks or even partially fall off. Auxiliary equipment such as fans and cable trays in the tunnel are also prone to loosening and producing abnormal noises under continuous vibration. At the same time, problems such as the accumulation of harmful gases may occur due to poor ventilation or equipment aging.

[0003] Currently, there is a monitoring scheme for tunnel safety. This scheme deploys sound pressure sensors and displacement sensors inside the tunnel. It determines the presence of abnormal events by checking if the amplitude of the sound signal exceeds a preset threshold. Simultaneously, the displacement sensor collects the displacement of a single point on the lining surface. When the displacement exceeds a warning value, an alarm is triggered. The two types of sensors operate independently and output acoustic and displacement alarms respectively. However, this scheme has several significant drawbacks. For example, acoustic judgment relies solely on amplitude thresholds, making it difficult to distinguish between the normal airflow sounds of a passing train and dangerous abnormal sounds such as structural loosening or lining detachment, leading to frequent false alarms. Displacement monitoring only focuses on single-point deformation, failing to reflect the overall stress state of the cross-section and exhibiting slow response to early, minor deformations. More importantly, the independent nature of acoustic and deformation monitoring results in alarm information lacking causal support, hindering subsequent hazard diagnosis and condition prediction.

[0004] More importantly, acoustic monitoring and deformation monitoring are independent of each other, resulting in alarm information lacking causal support, making it difficult to support subsequent hazard diagnosis and status prediction. As a result, it is difficult for operational safety management to upgrade from passive alarm response to proactive control decision-making based on multi-source data fusion, affecting the accurate execution of control measures such as train dispatching, maintenance and repair. Summary of the Invention

[0005] The purpose of this application is to provide an intelligent monitoring method and system for the safety of the high-speed railway tunnel operation environment, so as to solve the problem that the existing technology has poor alarm reliability and cannot support the diagnosis and prediction of hidden dangers due to the separation of acoustic monitoring and deformation monitoring.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides an intelligent monitoring method for the safety of the high-speed railway tunnel operating environment, comprising:

[0007] Acquire data on surrounding rock deformation, target sound, and tunnel environment within high-speed railway tunnels;

[0008] Based on the surrounding rock deformation data, a stress field is generated;

[0009] Based on the target sound data and the stress field, normal sound signals, abnormal sound signals and their corresponding abnormality types are identified, and the sound source location corresponding to each abnormal sound signal is determined.

[0010] Density-based clustering algorithms were used to perform clustering analysis on the tunnel environment data, stress field, abnormal sound signals, corresponding anomaly types, and sound source locations at the time when the abnormal sound signals occurred, and the clustering results were obtained.

[0011] Based on the clustering results, a potential hazard warning message is generated;

[0012] Based on the aforementioned hazard warning information, a digital twin model is used to perform state simulation of the stress field, the location of the sound source, and the tunnel environment, thereby obtaining the monitoring results of the high-speed railway tunnel's operating environment.

[0013] Secondly, this application provides an intelligent monitoring system for the safety of high-speed railway tunnel operation environment, including:

[0014] The acquisition module is used to acquire data on surrounding rock deformation, target sound data, and tunnel environment data within high-speed railway tunnels.

[0015] The first generation module is used to generate a stress field based on the surrounding rock deformation data;

[0016] The identification module is used to identify normal sound signals, abnormal sound signals and corresponding abnormality types based on the target sound data and the stress field, and to determine the sound source location corresponding to each abnormal sound signal.

[0017] The analysis module is used to perform cluster analysis on the tunnel environment data, stress field, abnormal sound signals and corresponding anomaly types and sound source locations at the time when the abnormal sound signals are acquired using a density-based clustering algorithm, and obtain the clustering results.

[0018] The second generation module is used to generate potential hazard warning information based on the clustering results;

[0019] The simulation module is used to perform state simulation of the stress field, the location of the sound source, and the tunnel environment based on the hidden danger warning information using a digital twin model, so as to obtain the monitoring results of the high-speed railway tunnel's operating environment.

[0020] Thirdly, this application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the steps of the intelligent monitoring method for safe operation of a high-speed railway tunnel as described in the first aspect above.

[0021] Fourthly, this application provides a computer storage medium storing a computer program, which, when executed by a computer, implements the steps of the intelligent monitoring method for safe operation of a high-speed railway tunnel as described in the first aspect above.

[0022] The intelligent monitoring method for the operational environment safety of high-speed railway tunnels provided in this application has the following beneficial effects: By acquiring surrounding rock deformation data, sound data, and environmental data, a stress field reflecting the overall stress state of the tunnel cross-section is constructed. Sound signal identification is correlated with the stress field in time and space to achieve accurate classification and source localization of abnormal sound signals. On this basis, a density-based clustering algorithm is used to collaboratively analyze multi-dimensional features, automatically eliminate interference from normal operating conditions such as train passage, focus on major hidden dangers such as loosening of tunnel auxiliary equipment, lining detachment, and environmental anomalies, and generate hidden danger warning information. Finally, a digital twin model is used to comprehensively deduce structural stress, sound field propagation, and environmental conditions, and output forward-looking operational environment monitoring results, thereby improving alarm credibility and hidden danger diagnosis capabilities. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart illustrating an intelligent monitoring method for the operational environment safety of a high-speed railway tunnel, provided as an embodiment of this application;

[0025] Figure 2 This is a schematic diagram illustrating a specific implementation of a method for identifying normal sound signals, abnormal sound signals, and corresponding abnormality types provided in an embodiment of this application.

[0026] Figure 3 This is a schematic diagram of the structure of an intelligent monitoring system for the safety of the high-speed railway tunnel operation environment provided in an embodiment of this application. Detailed Implementation

[0027] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] The core of this application is to provide an intelligent monitoring method for the safety of the high-speed railway tunnel operating environment. A flowchart of one specific implementation method is shown below. Figure 1 As shown, the method includes:

[0029] Step 101: Obtain data on the deformation of the surrounding rock, the target sound data, and the tunnel environment data inside the high-speed railway tunnel.

[0030] Step 102: Generate a stress field based on the surrounding rock deformation data.

[0031] Step 103: Based on the target sound data and stress field, identify normal sound signals, abnormal sound signals and their corresponding abnormality types, and determine the sound source location corresponding to each abnormal sound signal.

[0032] Step 104: Use a density-based clustering algorithm to perform cluster analysis on the tunnel environment data, stress field, abnormal sound signals and corresponding anomaly types and sound source locations at the time when the abnormal sound signals occur, and obtain the clustering results.

[0033] Step 105: Generate hazard warning information based on the clustering results.

[0034] Step 106: Based on the hazard warning information, use a digital twin model to perform state simulation of the stress field, sound source location, and tunnel environment status to obtain the monitoring results of the high-speed railway tunnel's operating environment.

[0035] In step 101 above, the surrounding rock deformation data refers to the measurement data collected by the sensing device that reflects the degree of stress deformation of the tunnel lining under the pressure of the surrounding rock and the dynamic load of the train. The surrounding rock refers to the rock and soil mass around the tunnel after excavation that is affected by construction disturbance and stress redistribution. The small deformations generated inside the surrounding rock during long-term operation will be transmitted to the lining surface.

[0036] The target sound data refers to the time-series data of the sound signals inside the tunnel collected by the acoustic sensing module deployed in the tunnel cross section. The acoustic sensing module consists of multiple sound sensing units, which are distributed in different sound collection positions within the tunnel cross section. The number of sound collection positions is at least three to meet the requirements of multi-channel signals for subsequent sound source localization. The target sound data includes various sound components such as the friction sound of train wheels and rails, the airflow sound generated by air compression and expansion, and the transient abnormal sound generated by local detachment of tunnel lining or loosening of equipment.

[0037] Tunnel environmental data refers to the measurement data of air state parameters inside the tunnel collected by environmental sensing modules deployed in the tunnel cross section. The environmental sensing module consists of multiple environmental sensing units, and its deployment location is the same as that of the acoustic sensing module. The tunnel environmental data includes the concentration of harmful gases, temperature, and dust concentration.

[0038] In this embodiment of the invention, the response optical signals of each measuring point to the deformation of the surrounding rock are collected by fiber optic grating strings deployed on the surface of the tunnel lining. The response optical signals are demodulated to obtain the wavelength offset (i.e., the deformation data of the surrounding rock). The measuring point refers to the spatial location of each grating segment distributed along the fiber direction on the fiber optic grating string. Each grating segment constitutes an independent sensing unit, which is used to convert the tensile or compressive deformation of the lining at its location into the offset of the center wavelength of the reflected light of that grating segment. The wavelength offset refers to the offset value of the center wavelength of the reflected light at each measuring point relative to the initial reference wavelength of that measuring point under stress-free conditions. The magnitude of the offset value reflects the degree of deformation of the surrounding rock at that measuring point.

[0039] Acoustic sensing modules deployed along the same cross-section as fiber optic gratings collect sound signals from inside the tunnel, converting them into digital audio sequences (i.e., target sound data). Environmental sensing modules collect air parameters from inside the tunnel, converting these parameters into digital measurements (i.e., tunnel environmental data). It should be noted that all three types of data are time-aligned using the same clock source during acquisition, ensuring that the surrounding rock deformation data, target sound data, and tunnel environmental data acquired at the same time have consistent time stamps.

[0040] Step 102 above specifically includes the following steps:

[0041] Step 201: Calculate the surrounding rock strain corresponding to each measuring point based on the strain sensitivity coefficient corresponding to the fiber grating string and the wavelength offset of each measuring point in the surrounding rock deformation data. The fiber grating string includes multiple measuring points connected in series along the fiber direction.

[0042] Step 202: Using the spatial coordinates of each measuring point as a reference, interpolate the surrounding rock strain between adjacent measuring points according to the distance between adjacent measuring points to obtain the preliminary strain distribution. The spatial coordinates of each measuring point are calibrated based on the tunnel engineering coordinate system.

[0043] Step 203: Based on the continuity of the surrounding rock deformation, the preliminary strain distribution is smoothed and corrected to obtain the target strain distribution.

[0044] Step 204: Based on the elastic modulus of the surrounding rock, convert the surrounding rock strain corresponding to each location point in the target strain distribution into stress values. Based on each stress value, construct a stress field, where each location point includes each measuring point and the interpolation points between adjacent measuring points.

[0045] In step 201 above, the strain sensitivity coefficient can be understood as a fixed conversion constant predetermined during the manufacturing or calibration stage of the fiber optic grating string. This constant characterizes the correspondence between the deformation per unit length and the offset of the center wavelength of the reflected light. The strain sensitivity coefficient can be used to convert optical measurements into mechanical deformation.

[0046] The surrounding rock strain refers to the deformation per unit length of the surrounding rock at each measuring point after being subjected to force along the fiber direction.

[0047] In this embodiment, the wavelength offset of all measuring points on the fiber optic grating string at the current moment is read from the surrounding rock deformation data, the strain sensitivity coefficient of the fiber optic grating string that has been measured during the factory calibration stage is obtained, the wavelength offset of each measuring point is divided by the strain sensitivity coefficient, and the surrounding rock strain corresponding to each measuring point is calculated.

[0048] In step 202 above, the spatial coordinates of the measuring point refer to the three-dimensional position coordinates of the measuring point on the tunnel lining section with reference to the tunnel engineering coordinate system. The spatial coordinates of each measuring point include three components: mileage value, horizontal offset, and height value.

[0049] The tunnel engineering coordinate system refers to a spatial reference system established with the control piles on the stable bedrock outside the tunnel entrance as the origin, the longitudinal direction of the tunnel as the mileage direction, the horizontal direction of the cross-section as the transverse direction, and the vertical direction as the height direction.

[0050] Preliminary strain distribution refers to the discrete strain distribution covering the tunnel lining section area formed by spatial interpolation of the surrounding rock strain at each measuring point.

[0051] In this embodiment, multiple interpolation points are sequentially determined at preset intervals within the gap area between two adjacent measuring points. An interpolation point refers to the position to be filled by spatial interpolation calculation between two adjacent original measuring points. Based on the spatial coordinates of each measuring point in the tunnel engineering coordinate system, the straight-line distance between two adjacent measuring points is calculated according to the spatial coordinates of any two adjacent measuring points. The weight coefficients of the surrounding rock strain of each of the two adjacent measuring points in the interpolation calculation are assigned according to the straight-line distance. For example, one way to assign weight coefficients is that for any interpolation point, the weight coefficients corresponding to measuring points closer to the interpolation point are taken as larger values, and the weight coefficients corresponding to measuring points farther away from the interpolation point are taken as smaller values. This application does not limit the specific size of the weight coefficients. In actual practice, another set of weight coefficients with different values ​​can be set according to the spatial distance between adjacent measuring points.

[0052] Based on the assigned weight coefficients, the surrounding rock strain of two adjacent measuring points is weighted and summed. The summation result is used as the estimated surrounding rock strain of the corresponding interpolation point. After performing the above interpolation calculation on all adjacent measuring points in sequence, the surrounding rock strain of each measuring point and the estimated surrounding rock strain of each interpolation point are summarized into a preliminary strain distribution.

[0053] In step 203 above, the continuity of surrounding rock deformation can be understood as the physical characteristic that when the surrounding rock, as a continuous medium, is subjected to external forces, the deformation of each point inside it smoothly transitions point by point along the spatial direction, and the deformation between adjacent positions does not change abruptly or jump.

[0054] The target strain distribution refers to the continuous strain distribution that conforms to the law of continuous deformation of the surrounding rock after smoothing and correcting the preliminary strain distribution.

[0055] In this embodiment, based on the continuity of surrounding rock deformation, the changes in surrounding rock strain in adjacent areas of the preliminary strain distribution are checked point by point to see if there are step-like jumps. One way to check this is to compare the difference in surrounding rock strain between two adjacent points with a preset smoothing threshold. When the difference in surrounding rock strain exceeds the preset smoothing threshold, it is determined that there is a jump at that point. Half of the jump amplitude is used as a correction amount to adjust the surrounding rock strain within a step range on both sides in the opposite direction, so that the strain between adjacent areas is restored to a smooth transition relationship. After the adjustment is completed, the target strain distribution is obtained. This application does not limit the size of the preset smoothing threshold. Specifically, it can be an empirical value preset based on the mechanical properties of the surrounding rock material and the dimensions of the tunnel lining structure.

[0056] In step 204 above, the elastic modulus of the surrounding rock refers to the proportional constant between stress and strain in the elastic deformation stage of the surrounding rock material. This constant is determined in advance through rock mechanics tests during the tunnel geological exploration stage.

[0057] Stress value refers to the magnitude of the force exerted on a unit area within the surrounding rock after the target strain distribution is subjected to stress.

[0058] The stress field refers to the continuous force field that reflects the stress deformation distribution of the tunnel cross section, formed by arranging stress values ​​point by point along the surface and circumference of the tunnel lining.

[0059] In this embodiment of the application, the elastic modulus of the surrounding rock of the tunnel is obtained from the tunnel geological survey report. The strain of the surrounding rock at each location point in the target strain distribution is multiplied by the elastic modulus of the surrounding rock, thereby converting each location point into a stress value. The stress values ​​at each location point are arranged along the tunnel lining surface and circumference to construct a stress field that reflects the current stress state of the section.

[0060] This application embodiment processes the wavelength offset of each measuring point in the surrounding rock deformation data through physical quantity conversion, spatial interpolation, smoothing correction, and mechanical quantity conversion. It gradually constructs a continuous stress field reflecting the overall stress deformation state of the tunnel lining section from discrete optical wavelength measurements, which can improve the structural safety assessment from single-point displacement monitoring to full-section stress field analysis.

[0061] In step 103 above, the step "based on target sound data and stress field, identify normal sound signals, abnormal sound signals, and corresponding abnormality types" is as follows: Figure 2 As shown, the specific steps include the following:

[0062] Step 301: Divide the target sound data into multiple sound segments in a preset order, and extract the duration, sound acquisition location and frequency distribution characteristics of each sound segment.

[0063] Step 302: Compare the frequency distribution characteristics of each sound segment with the preset acoustic feature template, and determine the sound segments with a matching degree greater than or equal to the preset matching threshold as normal sound signals, and the sound segments with a matching degree less than the preset matching threshold as abnormal sound signals.

[0064] Step 303: Extract the stress fluctuation amplitude corresponding to the duration segment of each abnormal sound signal from the stress field, and the stress fluctuation amplitude of each abnormal sound signal is located in the same tunnel section as the sound acquisition location. Based on the synchronous change of the stress fluctuation amplitude of each abnormal sound signal and the corresponding signal waveform, determine the abnormal type corresponding to each abnormal sound signal.

[0065] In step 301 above, the preset order refers to the temporal sequence of capturing sound segments from front to back along the time axis. The sound acquisition location refers to the placement of each sound sensing unit in the acoustic perception module within the tunnel cross-section.

[0066] Frequency distribution characteristics refer to the distribution of sound wave energy across multiple frequency ranges within a sound segment, used to describe the strength of different tonal components in that sound segment.

[0067] In this embodiment of the application, starting from the starting point of the target sound data in a preset order, multiple sound segments are sequentially extracted along the time direction with a fixed duration as the length of a segmentation window and a fixed step size as the interval between adjacent segmentation windows. Each sound segment contains all the sampling points within the corresponding segmentation window. For each sound segment, the start and end times of the corresponding segmentation window are recorded as the duration of the sound segment, and the spatial coordinates of the sound sensing unit that acquires the sound segment are recorded as the sound acquisition position of the sound segment.

[0068] Perform a Fourier transform on the sub-digital audio sequence within each sound segment to obtain the energy distribution spectrum of the sound segment at each frequency point; divide the frequency range covered by the energy distribution spectrum into multiple continuous frequency intervals, accumulate the energy values ​​of all frequency points within each frequency interval to obtain the energy value of that frequency interval; combine the energy values ​​of each frequency interval to form the frequency distribution characteristics of the sound segment.

[0069] In step 302 above, the preset acoustic feature template refers to an acoustic feature reference sample pre-constructed based on a large amount of measured sound data when a train passes through a tunnel normally. The preset acoustic feature template includes the energy distribution patterns of train wheel-rail friction sound, air compression sound, and airflow disturbance sound in the frequency domain.

[0070] Matching degree refers to the degree of consistency between the frequency distribution characteristics of a sound segment and the energy distribution of a preset acoustic feature template in each frequency range. The higher the matching degree, the closer the sound segment is to the sound of a train running normally.

[0071] The preset matching threshold refers to the pre-set matching degree limit value used to distinguish between normal sound signals and abnormal sound signals. In this application embodiment, the size of the threshold is not limited. Specifically, it can be determined based on the statistical analysis of the minimum matching degree between a large number of measured sound samples under normal train operation and the preset acoustic feature template.

[0072] In this embodiment, the frequency distribution characteristics of each sound segment are compared with the energy values ​​of a preset acoustic feature template for each frequency interval to calculate the matching degree. For example, one way to calculate the matching degree is to first calculate the difference between the energy value of the sound segment in each frequency interval and the energy value of the corresponding interval of the preset acoustic feature template, then take the absolute value of the difference for each frequency interval, sum the absolute values ​​of each frequency interval to obtain the sum of differences, and finally take the reciprocal of the sum of differences as the matching degree value. The matching degree is compared with a preset matching threshold. Sound segments with a matching degree greater than or equal to the preset matching threshold are determined as normal sound signals, and sound segments with a matching degree less than the preset matching threshold are determined as abnormal sound signals.

[0073] In step 303 above, stress fluctuation amplitude refers to the maximum change in stress value at the same location in the stress field from the average stress value during the duration of the abnormal acoustic signal.

[0074] The same tunnel cross section refers to the spatial cross section in the tunnel engineering coordinate system where the mileage value corresponding to the sound acquisition position is equal to the mileage value of each position point in the stress field.

[0075] Synchronous change refers to the fact that the fluctuation trend of stress fluctuation amplitude and the fluctuation trend of abnormal sound signal waveform have corresponding start and end times and similar change rhythms on the time axis.

[0076] Anomaly type refers to the classification of the nature of the hidden danger corresponding to the abnormal sound signal. Anomaly type can be lining detachment type, tunnel auxiliary equipment loosening type, airflow disturbance type.

[0077] In this embodiment of the application, based on the duration of each abnormal sound signal and the sound acquisition location, a target location point with the same mileage value as the sound acquisition location is located in the stress field. The stress value sequence of the target location point from the start time to the end time of the abnormal sound signal is extracted, and the difference between the maximum and minimum values ​​in the stress value sequence is taken as the stress fluctuation amplitude.

[0078] The stress fluctuation amplitude variation curves of each abnormal acoustic signal are time-aligned with the signal waveforms, and the correspondence between the two at the start time, end time, and fluctuation transition time is compared. When the stress fluctuation amplitude shows a synchronous change consistent with the fluctuation rhythm of the signal waveform of the abnormal acoustic signal within the duration, the stress fluctuation amplitude is further compared with a preset amplitude threshold. When the stress fluctuation amplitude exceeds the preset amplitude threshold, the abnormal type corresponding to the abnormal acoustic signal is determined to be lining detachment. When the stress fluctuation amplitude does not exceed the preset amplitude threshold, the abnormal type corresponding to the abnormal acoustic signal is determined to be loosening of tunnel auxiliary equipment. This application does not limit the size of the preset amplitude threshold, and its value can be determined based on the upper limit of stress fluctuation caused by train dynamic load on the tunnel lining structure under normal operating conditions. When the stress fluctuation amplitude remains stable within the duration and does not show a synchronous change corresponding to the signal waveform, the abnormal type corresponding to the abnormal acoustic signal is determined to be airflow disturbance.

[0079] This application embodiment compares the temporal characteristics of the sound signal with the spatial characteristics of the stress field, uses acoustic template matching to initially distinguish between normal and abnormal signals, and then verifies the synchronous changes of stress fluctuation amplitude and sound signal waveform. The abnormal signals are further subdivided into structurally related anomaly types and non-structural disturbance types, thereby providing a classification result of abnormal sound signals with physical causal support for subsequent cluster analysis.

[0080] In step 103 above, the step "determining the location of the sound source corresponding to each abnormal sound signal" specifically includes the following steps:

[0081] Step 311: Extract the start time of each abnormal sound signal arriving at the corresponding sound acquisition location from the target sound data;

[0082] Step 312: Calculate the time difference between any two sound acquisition locations based on the spatial coordinates of each sound acquisition location and the corresponding signal start time. The spatial coordinates of the sound acquisition locations are calibrated based on the tunnel engineering coordinate system.

[0083] Step 313: Using the spatial coordinates of any two sound acquisition locations within the high-speed rail tunnel as focal points, and the product of each time difference and the speed of sound propagation in the high-speed rail tunnel as distance differences, construct multiple hyperboloids.

[0084] Step 314: Solve the hyperboloid equations for each hyperboloid to obtain the location of the sound source corresponding to the abnormal sound signal.

[0085] In step 311 above, for each sub-digital audio sequence corresponding to an abnormal sound signal in the target sound data, the sampling point where the amplitude first exceeds the average amplitude of the background noise is taken as the signal start point, and the timestamp corresponding to the signal start point is recorded as the signal start time of the abnormal sound signal at the sound acquisition location.

[0086] In step 312 above, the spatial coordinates of each sound acquisition location in the acoustic sensing module in the tunnel engineering coordinate system are first obtained. The number of sound acquisition locations is at least three, and the sound acquisition locations are divided into multiple sound acquisition location pairs. For each sound acquisition location pair, the start time of each abnormal sound signal in the two sound acquisition locations is subtracted to obtain the time difference between the arrival of the corresponding abnormal sound signal at the two sound acquisition locations.

[0087] In step 313 above, a hyperboloid refers to a surface formed by all points in a three-dimensional space with reference to the tunnel engineering coordinate system, where the difference in distance to two fixed points is constant. The two fixed points are the foci of the hyperboloid, and the difference in distance is the focal length difference of the hyperboloid.

[0088] In this embodiment, the spatial coordinates of the two sound acquisition positions in each sound acquisition position pair are respectively used as the two foci of the hyperboloid. The time difference corresponding to each sound acquisition position pair is multiplied by the speed of sound propagation in the air of the high-speed rail tunnel to obtain the focal length difference. Based on the focal coordinates of the above-mentioned foci and the corresponding focal length difference, multiple hyperboloids are constructed.

[0089] In step 314 above, the hyperboloid equation system refers to a system of equations composed of multiple hyperboloid equations. Each hyperboloid equation describes the spatial geometric constraints with the two sound acquisition positions corresponding to that hyperboloid as foci and the distance difference between the two hyperboloids as the focal length difference. The unknown in the hyperboloid equation system is a three-dimensional spatial coordinate. This unknown appears in all surface equations simultaneously, indicating that the point must simultaneously satisfy the spatial distance relationship of all hyperbolas. Therefore, solving the hyperboloid equation system yields the unique sound source location coordinates corresponding to the abnormal sound signal. Since the number of sound acquisition positions in this application is at least three, and the number of hyperboloids is at least three, the hyperboloid equation system is an overdetermined system of equations. The sound source location coordinates can be obtained by solving the overdetermined system of equations.

[0090] The location of the sound source refers to the three-dimensional spatial coordinates of the physical source that generates the abnormal sound signal in the tunnel engineering coordinate system.

[0091] In this embodiment, the surface equations corresponding to multiple hyperboloids are combined to form a hyperboloid equation system. By solving this hyperboloid equation system, a spatial coordinate solution that simultaneously satisfies all hyperboloid equations is obtained. This spatial coordinate solution is the location of the sound source corresponding to the abnormal sound signal.

[0092] The embodiments of this application enable the sound source localization result to be calibrated based on the tunnel engineering coordinate system through the above steps, thereby achieving the unification of the sound source location and the stress field spatial coordinate system. The positioning accuracy is not affected by the fluctuation of the sound signal intensity, and it can reliably locate abnormal sound sources over long distances in the tunnel.

[0093] Step 104 above specifically includes the following steps:

[0094] Step 401: Calculate the rate of change of stress based on the stress values ​​at the same location in the stress field at different data acquisition times.

[0095] Step 402: The stress change rate, characteristic parameters of the abnormal sound signal, anomaly type, spatial coordinates of the sound source location, and concentration of harmful gases in the tunnel environment data corresponding to the data acquisition time of the abnormal sound signal are respectively regarded as a type of feature dimension data. Based on the feature data corresponding to different data acquisition times, multiple clusters are divided using a density-based clustering algorithm, and each cluster corresponds to a tunnel working condition.

[0096] Step 403: Statistically analyze the value distribution range of the feature data corresponding to each data acquisition time within each cluster. Based on the value distribution range, identify the tunnel working condition type of each cluster. Integrate the feature data corresponding to the clusters belonging to the target tunnel working condition type into clustering results. The target tunnel working condition types include loose tunnel auxiliary equipment, lining detachment, and environmental anomaly.

[0097] In step 401 above, the data acquisition time refers to the time point when the surrounding rock deformation data, target sound data and tunnel environment data are collected synchronously, and each collection corresponds to a data acquisition time.

[0098] The stress change rate refers to the change in stress value at the same location point at a later data acquisition time relative to the stress value at a previous data acquisition time, and the ratio of this change to the time interval between the two data acquisition times.

[0099] In this embodiment of the application, any point in the stress field is selected as the same point being processed. The stress values ​​corresponding to the point at each of the two adjacent data acquisition times are read. The stress value at the later data acquisition time is subtracted from the stress value at the earlier data acquisition time to obtain the stress change. The stress change is divided by the time interval between the two data acquisition times, and the resulting ratio is the stress change rate of the point at the two adjacent data acquisition times.

[0100] In step 402 above, the characteristic parameter refers to the numerical index that can quantify the acoustic characteristics of the abnormal sound signal. For example, the characteristic parameter can be the statistical quantity of the energy value of each frequency interval in the frequency distribution characteristics of the abnormal sound signal within the duration period.

[0101] Feature dimension data refers to a specific numerical value or category label extracted or recorded at the same time when abnormal sound signals are acquired, for stress change rate, characteristic parameters of abnormal sound signals, anomaly type, spatial coordinates of sound source location, and concentration of harmful gas. Each category of data constitutes a feature dimension of cluster analysis.

[0102] Density-based clustering algorithms refer to algorithms that divide data points into clusters based on the density distribution of data points in the feature space. This algorithm groups data points in high-density areas into the same cluster and treats points in low-density areas as noise or into different clusters. A cluster is a set of data points that are close to each other in the feature space, and may include clusters of equipment loosening conditions, lining falling off conditions, abnormal environmental conditions, airflow disturbance conditions, and normal train passage conditions.

[0103] The feature space refers to a multi-dimensional space spanned by the coordinate axes of stress change rate, characteristic parameters of abnormal sound signals, anomaly type, spatial coordinates of sound source location, and concentration of harmful gas. Each set of feature dimension data corresponds to a data point in the feature space, and the values ​​of the data points on each coordinate axis are the numerical values ​​or category labels of the corresponding feature dimension data.

[0104] In this embodiment, for each data acquisition time where an abnormal sound signal exists, the stress change rate at each location point at that time is obtained as the feature dimension data in the dimension of stress change rate; the average or maximum value of the energy value of each frequency interval in the frequency distribution characteristics of the abnormal sound signal at that time is extracted as the feature dimension data in the dimension of feature parameter; the category label of the abnormal type corresponding to the abnormal sound signal at that time is recorded as the feature dimension data in the dimension of abnormal type; the spatial coordinates of the sound source location corresponding to the abnormal sound signal at that time are obtained as the feature dimension data in the dimension of sound source location; and the concentration of harmful gases in the tunnel environment data at that time is obtained as the feature dimension data in the dimension of harmful gas concentration. The above five types of feature dimension data together constitute a set of feature dimension data corresponding to the data acquisition time.

[0105] A density-based clustering algorithm is used to cluster the feature dimension data corresponding to different data acquisition times. Each feature dimension data is divided into multiple clusters according to the density of distribution in the feature space composed of stress change rate, feature parameters, anomaly type, sound source location and harmful gas concentration. This makes the feature dimension data of each group within the same cluster closer to each other in the feature space, and the feature dimension data of different clusters farther apart. Each cluster obtained corresponds to a tunnel working condition.

[0106] For example, five abnormal sound events occurred in the tunnel within a day. Two of these were clanging sounds caused by airflow blowing loose cable trays as trains passed, two were muffled sounds caused by small-scale lining detachment, and one was an environmental event where the concentration of harmful gases suddenly increased after a diesel train passed. The characteristic dimensions of these five events—stress change rate, anomaly type, sound source location, and harmful gas concentration—were placed into a feature space. The two sets of data related to the airflow blowing the cable trays would cluster closely together in the feature space due to their low stress change rate and the sound source location being in the same area of ​​the sidewall. The two sets of data related to the lining detachment would cluster into another separate dense cluster due to their significantly higher stress change rate and the sound source location being concentrated near the arch. The single set of data related to the environmental anomaly would be identified as an independent cluster or noise point because its harmful gas concentration value was far from other data points. Through this density clustering, the five abnormal sound events were automatically classified into three types of operating condition clusters: loose tunnel auxiliary equipment cluster, lining detachment cluster, and environmental anomaly cluster, eliminating the need for manual labeling.

[0107] In step 403 above, the tunnel condition type refers to the category of the tunnel's operating status identified based on the value distribution range of the feature dimension data, including abnormal types (lining detachment type, tunnel auxiliary equipment loosening type, airflow disturbance type), environmental abnormal types, and normal train passage type.

[0108] In this embodiment of the application, the minimum and maximum values ​​of the stress change rate, the category label with the highest frequency of occurrence of the anomaly type, the spatial distribution range of the sound source location, and the minimum and maximum values ​​of the harmful gas concentration are statistically analyzed for each data acquisition time within each cluster. The above statistical results are used as the value distribution range of the corresponding cluster.

[0109] Based on the value distribution range of each cluster, the tunnel working condition type corresponding to each cluster is identified. One identification method is as follows: when the minimum and maximum values ​​of the stress change rate are both maintained in the low range below the preset stress fluctuation threshold, and the category label with the highest frequency of abnormal type occurrence is airflow disturbance type, and the value of harmful gas concentration is within the normal range, then the cluster is identified as an airflow disturbance working condition. Here, the low range can be understood as the stress change rate being below the stress fluctuation level caused by normal train operation or slight airflow disturbance, and the normal range can be understood as the harmful gas concentration being below the allowable concentration specified by the tunnel operation safety standard. This application does not limit the value range of the above-mentioned low range and normal range. Specifically, the low range can be set according to the statistical upper limit value of the stress fluctuation amplitude caused by wheel-rail vibration and airflow disturbance of the tunnel lining structure under normal train operation conditions, and the normal range can be set according to the allowable upper limit value of the concentration of harmful gases such as carbon monoxide and hydrogen sulfide specified in the tunnel operation safety standard, and so on.

[0110] When the stress change rate fluctuates significantly, the category label with the highest frequency of abnormality is lining detachment, and the spatial distribution of the sound source location is concentrated near the tunnel lining structure cross-section, the cluster is identified as a lining detachment condition. Here, significant fluctuation can be understood as the stress change rate rapidly changing from a stable state to a drastic change mode in time, with its maximum value exceeding the preset stress fluctuation threshold. Near the tunnel lining structure cross-section can be understood as the shortest distance between the coordinates of the sound source location and the tunnel lining design outline not exceeding the preset structural association range. This structural association range can be set based on the superposition value of the tunnel lining thickness and construction error. This application does not limit the specific value of this structural association range.

[0111] When the stress change rate exhibits moderate fluctuations, the most frequent category label for abnormal types is "loosening of tunnel auxiliary equipment," and the spatial distribution of the sound source is dispersed within the installation area of ​​the tunnel auxiliary equipment, this cluster is identified as a loosening condition of tunnel auxiliary equipment. Moderate fluctuations can be understood as the stress change rate transitioning from a stable state to a changing pattern over time, with its maximum value not exceeding the preset stress fluctuation threshold but consistently higher than the baseline level of stress change rate under normal train operation. The installation area of ​​tunnel auxiliary equipment can be understood as the installation location of fans, cable trays, and lighting fixtures clearly marked in the tunnel design drawings on the tunnel cross-section, and the surrounding area centered on the installation location with a preset radius. This application does not limit the specific dimensions of this preset range.

[0112] When the maximum value of the harmful gas concentration dimension exceeds the normal range and shows a continuous upward trend, the cluster is identified as an abnormal environmental condition.

[0113] When the stress change rate dimension remains in the low range below the preset stress fluctuation threshold, no abnormal type label corresponding to the abnormal sound signal appears, and the concentration of harmful gas is within the normal range, the cluster is identified as a normal train operation condition.

[0114] After completing the identification of tunnel working condition types for each cluster, the feature dimension data corresponding to the clusters with tunnel working condition types of loosening of tunnel auxiliary equipment, lining detachment, and abnormal environment are integrated into clustering results.

[0115] This application embodiment performs density clustering of multi-dimensional feature data in the feature space and identifies the tunnel working condition type of each cluster. It can automatically distinguish between normal operation, airflow disturbance, and abnormal state with structural or environmental safety hazards. It accurately filters out three types of working condition data with safety value from a large amount of monitoring data: loose tunnel auxiliary equipment, lining detachment, and environmental anomalies. The clustering results focus on core hazards and have no redundant interference.

[0116] Step 105 above specifically includes the following steps:

[0117] Step 501: Based on the clustering results, count the total number of data acquisition times within each cluster, and calculate the duration of the abnormal sound signal set corresponding to each cluster based on the time interval between adjacent data acquisition times. The abnormal sound signal set consists of multiple abnormal sound signals corresponding to the same abnormal type.

[0118] Step 502: Calculate the hazard growth rate based on the variation range between the stress change rates at the corresponding times of data acquisition for each adjacent data point within each cluster.

[0119] Step 503: Determine the hazard level for each cluster based on the total number, duration, and hazard growth rate of each cluster.

[0120] Step 504: Integrate the hazard level, anomaly type, and sound source location corresponding to each cluster to generate hazard warning information.

[0121] In step 501 above, the total number refers to the number of data acquisition times belonging to the same cluster in the clustering results, reflecting the number of times the hidden dangers corresponding to each cluster are triggered within the corresponding data acquisition time range.

[0122] The set of abnormal acoustic signals refers to the set of abnormal acoustic signals corresponding to all data acquisition times within the same cluster in the clustering results. These abnormal acoustic signals have the same abnormality type.

[0123] Duration refers to the length of time spanned from the earliest data acquisition time to the latest data acquisition time in the abnormal sound signal set.

[0124] In this embodiment of the application, the total number of data acquisition times contained in each cluster in the clustering results is counted; the data acquisition times in each cluster are arranged in chronological order, and the total span between the earliest and latest data acquisition times is calculated. The total span is used as the duration of the abnormal sound signal set corresponding to the cluster.

[0125] In step 502 above, the hazard growth rate refers to an indicator that reflects the trend of stress change rate within the same cluster over time. The larger the hazard growth rate, the more the structural stress response of this type of hazard tends to intensify during the monitoring period.

[0126] In this embodiment of the application, for each cluster in the clustering results, the data acquisition times of each cluster are arranged in chronological order, and the stress change rate value corresponding to each data acquisition time is extracted. The stress change rate difference between the next data acquisition time and the previous data acquisition time is calculated in turn. The absolute value of the stress change rate difference between each adjacent time is taken and accumulated to obtain a cumulative change. The cumulative change is divided by the duration corresponding to the cluster, and the ratio obtained is the hidden danger growth rate of the cluster.

[0127] In step 503 above, the hazard level refers to the classification result of the severity of the hazard based on a comprehensive assessment of three indicators: total number, duration, and hazard growth rate.

[0128] In this embodiment of the application, for each cluster in the clustering results, the three indicators of total quantity, duration and hazard growth rate are compared with multiple pre-divided quantity intervals, multiple duration intervals and multiple growth rate intervals. The level scores corresponding to the intervals in which the three indicators fall are accumulated to obtain the total hazard score. The hazard level corresponding to the cluster is determined based on the total hazard score.

[0129] One specific implementation method for interval comparison is as follows: The total quantity is divided into low-frequency, medium-frequency, and high-frequency intervals. The level score corresponding to the low-frequency interval is the first quantity score, the level score corresponding to the medium-frequency interval is the second quantity score, and the level score corresponding to the high-frequency interval is the third quantity score, with the first quantity score being the smallest and the third quantity score being the largest. The duration is divided into short-term, medium-term, and long-term intervals. The level score corresponding to the short-term interval is the first duration score, the level score corresponding to the medium-term interval is the second duration score, and the level score corresponding to the long-term interval is the third duration score, with the first duration score being the smallest and the third duration score being the largest. The hazard growth rate is divided into stable, slow-increasing, and rapid-increasing intervals. The level score corresponding to the stable interval is the first growth rate score, the level score corresponding to the slow-increasing interval is the second growth rate score, and the level score corresponding to the rapid-increasing interval is the third growth rate score, with the first growth rate score being the smallest and the third growth rate score being the largest.

[0130] The total hazard score is obtained by summing the level scores corresponding to the intervals into which the total number, duration, and hazard growth rate fall. For example, when the total number falls into the high-frequency interval, the duration falls into the long-term interval, and the hazard growth rate falls into the rapid-increase interval, the three level scores corresponding to this cluster are the highest scores in their respective dimensions, and the sum of these scores yields the highest total hazard score. The total hazard score is then compared with multiple pre-divided total score intervals. For example, the total score intervals can be divided into low-risk, medium-risk, and high-risk intervals. When the total hazard score falls into the low-risk interval, the hazard level is determined to be at the general concern level; when it falls into the medium-risk interval, the hazard level is determined to be at the key investigation level; and when it falls into the high-risk interval, the hazard level is determined to be at the immediate alarm level. This application does not limit the specific boundary values ​​of the above interval divisions or the specific numerical values ​​of the level scores corresponding to each interval; these can be set according to the actual operation of the tunnel.

[0131] In step 504 above, the hazard warning information refers to the structured warning content formed by summarizing the hazard level, anomaly type and sound source location of the same cluster.

[0132] In this embodiment of the application, for each cluster in the clustering results, the hazard level, anomaly type, and spatial coordinates of the sound source location corresponding to the data acquisition time of each cluster are uniformly recorded to form a prompt information containing the hazard level, anomaly type, and sound source location, and the prompt information of all clusters is summarized into hazard prompt information.

[0133] Step 106 above specifically includes the following steps:

[0134] Obtain the hazard level, anomaly type, and sound source location for each type of hazard from the hazard warning information, and perform the following deduction operation for each type of hazard:

[0135] Step 601: When the sound source location corresponding to the same type of hidden danger includes multiple coordinate points, count the frequency of each coordinate point in the abnormal sound signal set corresponding to the current hidden danger, and take the coordinate point with the highest frequency as the representative sound source location corresponding to the current hidden danger.

[0136] Step 602: Using each representative sound source location as the load application point in the digital twin model, and using the preset load amplitude corresponding to the corresponding anomaly type and the load duration corresponding to the corresponding hazard level as input conditions, the stress distribution of the tunnel lining structure is deduced using the digital twin model to obtain the first change trajectory of the stress field corresponding to various hazards within the preset time period.

[0137] Step 603: Using each representative sound source location as the starting point of sound field propagation in the digital twin model, and using the geometric boundary of the tunnel lining structure and the material acoustic parameters as constraints on sound field propagation in the digital twin model, the spatial propagation path of the abnormal sound signal at the corresponding sound source location within a preset time period is deduced using the digital twin model, and the second change trajectory of the sound field coverage area corresponding to various hidden dangers is obtained.

[0138] Step 604: Based on tunnel environment data, use a digital twin model to deduce gas diffusion and temperature changes inside the high-speed railway tunnel, and obtain the third trajectory of tunnel environment state changes.

[0139] Step 605: Based on the first, second, and third change trajectories corresponding to various hidden dangers, generate the monitoring results of the high-speed railway tunnel's operating environment.

[0140] In step 106 above, each type of hazard refers to a hazard record in the hazard warning information that consists of a hazard level, anomaly type, and sound source location. Different groups of hazard records correspond to different anomaly types or different areas where sound sources are concentrated.

[0141] This step, which performs simulations for each type of hazard, ensures that when multiple hazards with different sound source locations and anomaly types are identified in the hazard warning information, each hazard can independently undergo structural stress and sound field propagation simulations based on its corresponding load conditions. This avoids the distortion of simulation results caused by the mixing of load conditions and sound source locations of multiple hazards. Specifically, step 106 may include steps 601-605, and more specifically:

[0142] In step 601 above, the frequency of occurrence refers to the number of times each spatial coordinate is located among the different spatial coordinates obtained after sound source localization of each abnormal sound signal in the set of abnormal sound signals corresponding to the same type of hidden danger. The representative sound source location refers to the coordinate point that best represents the spatial distribution characteristics of this type of hidden danger selected from multiple sound source coordinate points.

[0143] In this embodiment, the sound source location corresponding to each hazard is read from the hazard warning information. For hazards with two or more coordinate points, the number of times each coordinate point appears in the abnormal sound signal set corresponding to that type of hazard is counted, and the coordinate point with the highest frequency is determined as the representative sound source location. If there are multiple coordinate points with the highest frequency, the coordinate point located at the time of data acquisition with the largest corresponding stress fluctuation amplitude is selected as the representative sound source location.

[0144] In step 602 above, the load application point refers to the location in the digital twin model where an external force is applied to the tunnel lining structure. The load application point corresponds to the mapping location of the representative sound source in the digital twin model. The preset load amplitude refers to the load size pre-set according to the anomaly type. For example, a larger load amplitude corresponds to the lining detachment type, and a smaller load amplitude corresponds to the loosening type of tunnel auxiliary equipment. The specific value of the preset load amplitude is not limited in this application.

[0145] The load duration refers to the duration of load application determined according to the hazard level; the higher the hazard level, the longer the load duration. The first variation trajectory refers to the continuous trajectory of stress value changes over time at various points in the stress field within a preset time period.

[0146] In this embodiment, representative sound source locations are mapped to load application points in a digital twin model. Based on the anomaly type corresponding to the current hazard, a corresponding preset load amplitude is retrieved from a preset load lookup table. Based on the hazard level corresponding to the current hazard, a corresponding load duration is retrieved from a preset duration lookup table. Within the load duration, a gradually increasing load is applied to the load application point, with the load amplitude serving as the magnitude of the force. The load gradually increases from zero to the load amplitude and then remains constant until the load duration ends. During the load application process, a preset simulation step is performed at intervals along the time direction. Based on the displacement change of the load application point and the elastic modulus of the surrounding rock, the stress increment at the load application point and its surrounding locations within the simulation step is calculated. The stress increment is added to the stress value at the end of the previous simulation step to obtain the stress value at the end of the current simulation step. The stress values ​​at the end of each simulation step are connected in chronological order to obtain the first trajectory of the stress value at each location point in the stress field over a preset time period.

[0147] In step 603 above, the geometric boundary refers to the spatial contour shape of the inner surface of the tunnel lining structure in the digital twin model. Material acoustic parameters refer to the acoustic characteristic parameters of the tunnel lining material and air medium in the digital twin model regarding sound wave reflection, absorption, and transmission, such as acoustic impedance. The second variation trajectory refers to the continuous trajectory of the sound field coverage area changing over time within a preset time period.

[0148] One specific implementation method is as follows: Using the location of a representative sound source as the starting point for sound wave propagation, the frequency distribution characteristics in the characteristic parameters of the abnormal sound signal are set as the sound wave frequency of the sound source during the simulation process. A preset simulation step size is set at each interval along the time direction, with the simulation time as the calculation time point of the current step size. Based on the position and propagation direction of the sound wave in the tunnel space at the previous simulation time, the position where the sound wave first intersects the geometric boundary of the tunnel lining in the propagation direction is determined as the sound wave reflection point at the current simulation time. Based on the reflection coefficient in the material's acoustic parameters, the sound pressure value of the sound wave reaching the reflection point before reflection is multiplied by the reflection coefficient to obtain the sound pressure value of the reflected sound wave. The propagation direction of the reflected sound wave is determined according to the rule that the incident angle equals the reflection angle. Simultaneously, the sound pressure attenuation on the propagation path is calculated based on the absorption coefficient of the air medium, and the target sound pressure value is calculated by subtracting the sound pressure attenuation from the original sound pressure value. The sound wave propagation continues to be tracked to the reflection point at the next simulation time until the target sound pressure value is lower than a preset sound pressure threshold. The spatial region where the sound pressure level exceeds the preset sound pressure threshold at each simulation moment is determined as the sound field coverage range at the corresponding simulation moment. The sound field coverage ranges at each simulation moment are connected in chronological order to obtain the second change trajectory of the sound field coverage range within the preset time period.

[0149] In step 604 above, the initial value refers to the concentration and temperature of the harmful gas at the start of the simulation. The third change trajectory refers to the continuous trajectory of the distribution of harmful gas concentration and temperature in the tunnel over a preset time period.

[0150] In this embodiment, the concentration and temperature of harmful gases are extracted from the tunnel environment data at the most recent data acquisition time, serving as the initial values ​​of the air environment state in the digital twin model. Using the internal spatial geometry of the tunnel in the digital twin model and a preset ventilation airflow velocity as constraints for gas diffusion and temperature conduction, a preset simulation step size is established at each time interval along the time direction. The simulation time is used as the calculation time point for the current step size. Based on the harmful gas concentration and temperature distributions at the previous simulation time, the changes in harmful gas concentration due to gas diffusion and the temperature changes due to heat exchange between the air inside the tunnel and the lining surface are calculated at the current simulation time. These changes are then superimposed with the concentration and temperature values ​​from the previous simulation time to obtain the harmful gas concentration and temperature distributions at the current simulation time. Connecting the concentration and temperature distributions at each simulation time in chronological order yields the third trajectory of the tunnel environment state within a preset time period.

[0151] In step 605 above, the operational environment monitoring result refers to the comprehensive monitoring output formed by summarizing the three types of change trajectories: stress simulation, sound field simulation, and environmental simulation.

[0152] In this embodiment, the first, second, and third change trajectories corresponding to various hidden dangers are aligned on the time axis, so that the stress field distribution, sound field coverage, harmful gas concentration distribution, and temperature distribution at the same simulation moment correspond one-to-one. The time information, stress field distribution, sound field coverage, harmful gas concentration distribution, and temperature distribution of each simulation moment after alignment are integrated into the monitoring content corresponding to that simulation moment. The monitoring content corresponding to each simulation moment is arranged in chronological order to generate the operational environment monitoring results.

[0153] This application embodiment achieves proactive monitoring by predicting potential structural stress changes, sound field propagation range changes, gas diffusion and temperature changes caused by hidden dangers in the future, extending from the current hidden danger alarm to the prediction of future conditions.

[0154] Figure 3 This application provides a schematic diagram of a specific implementation of an intelligent monitoring system for the safety of high-speed railway tunnel operation, as illustrated in the embodiments of this application. Figure 3 The system may include:

[0155] The acquisition module 31 is used to acquire data on the deformation of the surrounding rock, the target sound data, and the tunnel environment data inside the high-speed railway tunnel.

[0156] The first generation module 32 is used to generate a stress field based on the surrounding rock deformation data;

[0157] The identification module 33 is used to identify normal sound signals, abnormal sound signals and corresponding abnormality types based on the target sound data and the stress field, and to determine the sound source location corresponding to each abnormal sound signal.

[0158] Analysis module 34 is used to perform cluster analysis on the tunnel environment data, stress field, abnormal sound signal and corresponding anomaly type and sound source location corresponding to the data acquisition time when the abnormal sound signal occurs using a density-based clustering algorithm, and obtain clustering results;

[0159] The second generation module 35 is used to generate hidden danger warning information based on the clustering results;

[0160] The simulation module 36 is used to perform state simulation of the stress field, the location of the sound source, and the tunnel environment based on the hidden danger warning information using a digital twin model, so as to obtain the monitoring results of the high-speed railway tunnel's operating environment.

[0161] This application provides an intelligent monitoring system for the safety of the high-speed railway tunnel operating environment, which is used to implement the aforementioned intelligent monitoring method for the safety of the high-speed railway tunnel operating environment. Therefore, the specific implementation of the intelligent monitoring system for the safety of the high-speed railway tunnel operating environment can be found in the embodiment section of the intelligent monitoring method for the safety of the high-speed railway tunnel operating environment described above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0162] This application also provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the steps of the intelligent monitoring method for safe operation of a high-speed railway tunnel as described above.

[0163] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, implements the steps of any of the above-described intelligent monitoring methods for safe operation of high-speed railway tunnels.

[0164] In one exemplary embodiment, the computer storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0165] The embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the embodiments of the intelligent monitoring method for safe operation of high-speed railway tunnels.

[0166] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0167] The above provides a detailed description of the intelligent monitoring method and system for the operational safety of high-speed railway tunnels provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. An intelligent monitoring method for the safety of high-speed railway tunnel operation environment, characterized in that, include: Acquire data on surrounding rock deformation, target sound, and tunnel environment within high-speed railway tunnels; Based on the surrounding rock deformation data, a stress field is generated; Based on the target sound data and the stress field, normal sound signals, abnormal sound signals and their corresponding abnormality types are identified, and the sound source location corresponding to each abnormal sound signal is determined. Density-based clustering algorithm was used to perform clustering analysis on the tunnel environment data, stress field, abnormal sound signals and corresponding anomaly types and sound source locations at the time when the abnormal sound signals were acquired, and the clustering results were obtained. Based on the clustering results, a potential hazard warning message is generated; Based on the aforementioned hazard warning information, a digital twin model is used to perform state simulation of the stress field, the location of the sound source, and the tunnel environment, thereby obtaining the monitoring results of the high-speed railway tunnel's operating environment.

2. The method according to claim 1, characterized in that, Based on the surrounding rock deformation data, a stress field is generated, including: Based on the strain sensitivity coefficient corresponding to the fiber grating string and the wavelength offset of each measuring point in the surrounding rock deformation data, the surrounding rock strain corresponding to each measuring point is calculated. The fiber grating string includes multiple measuring points connected in series along the fiber direction. Using the spatial coordinates of each measuring point as a reference, the surrounding rock strain between adjacent measuring points is interpolated according to the distance between adjacent measuring points to obtain the preliminary strain distribution. The spatial coordinates of each measuring point are calibrated based on the tunnel engineering coordinate system. Based on the continuity of the surrounding rock deformation, the preliminary strain distribution is smoothed and corrected to obtain the target strain distribution; Based on the elastic modulus of the surrounding rock, the surrounding rock strain corresponding to each location point in the target strain distribution is converted into stress values. Based on each stress value, a stress field is constructed, wherein each location point includes each measuring point and the interpolation point between adjacent measuring points.

3. The method according to claim 1, characterized in that, Based on the target sound data and the stress field, normal sound signals, abnormal sound signals, and corresponding abnormality types are identified, including: The target sound data is divided into multiple sound segments in a preset order, and the duration, sound acquisition location, and frequency distribution characteristics of each sound segment are extracted. The frequency distribution characteristics of each sound segment are compared with the preset acoustic feature template. Sound segments with a matching degree greater than or equal to the preset matching threshold are judged as normal sound signals, and sound segments with a matching degree less than the preset matching threshold are judged as abnormal sound signals. The stress fluctuation amplitude corresponding to the duration of each of the abnormal sound signals and located on the same tunnel section as the sound acquisition location is extracted from the stress field. Based on the synchronous change of the stress fluctuation amplitude of each of the abnormal sound signals and the corresponding signal waveform, the abnormality type corresponding to each of the abnormal sound signals is determined.

4. The method according to claim 1, characterized in that, Determining the location of the sound source corresponding to each of the aforementioned abnormal sound signals includes: Extract the start time of each abnormal sound signal arriving at the corresponding sound acquisition location from the target sound data; Based on the spatial coordinates of each sound acquisition location and the corresponding signal start time, the time difference between the arrival of each abnormal sound signal at any two sound acquisition locations is calculated, wherein the spatial coordinates of the sound acquisition locations are calibrated based on the tunnel engineering coordinate system; Using the spatial coordinates of any two sound acquisition locations within the high-speed rail tunnel as the focal points, and the product of the time difference and the speed of sound propagation in the high-speed rail tunnel as the distance difference, multiple hyperboloids are constructed. The hyperboloid equations of each hyperboloid are solved to obtain the location of the sound source corresponding to the abnormal sound signal.

5. The method according to claim 1, characterized in that, Density-based clustering algorithms were used to perform clustering analysis on the tunnel environmental data, stress field, abnormal acoustic signals, corresponding anomaly types, and sound source locations at the time of data acquisition when abnormal acoustic signals occurred. The clustering results included: The rate of change of stress is calculated based on the stress values ​​at the same location point in the stress field at different data acquisition times. The stress change rate, characteristic parameters of the abnormal sound signal, anomaly type, spatial coordinates of the sound source location, and concentration of harmful gases in the tunnel environment data corresponding to the data acquisition time of the abnormal sound signal are respectively used as a type of feature dimension data. Based on the feature data corresponding to different data acquisition times, multiple clusters are divided using a density-based clustering algorithm, and each cluster corresponds to a tunnel working condition. The distribution range of feature data values ​​corresponding to each data acquisition time within each cluster is statistically analyzed. Based on the value distribution range, the tunnel working condition type of each cluster is identified. The feature data corresponding to the clusters belonging to the target tunnel working condition type are integrated into the clustering result. The target tunnel working condition type includes the loosening type of tunnel auxiliary equipment, the lining falling off type, and the abnormal environment type.

6. The method according to claim 5, characterized in that, Based on the clustering results, hazard warning information is generated, including: Based on the clustering results, the total number of data acquisition times within each cluster is counted, and the duration of the abnormal sound signal set corresponding to each cluster is calculated based on the time interval between adjacent data acquisition times. The abnormal sound signal set consists of multiple abnormal sound signals corresponding to the same abnormal type. The hazard growth rate is calculated based on the variation range between the stress change rates at the corresponding times of data acquisition for each adjacent data point within each cluster. The hazard level for each cluster is determined based on the total number of clusters, the duration of the hazard, and the hazard growth rate. The hazard level, anomaly type, and sound source location corresponding to each cluster are integrated to generate hazard warning information.

7. The method according to claim 1, characterized in that, Based on the aforementioned hazard warning information, a digital twin model is used to perform state simulation of the stress field, the location of the sound source, and the tunnel environmental conditions, resulting in the monitoring results of the high-speed railway tunnel's operational environment, including: Obtain the hazard level, anomaly type, and sound source location corresponding to each type of hazard from the hazard warning information, and perform the following deduction operation for each type of hazard: When the sound source location corresponding to the same type of hidden danger includes multiple coordinate points, the frequency of each coordinate point in the abnormal sound signal set corresponding to the current hidden danger is counted, and the coordinate point with the highest frequency of occurrence is taken as the representative sound source location corresponding to the current hidden danger. Using each representative sound source location as the load application point in the digital twin model, and taking the preset load amplitude corresponding to the corresponding anomaly type and the load duration corresponding to the corresponding hazard level as input conditions, the stress distribution of the tunnel lining structure is deduced using the digital twin model, and the first change trajectory of the stress field corresponding to various hazards within the preset time period is obtained. Using each representative sound source location as the starting point of sound field propagation in the digital twin model, and the geometric boundary and material acoustic parameters of the tunnel lining structure as the constraints of sound field propagation in the digital twin model, the spatial propagation path of the abnormal sound signal at the corresponding sound source location within a preset time period is deduced using the digital twin model, and the second change trajectory of the sound field coverage corresponding to various hidden dangers is obtained. Based on the tunnel environment data, a digital twin model is used to deduce the gas diffusion and temperature changes inside the high-speed railway tunnel, and a third trajectory of tunnel environment change is obtained. Based on the first, second, and third change trajectories corresponding to various hidden dangers, the monitoring results of the high-speed railway tunnel's operating environment are generated.

8. An intelligent monitoring system for the safety of high-speed railway tunnel operation, characterized in that, include: The acquisition module is used to acquire data on surrounding rock deformation, target sound data, and tunnel environment data within high-speed railway tunnels. The first generation module is used to generate a stress field based on the surrounding rock deformation data; The identification module is used to identify normal sound signals, abnormal sound signals and corresponding abnormality types based on the target sound data and the stress field, and to determine the sound source location corresponding to each abnormal sound signal. The analysis module is used to perform cluster analysis on the tunnel environment data, stress field, abnormal sound signals and corresponding anomaly types and sound source locations at the time when the abnormal sound signals are acquired using a density-based clustering algorithm, and obtain the clustering results. The second generation module is used to generate potential hazard warning information based on the clustering results; The simulation module is used to perform state simulation of the stress field, the location of the sound source, and the tunnel environment based on the hidden danger warning information using a digital twin model, so as to obtain the monitoring results of the high-speed railway tunnel's operating environment.

9. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement an intelligent monitoring method for the safety of the high-speed railway tunnel operating environment as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The system contains a computer program that, when executed by a computer, implements an intelligent monitoring method for the safety of the high-speed railway tunnel operating environment as described in any one of claims 1 to 7.

Images