Expressway concealed goaf global perception prediction and emergency linkage early warning method

Abstract

The present application belongs to the field of traffic safety and intelligent monitoring technology, and discloses a highway hidden goaf global perception prediction and emergency linkage early warning method, which comprises the following steps: continuously collecting multi-physical field signals such as roadbed microseismic vibration, ground resistivity and pore water pressure, extracting coupled evolution characteristics through cross-field time sequence correlation analysis, and identifying hidden abnormal sections; starting directional detection according to the marking information, and inversely generating a goaf spatial distribution map; combining the map and the characteristic sequence to construct a instability criterion set, substituting real-time signals into the instability criterion set to perform approximation degree operation; when the approximation degree reaches the graded response threshold, corresponding grade warning instructions are automatically generated and sent to the traffic control terminal; the real-time performance and accuracy of the goaf instability risk warning are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of traffic safety and intelligent monitoring technology, and more specifically, to a method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways. Background Technology

[0002] Existing goaf detection technologies face bottlenecks in the early identification of concealed goaf areas, a problem particularly pronounced in high-traffic environments. Specifically, when underground goaf areas are located deep within the roadbed and the overlying soil and rock layers are relatively intact, traditional surface deformation monitoring methods often fail to capture early, subtle changes. These concealed goaf areas may experience sudden instability under long-term loads and environmental influences. For instance, when mountain highways traverse historical mining areas, abandoned small mines and irregular mining tunnels often lack detailed records. Under the influence of rainy season groundwater level changes or repeated heavy traffic, the overlying soil and rock in these concealed goaf areas may undergo gradual damage, making them difficult to detect in a timely manner. This concealment not only... This makes it impossible for existing fixed-point monitoring equipment to effectively cover potential risk areas, and it also fails to provide sufficient early warning time when the goaf is in a critical state of instability. The current mainstream approach to solving this problem relies on high-density drilling verification and long-term settlement observation, obtaining underground structure information through physical exploration and establishing a monitoring network. However, this solution based on traditional exploration not only requires a large investment of manpower and resources and a long monitoring cycle, but is also limited by the spatial limitations of exploration points and the insufficient timeliness of monitoring data. This makes it difficult to adapt to the large-scale and dynamic safety management needs of highways, and it cannot meet the urgent requirements of modern intelligent transportation for real-time risk prediction. Ultimately, this restricts the intelligent development of the safety prevention and control system for highway goaf areas.

[0003] In view of this, the present invention proposes a method for comprehensive perception, prediction and emergency response early warning of hidden mining subsidence areas along highways to solve the above problems. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways, comprising: Step S1: Continuously collect multi-physics field original response signals of deep soil and rock masses in the roadbed. The multi-physics field original response signals include micro-vibration waveform signals, ground resistivity change signals and pore water pressure fluctuation signals. Step S2: Perform cross-physical field time-series correlation analysis on the original response signals of the multi-physics field, extract the coupling evolution feature sequence between signals of different physical fields, identify the progressive mechanical degradation sections in the deep soil and rock mass of the roadbed based on the coupling evolution feature sequence, and generate hidden anomaly section marking information. Step S3: Initiate a directional detection and scanning task based on the marking information of the hidden anomaly section. Obtain deep structural echo data of the anomaly section through the coordinated work of road traffic load excitation and stratum response sensing array. Perform cavity geometry inversion calculation on the deep structural echo data to generate a spatial distribution map of the hidden goaf. Step S4: Based on the spatial distribution map of the hidden goaf and the coupled evolution characteristic sequence, construct the instability critical state criterion set, substitute the real-time collected multi-physics field original response signal into the instability critical state criterion set to perform time-by-time approximation degree calculation, and generate goaf instability approximation degree time series data. Step S5: When the proximity value of any segment in the time series data of the instability proximity of the goaf reaches the preset graded response threshold, an emergency linkage early warning instruction of the corresponding level is generated according to the threshold level reached and sent to the highway traffic control facility terminal.

[0005] The technical effects and advantages of this invention's method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways: This invention extracts coupled evolutionary feature sequences through cross-physics field temporal correlation analysis, enabling the identification of early, subtle changes in the progressive mechanical degradation of deep roadbed soil and rock. This overcomes the technical bottleneck of insufficient sensitivity of traditional single monitoring methods to hidden deep anomalies. By constructing a set of instability critical state criteria that integrates structural vulnerability levels and dynamic degradation rate indicators, it performs time-period approximation calculations on real-time multiphysics field signals and triggers graded emergency linkage early warning commands. This solves the technical problems of insufficient timeliness and inability to provide sufficient early warning time in instability critical states in traditional long-term observation methods. It realizes the transformation from passive response to proactive prediction of hidden mining areas on highways, avoiding the drawbacks of traditional methods that rely on high-density drilling verification, large-scale investment of manpower and resources, and long monitoring cycles. It significantly improves the real-time performance and accuracy of instability risk early warning in mining areas and promotes the intelligent development of the safety control system for mining areas on highways. Attached Figure Description

[0006] Figure 1 This is a schematic diagram of the method for comprehensive perception, prediction, and emergency response early warning of hidden mining areas on highways according to the present invention. Detailed Implementation

[0007] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Example

[0008] Please see Figure 1As shown in this embodiment, the method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways includes: Step S1: Continuously collect multi-physics field original response signals of deep soil and rock masses in the roadbed. The multi-physics field original response signals include micro-vibration waveform signals, ground resistivity change signals and pore water pressure fluctuation signals. When highways traverse historical mining areas, the underground hidden goaf areas, lacking detailed records, may experience sudden instability under long-term loads and environmental factors. Traditional surface deformation monitoring methods cannot capture early, subtle changes. Therefore, this solution deploys a ground response sensing array at the road shoulder to continuously collect multi-physics field raw response signals from the deep soil and rock mass of the roadbed. These signals include micro-vibration waveforms, resistivity changes, and pore water pressure fluctuations, enabling all-weather detection of early anomalies in hidden goaf areas.

[0009] Specifically, surface-anchored sensing probe groups are driven at equal intervals along the outer slope toe of the highway shoulder to construct a ground response sensing array. The initial spacing between adjacent surface-anchored sensing probe groups is set at 10 meters. Implementers can adjust the spacing from 5 to 20 meters depending on the complexity of the geological conditions of the road section. Smaller spacing is used in sections with complex geological conditions and frequent historical mining activities to increase sensing density, while larger spacing is used in sections with relatively simple geological conditions to reduce construction costs. Each surface-anchored sensing probe group includes micro-vibration pickup elements fixed at different depths on the probe rod, ring electrode pairs, and micro-permeability sensing units. These three types of sensing elements are respectively used to collect micro-vibration waveform signals, ground resistivity change signals, and pore water pressure fluctuation signals. These three types of signals reflect the changes in the mechanical state of the underground rock and soil from different physical field perspectives. The surface-anchored sensing probe group is implanted into the natural soil outside the road shoulder through static pressure penetration. The penetration depth is determined by penetrating the influence range of the roadbed and reaching the underlying stable strata. The specific depth is determined by the stratum profile in the roadbed survey report, which is usually 8 to 25 meters below the roadbed surface. The top of the surface-anchored sensing probe group is flush with the ground surface, does not occupy the driving lane and road shoulder passage space, and does not affect normal traffic operation.

[0010] Adjacent surface-anchored sensing probe groups are connected to the roadside aggregation node via shallowly buried signal transmission cables along the toe of the road shoulder slope to avoid disturbing the road shoulder structure layer during construction. The roadside aggregation node timestamps and synchronizes the raw electrical signals received from each surface-anchored sensing probe group. Specifically, the roadside aggregation node has a built-in navigation satellite system timing module that uses the satellite clock as a unified time reference. A high-precision timestamp is added to each access signal at the recording time, with a time synchronization accuracy better than 1 millisecond. This generates a synchronized data stream with a unified time reference, ensuring the time alignment accuracy of each signal channel during subsequent cross-physical field time-series correlation analysis.

[0011] The roadside convergence node separates the synchronously acquired data stream into micro-vibration waveform signal channels, ground resistivity change signal channels, and pore water pressure fluctuation signal channels according to preset physical field classification labels. Data from each channel is arranged and encapsulated according to the spatial numbering of the surface-anchored sensing probe group, generating multi-physics original response signals. These signals are then uploaded to the edge computing unit for further processing via a wireless communication module. The standard acquisition frequency for the micro-vibration waveform signal is set to 200 Hz, while the acquisition frequencies for the ground resistivity change signal and the pore water pressure fluctuation signal are both set to 10 Hz to match the characteristic frequency bands of each physical field signal.

[0012] It should be noted that the range and sensitivity of each sensing element in this embodiment are configured according to the physical properties of the stratum.

[0013] Step S2: Perform cross-physical field time-series correlation analysis on the original response signals of the multi-physics field, extract the coupling evolution feature sequence between signals of different physical fields, identify the progressive mechanical degradation sections in the deep soil and rock mass of the roadbed based on the coupling evolution feature sequence, and generate hidden anomaly section marking information. Because the amplitude of changes in a single physical field signal is often weak during the early, gradual instability process of a concealed goaf, making it difficult to distinguish from background noise, while the coordinated changes of multiple physical field signals within the same time period can reliably indicate a substantial mechanical degradation event within the soil and rock mass. Therefore, this scheme performs cross-physical field time-series correlation analysis on the original multi-physical field response signals, and identifies gradual mechanical degradation sections by extracting the coupling evolution characteristics between different physical field signals.

[0014] Specifically, the micro-vibration waveform signal is segmented according to the depth layers of the surface-anchored sensing probe group. When involving depth-layered analysis, the surface-anchored sensing probe group is simply referred to as the vertical sensing probe rod group; these are different ways of describing the same device. The name "vertical sensing probe rod group" emphasizes its structural characteristics of being vertically embedded into the strata and arranged along the depth direction. Each depth layer corresponds to the stratum segment between two adjacent micro-vibration pickup elements on the probe rod. The waveform energy attenuation rate for each depth layer within a continuous acquisition cycle is extracted. The waveform energy attenuation rate is calculated as follows: the root mean square amplitude of the micro-vibration waveform at that depth layer is taken in each acquisition cycle. The ratio of the difference in root mean square amplitude between two adjacent acquisition cycles to the root mean square amplitude of the previous acquisition cycle is the waveform energy attenuation rate. A positive value indicates energy attenuation, and a negative value indicates energy enhancement. The acquisition cycle is set to 60 seconds, which can be adjusted within the range of 30 to 300 seconds according to the actual monitoring accuracy requirements. Simultaneously, resistivity temporal gradient values ​​are extracted from the resistivity change signal according to the same depth layer. The resistivity temporal gradient value is the ratio of the difference in resistivity measurements at the same depth layer within adjacent acquisition cycles to the acquisition cycle duration, reflecting the rate of change in formation water content or structural looseness. Pore water pressure fluctuation signals are extracted from the pore water pressure fluctuation signal according to the same depth layer, including pressure oscillation frequency and amplitude envelope. The pressure oscillation frequency is the number of fluctuations in the pore water pressure signal exceeding the reference value ±5 kPa threshold within a single acquisition cycle, and the amplitude envelope is the difference between the maximum and minimum values ​​of the pressure signal within the same acquisition cycle.

[0015] For the same depth layer, the waveform energy attenuation rate, resistivity temporal gradient value, and pressure oscillation frequency are compared cycle by cycle along the time axis. The specific method for obtaining coupled response event points includes: determining the direction of change of the waveform energy attenuation rate between adjacent acquisition cycles. The direction of change indicator is one of three states: positive increase, negative decrease, or stable. A positive increase indicates that the waveform energy attenuation rate has increased compared to the previous cycle, and a negative decrease indicates that it has decreased compared to the previous cycle. A change in the absolute value of either indicator exceeding 0.5% is considered non-stationary; otherwise, it is considered stable. Simultaneously, the direction of change of the resistivity temporal gradient value and the pressure oscillation frequency between adjacent acquisition cycles is determined separately, using the same determination method as for the waveform energy attenuation rate. The stability determination threshold for the resistivity temporal gradient value is 0.2 ohm-meters per minute, and the stability determination threshold for the pressure oscillation frequency is 1 oscillation per cycle.

[0016] The ternary consistency determination is performed on the direction indicators of waveform energy attenuation rate, resistivity temporal gradient value, and pressure oscillation frequency. When all three direction indicators point to a positive increase or a negative decrease within the same acquisition period, it is determined to be a strong coordinated change trend, indicating that the three physical field signals change synchronously in the same direction, and the mechanical state of the soil and rock undergoes a coordinated response. When two of the three direction indicators point to the same direction and the third remains stable, it is determined to be a secondary coordinated change trend, indicating that the two physical field signals change synchronously while the third signal changes insignificantly, which also has indicative significance. The acquisition periods determined to be strong or secondary coordinated change trends are marked as coupling response event points.

[0017] For each coupled response event point, the waveform energy attenuation rate, resistivity temporal gradient value, and pressure oscillation frequency are read during the acquisition period. The values ​​of the three physical field signals are then weighted, mapped, and summed to obtain the event intensity value for that coupled response event point. The specific weighting method is as follows: First, the values ​​of the three physical field signals are divided by their respective dimensionally normalized reference values ​​to eliminate dimensional differences. These normalized reference values ​​correspond to the typical response amplitude of each physical field signal under progressive degradation, rather than the background fluctuation amplitude under normal operating conditions. The normalized reference value for the waveform energy attenuation rate is 0.05, the normalized reference value for the resistivity temporal gradient value is 1.0 ohm-meter per minute, and the normalized reference value for the pressure oscillation frequency is 3 times per cycle. These normalized reference values ​​are set based on the typical response amplitude of each physical field signal under progressive formation degradation conditions and can be adjusted by the implementer based on field calibration data. Then, the normalized values ​​of the three physical field signals are multiplied by their corresponding weighting coefficients and summed. The weighting coefficient for the rate of decrease is 0.4, the weighting coefficient for the gradient value of resistivity time series change is 0.35, and the weighting coefficient for the pressure oscillation frequency is 0.25. The sum of the three weighting coefficients is 1.0. The micro-vibration waveform signal is the most direct and sensitive mechanical response to the initiation and propagation of micro-cracks inside the rock and soil mass. It is the most core physical field indicator reflecting the progressive failure of the overlying rock mass in the goaf area. Therefore, it is given the highest weight of 0.4. The resistivity change signal can reflect the change of the water state of the strata and the degree of fracture connectivity. It is an important indirect indicator of the loosening of the rock and soil structure. It is given the second highest weight of 0.35. The pore water pressure fluctuation signal mainly reflects the disturbance state of the groundwater seepage field. Its change is relatively lagging and is greatly affected by external environmental factors such as rainfall. Therefore, it is given the lowest weight of 0.25. Furthermore, for coupled response event points identified as exhibiting a strong coordinated change trend, their weighted summation result is directly used as the event intensity value. For coupled response event points identified as exhibiting a secondary coordinated change trend, since only two physical field signals show coordinated changes while the third signal remains stable, their reliability in indicating soil and rock mechanical degradation is lower than that of strong coordinated change trends. Therefore, the weighted summation result is multiplied by a reduction factor of 0.6 to obtain the event intensity value, in order to distinguish the information content difference between the two types of coordinated trends. The event intensity values ​​of each coupled response event point are arranged in chronological order to form an event intensity value sequence corresponding to the coupled evolution characteristic sequence, which is used for the subsequent construction of the instability critical state criterion set.

[0018] Coupled response event points appearing sequentially within a preset statistical time window are chained together to form a coupled evolution characteristic sequence. The statistical time window is set to 24 hours, which can be adjusted by the implementer within the range of 6 to 72 hours according to the monitoring cycle requirements. The coupled evolution characteristic sequence is scanned segment by segment along the longitudinal direction of the roadbed at the corresponding positions of each vertical sensing probe group. Continuous sections where the cumulative event density of the coupled evolution characteristic sequence exceeds a preset density threshold are identified as progressive mechanical degradation sections. The preset density threshold is set at 5 or more coupled response event points appearing within each 24-hour statistical time window, which can be adjusted by the implementer within the range of 3 to 10 based on the engineering geological background and historical monitoring data. The start and end station numbers and depth range of the progressive mechanical degradation sections are recorded, generating hidden anomaly section marker information. Through the above cross-physical field collaborative analysis, the short-term disturbance of a single physical field caused by traffic load and the continuous collaborative response of multiple physical fields caused by the progressive degradation of soil and rock are effectively distinguished, solving the technical problem of difficulty in identifying early weak change signals in hidden mining subsidence areas.

[0019] Step S3: Initiate a directional detection and scanning task based on the marking information of the hidden anomaly section. Obtain deep structural echo data of the anomaly section through the coordinated work of road traffic load excitation and stratum response sensing array. Perform cavity geometry inversion calculation on the deep structural echo data to generate a spatial distribution map of the hidden goaf. After identifying the progressively degraded mechanical zones, it is necessary to further clarify the specific spatial morphology of the underground goaf to provide accurate geometric parameters for subsequent instability risk assessment. This scheme utilizes the wheel loads of heavy-duty vehicles in actual traffic as a natural excitation source, combined with a sensing array under high-frequency acquisition conditions, to acquire deep structural echo data using a passive source detection method, achieving precise detection of hidden goaf areas without interrupting traffic operations.

[0020] Specifically, the starting and ending chainages of the progressively degraded mechanical zone are read from the hidden anomaly zone marking information. The vertical sensing probe rods within the corresponding chainage range are switched to high-frequency acquisition mode. In high-frequency acquisition mode, the sampling frequency of the micro-vibration pickup element is increased to N times the conventional acquisition frequency, where N is taken as an empirical value of 5, i.e., the high-frequency acquisition frequency is 1000 Hz, to meet the Nyquist sampling requirements for capturing the high-frequency reflected wave group in the transient vibration response of the stratum. The implementer can adjust the value of N according to the target detection depth, with a value ranging from 3 to 10: when the detection depth is shallow, the dominant frequency of the target reflected wave is high, requiring a higher sampling frequency to meet the waveform restoration requirements, so a larger value of N can be taken; when the detection depth is large, the dominant frequency of the target reflected wave decreases after propagation through the stratum, requiring a relatively lower sampling frequency, so a smaller value of N can be taken.

[0021] During the road passage period in the progressively degraded mechanical zone, the wheel load of heavy-duty vehicles passing through is used as the natural excitation source. A road pressure sensor is embedded in the road structure layer to detect vehicle axle load and passage time. When the road pressure sensor detects that a heavy-duty vehicle axle has passed through the progressively degraded mechanical zone, the judgment condition is that the single axle load exceeds 100 kN. Simultaneously, the vertical sensing probe group in high-frequency acquisition mode is triggered to record the transient vibration response waveform of the stratum caused by the wheel load impact. The recording duration is set to 5 seconds after the wheel load passes, to fully capture the arrival time of reflected waves from each interface in the stratum. The transient vibration response waveforms recorded by multiple vertical sensing probe groups under the same wheel load excitation event are arranged according to spatial location and reception time difference, forming a passive source detection record set with the wheel load application point as the virtual seismic source and the vibration pickup elements of each probe group as the receiving point. This passive source detection record set is the deep structure echo data.

[0022] The arrival time and amplitude characteristics of each transient vibration response waveform are extracted from the deep structure echo data. The arrival time is the moment when the waveform first exceeds three times the root mean square amplitude of the background noise. The propagation velocity distribution of the direct wave is calculated according to the geometric relationship between the virtual source position and the receiver position. Specifically, the average wave velocity of the path corresponding to each receiver point is the ratio of the spatial straight-line distance from the virtual source to the receiver point to the difference between the arrival time and the start recording time. The spatial straight-line distance takes into account the horizontal offset and the vertical depth difference between the virtual source and the receiver point to ensure the physical accuracy of the wave velocity calculation. The average wave velocity of all receiver points is spatially interpolated to establish the initial velocity structure profile of the progressive mechanical degradation section.

[0023] In the initial velocity structure profile, regions with abnormally low wave velocity values ​​are scanned layer by layer. The marking method for suspected cavity response regions includes: extracting the transverse wave velocity distribution curve of each depth layer along the depth direction of the initial velocity structure profile, reading the wave velocity value point by point according to the spatial location of each transverse wave velocity distribution curve, and simultaneously reading the wave velocity values ​​of the adjacent layers above and below the same depth layer at the same spatial location, and taking the arithmetic mean of the wave velocity values ​​of the adjacent layers above and below as the reference wave velocity value of the adjacent layer at that spatial location. The actual wave velocity value at this spatial location is calculated by subtracting the wave velocity value of the adjacent layer to obtain the wave velocity deviation. When the wave velocity deviation is negative and its absolute value exceeds the deviation threshold obtained by multiplying the wave velocity value of the adjacent layer by a preset ratio, the spatial location is marked as a wave velocity abnormally low point. The preset ratio is 20%, that is, when the wave velocity at a certain location is more than 20% lower than the wave velocity of the adjacent layer, it is judged as an abnormally low point. The wave velocity in the cavities or loose filling areas of the formation is usually 30% to 60% lower than that of the surrounding intact rock and soil. 20% is taken as a conservative judgment threshold to avoid missed detection. The implementer can adjust it within the range of 10% to 30% according to the local formation physical parameters. Scanning adjacent low-value points of wave velocity anomalies along the transverse wave velocity distribution curve, spatially continuous low-value points of wave velocity anomalies are grouped into low-value wave velocity anomaly clusters. When the number of low-value wave velocity anomalies in the same low-value wave velocity anomaly cluster reaches the preset minimum number of consecutive points, the spatial range covered by the low-value wave velocity anomaly cluster is determined as a suspected cavity response zone. The minimum number of consecutive points is 3. A single low-value wave velocity anomaly may be caused by measurement error. Only 3 or more consecutive low-value points can reliably indicate an anomaly of a certain scale in space. The implementer can adjust the number of consecutive points within the range of 2 to 5 according to the detection spacing and the minimum cavity size of the target. Extract the horizontal boundary coordinates and vertical depth interval of the suspected cavity response zone.

[0024] Time difference analysis was performed on the reflected wave groups in the transient vibration response waveforms corresponding to the suspected cavity response zone. The method for identifying reflected wave groups was as follows: in the waveforms after the first arrival time, signal segments with amplitudes exceeding 30% of the first arrival wave amplitude and whose time difference with the first arrival time was greater than the minimum two-way travel time corresponding to the stratum thickness were identified as reflected wave groups originating from the underground interface. The depth positions of the cavity roof and floor interfaces were determined based on the time difference of the reflected interfaces. Specifically, the depth value of the reflected interface was obtained by dividing the difference between the arrival time and the first arrival time of the reflected wave group by 2 and then multiplying it by the average wave velocity of the corresponding path. The three-dimensional boundary contours of each suspected cavity were delineated by combining the horizontal boundary coordinates. All the delineated three-dimensional boundary contours were assembled according to the longitudinal mileage coordinates of the roadbed to generate a spatial distribution map of the hidden goaf area. The map recorded the mileage coordinate range, roof depth, floor depth, and lateral width of each hidden goaf area.

[0025] Step S4: Based on the spatial distribution map of the hidden goaf and the coupled evolution characteristic sequence, construct the instability critical state criterion set, substitute the real-time collected multi-physics field original response signal into the instability critical state criterion set to perform time-by-time approximation degree calculation, and generate goaf instability approximation degree time series data. After obtaining the precise spatial morphological parameters of the hidden goaf, it is necessary to establish a set of instability critical state criteria that combines spatial geometric parameters with dynamic monitoring signals in order to achieve a quantitative and dynamic assessment of the instability risk of the goaf.

[0026] Specifically, the roof thickness, cavity span, and cavity depth of each hidden goaf are extracted from the spatial distribution map of the goaf. The structural vulnerability level parameter of the hidden goaf is determined based on the ratio of the roof thickness to the cavity span. When the ratio is less than 0.5, it is classified as highly vulnerable, indicating that the roof is relatively weak and prone to through-hole failure under load. When the ratio is between 0.5 and 1.0, it is classified as moderately vulnerable, indicating that the roof has a certain bearing capacity but is at risk of progressive degradation under sustained load. When the ratio is greater than 1.0, it is classified as low vulnerable, indicating that the roof is relatively thick and has a low risk of instability in the short term. The above ratio range is based on the empirical criteria for evaluating the stability of surrounding rock in underground engineering, and implementers can modify it according to the measured results of local geotechnical mechanical parameters.

[0027] For each type of structural vulnerability level parameter corresponding to a hidden goaf, historical coupled response events that have occurred in the coupled evolution characteristic sequence of the corresponding segment are traced back. The increasing trend of the interval duration and intensity of these historical coupled response events is statistically analyzed. The method for calculating the shortening rate of the interval duration is as follows: extract the occurrence times of the most recent M coupled response events, calculate the time interval sequence between adjacent event points, perform linear fitting on the time interval sequence, and the absolute value of the slope of the fitted line is the interval duration shortening rate. M is set to 10, and the implementer can adjust it within the range of 5 to 20 according to the historical event density. When the number of historical coupled response events is less than M, the actual number of available event points is used for fitting, and the criterion entries for this segment are marked as pending calibration. As monitoring data continues to accumulate, the critical values ​​in the criterion entries are automatically updated. Once the number of historical event points reaches M, they are converted into formal criterion entries to participate in subsequent approximation degree calculations. The slope of the intensity increasing trend is calculated as follows: extract the event intensity value sequence corresponding to the above M coupled response event points, perform linear fitting on the event intensity value sequence, and the slope of the fitted line is the slope of the intensity increasing trend. The rate of shortening of the occurrence interval and the slope of the intensity increasing trend are used as indicators of the dynamic degradation rate of the vulnerability level of this type of structure.

[0028] Structural vulnerability level parameters and dynamic degradation rate indicators are combined to form criteria entries. Each criterion entry includes the structural vulnerability level, the corresponding critical value for coupled response event density, and the corresponding critical value for degradation rate. Specifically, the critical value for coupled response event density for high vulnerability level is 8 coupled response event points per 24 hours, and the critical value for degradation rate is an interval shortening rate exceeding 0.1 hours per event point. The critical values ​​for medium vulnerability level are 12 event points per 24 hours and a shortening rate exceeding 0.05 hours per event point. The critical values ​​for low vulnerability level are 18 event points per 24 hours and a shortening rate exceeding 0.03 hours per event point. Highly vulnerable goaf areas may become unstable after a relatively small accumulation of coupled response events; therefore, a lower trigger threshold is set. Implementers can calibrate and correct each critical value based on historical monitoring data. All criterion entries are compiled to form a set of instability critical state criteria.

[0029] The real-time acquired multi-physics field raw response signals are processed using the same cross-physics field temporal correlation analysis method as the coupled evolution feature sequence to obtain real-time coupled response event points. The density value of real-time coupled response event points and the real-time degradation rate value within the most recent preset time period are statistically analyzed. The preset time period is the most recent 24 hours.

[0030] Retrieve criteria entries from the instability critical state criterion set that match the structural vulnerability level parameters of the current segment, and read the critical values ​​of coupling response event density and degradation rate from these criterion entries. Use the ratio of the real-time coupling response event point density value to the coupling response event density critical value as the density approximation component, and the ratio of the real-time degradation rate value to the degradation rate critical value as the rate approximation component. When the calculated value of either the density or rate approximation component exceeds 1.0, it is truncated to 1.0 before participating in subsequent comparison calculations, ensuring that the approximation value remains within the range of 0 to 1.0. Take the larger value between the density and rate approximation components as the segment approximation for the current time period. The reason for taking the maximum value instead of the average is that either component reaching the critical value first indicates that the instability risk has entered the corresponding level. Taking the maximum value ensures that the most sensitive risk signals are responded to first, avoiding the underreporting of high-risk states due to the low average value of the two components. The segment proximity degrees of continuous time periods are arranged in chronological order to generate time series data on the proximity degree of goaf instability. The time resolution of the time series data is the same as the acquisition period, which is 60 seconds, to support real-time comparison of subsequent graded response thresholds.

[0031] Step S5: When the proximity value of any segment in the time series data of the instability proximity of the goaf reaches the preset graded response threshold, an emergency linkage early warning instruction of the corresponding level is generated according to the threshold level reached and sent to the highway traffic control facility terminal. When the proximity value of any segment in the time series data of the instability proximity of the goaf reaches the preset graded response threshold, an emergency linkage early warning instruction of the corresponding level is generated according to the threshold level reached and sent to the highway traffic control facility terminal, so as to realize the closed-loop linkage response from risk perception to traffic intervention and make up for the lack of sufficient early warning time of the traditional monitoring system.

[0032] The tiered response thresholds include, in ascending order, a concern level threshold, a warning level threshold, and an emergency level threshold. The concern level threshold is set at a segment proximity value of 0.5, the warning level threshold at 0.7, and the emergency level threshold at 0.9. These threshold values ​​are based on the following criteria: in the instability critical state criterion set, the segment proximity value corresponding to the critical value is 1.0. A segment proximity value of 0.5 indicates that the current monitoring indicator has reached half of the critical value, indicating a state requiring attention but with a sufficient window for intervention; a segment proximity value of 0.7 indicates that the risk is accumulating rapidly, requiring proactive intervention; and a segment proximity value of 0.9 indicates that the instability critical state is extremely close, requiring immediate emergency response. These threshold values ​​aim to control the false alarm rate while ensuring adequate advance warning. Implementers can adjust these values ​​within a reasonable range based on the trade-off between the costs of false alarms and missed alarms in specific projects.

[0033] When the proximity value of a section reaches the attention level threshold of 0.5 for the first time, an attention level warning instruction is generated. The attention level warning instruction is sent to the road section information board display system and the road section inspection and dispatch terminal. The road section information board display system is instructed to publish speed limit information on the variable message sign in front of the corresponding progressive mechanical degradation section, reducing the speed limit from the normal speed limit by 20 kilometers per hour. At the same time, the road section inspection and dispatch terminal is instructed to arrange on-site verification tasks, dispatching inspection personnel to arrive at the site within 2 hours to conduct visual inspection and portable ground-penetrating radar verification. The risk status of the hidden mining subsidence area is further confirmed through manual verification to prevent false alarms from causing unnecessary traffic interference.

[0034] When the proximity value of a section continues to rise and reaches the warning threshold of 0.7, a warning-level alert is generated. In addition to the targets of the attention-level alert, the warning-level alert is also sent to the toll station entrance control equipment and traffic guidance screens on adjacent road sections. The toll station entrance control equipment is instructed to implement flow control for vehicles entering the progressively degraded mechanical zone, allowing only vehicles with an axle load below 80 kN to pass. Heavy-load vehicles are guided to detours to parallel lanes. Adjacent road section traffic guidance screens are also instructed to publish detour guidance information, informing drivers of the geological risks on the current road section and recommending alternative routes. This proactively reduces the frequency of load application above the progressively degraded mechanical zone, slows the rate of soil and rock degradation, and buys more time for emergency response.

[0035] When the segment proximity value reaches the emergency level threshold of 0.9, an emergency level warning command is generated. In addition to the targets of the alert level warning command, the emergency level warning command is also sent to the emergency rescue command center. An emergency rescue dispatch request is sent to the emergency rescue command center. The emergency rescue dispatch request includes the precise mileage coordinates of the progressively degraded mechanical section, the spatial geometric parameters of the hidden goaf, and a real-time status summary of the current segment proximity value and coupled evolution characteristic sequence. This allows the emergency rescue command center to allocate engineering rescue forces and professional geological disaster response teams. At the same time, a full closure traffic control measure is triggered for the progressively degraded mechanical section. The toll station entrance control equipment prevents all vehicles from entering the section, and the information board system sends emergency evacuation guidance information to vehicles en route within the progressively degraded mechanical section.

[0036] This embodiment realizes a complete technical closed loop from continuous acquisition of multi-physics field signals, cross-field collaborative anomaly identification, fine detection of passive sources, quantitative risk assessment to hierarchical linkage early warning response. It effectively solves the safety and prevention problems caused by the strong concealment, difficulty in monitoring coverage, and short early warning time of hidden mining areas on highways, and provides a systematic technical solution for the urgent need of modern intelligent transportation for real-time risk prediction.

[0037] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways, characterized in that: include: Step S1: Continuously collect multi-physics field original response signals of deep soil and rock masses in the roadbed. The multi-physics field original response signals include micro-vibration waveform signals, ground resistivity change signals and pore water pressure fluctuation signals. Step S2: Perform cross-physical field time-series correlation analysis on the original response signals of the multi-physics field, extract the coupling evolution feature sequence between signals of different physical fields, identify the progressive mechanical degradation sections in the deep soil and rock mass of the roadbed based on the coupling evolution feature sequence, and generate hidden anomaly section marking information. Step S3: Initiate a directional detection and scanning task based on the marking information of the hidden anomaly section. Obtain deep structural echo data of the anomaly section through the coordinated work of road traffic load excitation and stratum response sensing array. Perform cavity geometry inversion calculation on the deep structural echo data to generate a spatial distribution map of the hidden goaf. Step S4: Based on the spatial distribution map of the hidden goaf and the coupled evolution characteristic sequence, construct the instability critical state criterion set, substitute the real-time collected multi-physics field original response signal into the instability critical state criterion set to perform time-by-time approximation degree calculation, and generate goaf instability approximation degree time series data. Step S5: When the proximity value of any segment in the time series data of the instability proximity of the goaf reaches the preset graded response threshold, an emergency linkage early warning instruction of the corresponding level is generated according to the threshold level reached and sent to the highway traffic control facility terminal.

2. The method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways according to claim 1, characterized in that, Methods for obtaining the original response signal of a multiphysics field include: A ground response sensing array is constructed by driving ground anchor-type sensing probe groups into the toe of the outer slope of the highway shoulder in an equally spaced manner. Each ground anchor-type sensing probe group includes a micro-vibration pickup element, a ring electrode pair and a micro-pressure sensing unit fixed at different depths of the probe rod. Adjacent surface anchored sensing probe groups are connected to the roadside aggregation node via signal transmission cables shallowly buried along the surface of the road shoulder slope. The roadside aggregation node timestamps and calibrates the original electrical signals received from each surface anchored sensing probe group to generate a synchronously collected data stream with a unified time reference. The roadside convergence node separates the synchronously acquired data stream into micro-vibration waveform signal channels, ground resistivity change signal channels, and pore water pressure fluctuation signal channels according to the preset physical field classification labels. The data of each channel is arranged and packaged according to the spatial number of the surface anchored sensing probe group to generate multi-physical field original response signals.

3. The method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways according to claim 2, characterized in that, The methods for generating hidden anomaly segment marker information include: The micro-vibration waveform signal is segmented according to the depth of the vertical sensing probe rod group, and the waveform energy attenuation rate of each depth layer in the continuous acquisition cycle is extracted. At the same time, the resistivity change signal is extracted according to the same depth layer to extract the resistivity time-series change gradient value, and the pore water pressure fluctuation signal is extracted according to the same depth layer to extract the pressure oscillation frequency and amplitude envelope. For the same depth layer, the waveform energy attenuation rate, resistivity temporal change gradient value and pressure oscillation frequency are compared cycle by cycle along the time axis. When the three show a coordinated change trend in the same acquisition cycle, the acquisition cycle of that depth layer is marked as a coupling response event point. The coupling response event points that appear in the order of time within the preset statistical time window are connected in series to form a coupling evolution feature sequence. The coupled evolution feature sequence of each vertical sensing probe rod group is scanned segment by segment along the longitudinal direction of the roadbed. The continuous segment where the cumulative event density of the coupled evolution feature sequence exceeds the preset density threshold is identified as a progressive mechanical degradation segment. The starting and ending station numbers and depth range of the progressive mechanical degradation segment are recorded to generate hidden abnormal segment marking information.

4. The method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways according to claim 3, characterized in that, The methods for obtaining coupled response event points include: The direction of change of the waveform energy attenuation rate between adjacent acquisition cycles is determined. The direction of change is determined as one of three states: positive increase, negative decrease, or stable. Simultaneously, the direction of change of the resistivity time-series gradient value and the pressure oscillation frequency between adjacent acquisition cycles are determined. The three-dimensional consistency determination is performed on the change direction indicators of waveform energy decay rate, resistivity time-series change gradient value, and pressure oscillation frequency. When all three change direction indicators point to positive increase or negative decrease within the same acquisition period, it is determined to be a strong cooperative change trend. When two of the three change direction indicators point to the same direction and the third remains stable, it is determined to be a secondary cooperative change trend. The acquisition period determined to be a strong cooperative change trend or a secondary cooperative change trend is marked as a coupling response event point. For each coupled response event point, the waveform energy decay rate, resistivity time-series change gradient value, and pressure oscillation frequency values ​​are read during the acquisition period. The values ​​of the three physical field signals are weighted, mapped, and summed to obtain the event intensity value of the coupled response event point. The event intensity values ​​of each coupled response event point are arranged in chronological order to form an event intensity value sequence corresponding to the coupled evolution characteristic sequence.

5. The method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways according to claim 4, characterized in that, Methods for acquiring deep structure echo data include: The starting and ending chainages of the progressively degraded mechanical zone are read from the hidden anomaly zone marker information. The vertical sensing probe groups within the range corresponding to the starting and ending chainages are switched to high-frequency acquisition mode. During the road traffic period in the progressively degraded mechanical zone, the wheel load of the actual heavy-duty vehicle is used as the natural excitation source. When the road pressure sensing device detects that the wheel axle of the heavy-duty vehicle passes through the progressively degraded mechanical zone, the vertical sensing probe groups in high-frequency acquisition mode are simultaneously triggered to record the transient vibration response waveform of the stratum caused by the wheel load impact. The transient vibration response waveforms recorded by multiple vertical sensing probe groups under the same wheel load excitation event are arranged according to spatial position and reception time difference to form a passive source detection record set with the wheel load application point as the virtual seismic source and the vibration pickup element of each probe group as the receiving point. The passive source detection record set is the deep structure echo data.

6. The method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways according to claim 5, characterized in that, The cavity geometry of the deep structure echo data is inverted and solved to generate a spatial distribution map of the hidden goaf, including: The arrival time and amplitude characteristics of each transient vibration response waveform are extracted from the deep structure echo data. The direct wave propagation velocity distribution is calculated according to the geometric relationship between the virtual source position and the receiver position. The initial velocity structure profile of the progressive mechanical degradation section is established based on the direct wave propagation velocity distribution. In the initial velocity structure profile, the regions with abnormally low wave velocity values ​​are scanned layer by layer. Regions with wave velocity values ​​lower than the average wave velocity of adjacent layers by a preset proportion are marked as suspected cavity response regions. The horizontal range boundary coordinates and vertical depth range of the suspected cavity response regions are extracted. Time difference analysis is performed on the reflected wave groups in the transient vibration response waveform corresponding to the suspected cavity response zone. The depth position of the cavity top plate interface and bottom plate interface is determined according to the time difference of the reflection interface. The three-dimensional boundary contour of each suspected cavity is delineated by combining the horizontal range boundary coordinates. All the delineated three-dimensional boundary contours are assembled according to the longitudinal mileage coordinates of the roadbed to generate a spatial distribution map of the hidden goaf area.

7. The method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways according to claim 6, characterized in that, The marking methods for suspected cavity response areas include: The transverse wave velocity distribution curves of each depth layer are extracted layer by layer along the depth direction of the initial velocity structure profile. The wave velocity values ​​of each transverse wave velocity distribution curve are read point by point according to the spatial location. Simultaneously, the wave velocity values ​​of the adjacent layers above and below the same depth layer at the same spatial location are read. The arithmetic mean of the wave velocity values ​​of the adjacent layers above and below is taken as the reference wave velocity value of the adjacent layer at that spatial location. The actual wave velocity value at this spatial location is compared with the reference wave velocity value of the adjacent layer to obtain the wave velocity deviation. When the wave velocity deviation is negative and its absolute value exceeds the deviation threshold obtained by multiplying the reference wave velocity value of the adjacent layer by a preset ratio, the spatial location is marked as a point with abnormally low wave velocity. Scan adjacent low-value points of wave velocity anomalies along the transverse wave velocity distribution curve, and group the spatially continuous low-value points of wave velocity anomalies into a low-value cluster of wave velocity anomalies. When the number of low-value points of wave velocity anomalies contained in the same low-value cluster of wave velocity anomalies reaches the preset minimum number of continuous points, the spatial range covered by the low-value cluster of wave velocity anomalies is determined as a suspected cavity response zone.

8. The method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways according to claim 7, characterized in that, A set of criteria for determining instability critical states is constructed based on the spatial distribution map of concealed goaf areas and coupled evolutionary feature sequences, including: The roof thickness, cavity span, and cavity depth of each hidden goaf are extracted from the spatial distribution map of the hidden goaf. The structural vulnerability level parameter of the hidden goaf is determined according to the ratio of the roof thickness to the cavity span. The structural vulnerability level parameter is divided into three categories: high vulnerability, medium vulnerability, and low vulnerability. For the hidden goaf corresponding to each type of structural vulnerability level parameter, the historical coupling response event points that have occurred in the coupling evolution characteristic sequence of the segment are traced back. The occurrence interval and intensity increasing trend of the historical coupling response event points are statistically analyzed. The rate of shortening of the occurrence interval and the slope of the intensity increasing trend are used as the dynamic degradation rate index of this type of structural vulnerability level. The structural vulnerability level parameter and the dynamic degradation rate index are combined to form a criterion entry. Each criterion entry includes the structural vulnerability level, the corresponding critical value of the coupled response event density, and the corresponding critical value of the degradation rate. All criterion entries are compiled to form an unstable critical state criterion set.

9. The method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways according to claim 8, characterized in that, The real-time acquired multi-physics field raw response signals are substituted into the instability critical state criterion set for time-by-time approximation degree calculation to generate time-series data of instability approximation degree of the goaf, including: The real-time acquired multi-physics field raw response signals are processed using the same cross-physics field temporal correlation analysis method as the coupled evolution feature sequence to obtain real-time coupled response event points. The density value of real-time coupled response event points and the real-time degradation rate value within the most recent preset time period are statistically analyzed. Retrieve criteria entries from the instability critical state criterion set that match the structural vulnerability level parameters of the current section, and read the critical values ​​of coupling response event density and degradation rate from the criterion entries; use the ratio of real-time coupling response event point density value to coupling response event density critical value as the density approximation component, and the ratio of real-time degradation rate value to degradation rate critical value as the rate approximation component, and take the larger value between the density approximation component and the rate approximation component as the section approximation for the current time period, and arrange the section approximation for continuous time periods in chronological order to generate time series data of instability approximation for goaf.

10. The method for comprehensive perception, prediction, and emergency response early warning of hidden mining subsidence areas along highways according to claim 9, characterized in that, Based on the threshold level reached, corresponding emergency response warning instructions are generated and sent to the highway traffic control facility terminals, including: The graded response thresholds include the attention level threshold, the alert level threshold, and the emergency level threshold, which are set sequentially from low to high. When the comprehensive proximity value first reaches the attention level threshold, an attention level warning instruction is generated. When the comprehensive proximity value continues to rise and reaches the alert level threshold, an alert level warning instruction is generated. When the comprehensive proximity value reaches the emergency level threshold, an emergency level warning instruction is generated. The alert level warning is issued to the road information board display system and the road inspection and dispatch terminal, instructing the road information board display system to issue speed limit reminder information, and at the same time instructing the road inspection and dispatch terminal to arrange on-site verification tasks; The alert-level warning instruction, in addition to the targets of the attention-level warning instruction, is also issued to the toll station entrance control equipment and the traffic guidance screens of adjacent road sections. It instructs the toll station entrance control equipment to implement flow restriction for vehicles entering the progressive mechanical degradation zone, and instructs the traffic guidance screens of adjacent road sections to publish detour guidance information. The emergency-level warning instruction, in addition to the target group of the alert-level warning instruction, is also sent to the emergency rescue command center, sending an emergency rescue dispatch request to the emergency rescue command center.

Images