Distributed optical fiber sensing system for health monitoring of hydraulic engineering structure

Abstract

The invention discloses a distributed optical fiber sensing system for water conservancy project structure health monitoring, relates to the technical field of water conservancy project safety monitoring, and provides the following scheme: the distributed optical fiber sensing system comprises a signal acquisition and preprocessing module which is used for dividing distributed optical fiber segments according to a fixed length, acquiring vibration signals of central points of the optical fiber segments in a fixed period, and sending the vibration signals to a data processing module; generating a sampling point sequence with uniform time intervals; and the feature extraction and criterion construction module is connected with the signal acquisition and preprocessing module and is used for detecting each sampling point in the sampling point sequence. According to the distributed optical fiber sensing system for water conservancy project structure health monitoring provided by the invention, through double-level criterion construction associated with frequency domain characteristics and spatial topology, a dynamic heat point classification mechanism based on signal propagation characteristics and collaborative optimization of a hierarchical transmission strategy; the problems of false signal flooding and transmission resource waste caused by multi-dimensional criterion isolation and spatial correlation deficiency in the prior art are effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of water conservancy project safety monitoring technology, and more specifically, the present application relates to a distributed optical fiber sensing system for water conservancy project structure health monitoring. Background Art

[0002] In the field of structural health monitoring of water conservancy projects, distributed fiber optic sensing systems, due to their advantages of long distance, distributed distribution, and high sensitivity, have been widely used to monitor vibration, strain, and other signals in structures such as dams, levees, and water pipelines. Existing technologies typically use fiber optic sensing segments divided into fixed lengths, collect vibration signals at a constant sampling period, and use preset thresholds to determine whether the structure is abnormal. Furthermore, to ensure data integrity, these systems often employ a unified transmission strategy, indiscriminately uploading the raw signals or simply compressed feature data from all sampling points to the terminal processing unit.

[0003] Although existing systems introduce multi-parameter thresholds, the criteria for each dimension are isolated. They neither consider the physical correlation between parameters nor verify the authenticity of the signal's spatial propagation characteristics and the intersection of frequency domain features. This leads to excessive upload of isolated or false high-heat signals, which occupy transmission bandwidth, hinder the real-time transmission of key abnormal signals, and ultimately cause early warning delays in the health monitoring system. Therefore, we propose a distributed fiber optic sensing system for structural health monitoring of water conservancy projects to solve this problem. Summary of the Invention

[0004] In order to solve the above technical problems, a distributed optical fiber sensing system for water conservancy project structure health monitoring is provided. This technical solution solves the problems raised in the above background technology.

[0005] In order to achieve the above object, the technical solution of the present invention is as follows:

[0006] The present application provides a distributed optical fiber sensing system for structural health monitoring of water conservancy projects, the system comprising:

[0007] The signal acquisition and preprocessing module is used to divide the distributed optical fiber into segments according to fixed lengths, collect the vibration signal of the center point of each optical fiber segment at a fixed period, and generate a sequence of sampling points with uniform time intervals;

[0008] The feature extraction and criterion construction module is connected to the signal acquisition and preprocessing module. It is used to extract the amplitude of the vibration signal of each sampling point in the sampling point sequence, calculate the frequency band energy of the sliding window where it is located, and construct a two-level threshold criterion;

[0009] The spatial analysis and hotspot classification module is connected to the feature extraction and criterion construction module. It is used to perform neighborhood expansion and region merging on sampling points that meet the dual-level threshold criteria based on the spatial topological neighborhood of the fiber segment center point to determine candidate areas. It combines the frequency band energy threshold and spatial proximity to filter out non-candidate area sampling points and classify the sampling points into different hotspot categories.

[0010] The transmission strategy decision module is connected to the spatial analysis and hot spot division module. It is used to aggregate the sampling points within a fixed period according to the hot spot category, generate different types of transmission segments, and allocate transmission strategies according to the transmission segment type.

[0011] Compared with the prior art, the present invention has the following beneficial effects:

[0012] The distributed fiber optic sensing system for structural health monitoring of water conservancy projects provided in this application is constructed through a two-level criterion associated with frequency domain characteristics and spatial topology, a dynamic hot point classification mechanism based on signal propagation characteristics, and collaborative optimization of hierarchical transmission strategies, which effectively solves the problems of false signal proliferation and waste of transmission resources caused by the isolation of multi-dimensional criteria and the lack of spatial correlation in the existing technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] The disclosure of the present invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Among them:

[0014] Figure 1 A flowchart of the steps executed by the system of the present invention;

[0015] Figure 2 This is a structural block diagram of the distributed optical fiber sensing system for water conservancy project structural health monitoring proposed in the present invention. DETAILED DESCRIPTION

[0016] It is easy to understand that according to the technical solution of the present invention, without changing the essential spirit of the present invention, a person skilled in the art can propose a variety of interchangeable structural modes and implementation modes. Therefore, the following specific embodiments and drawings are only exemplary descriptions of the technical solution of the present invention and should not be regarded as the entire invention or as a limitation or restriction of the technical solution of the present invention.

[0017] In existing technologies, the health and safety of water conservancy projects requires that the monitoring system has ultra-low latency (millisecond level) risk perception and early warning capabilities to support rapid emergency decision-making. Distributed fiber optic sensing is the core monitoring technology, providing the advantage of continuous spatial perception. However, its raw optical signal data volume is extremely large, often reaching GB per second. The existing mainstream architecture relies on the complete transmission of all raw DFS data to the remote cloud / central server for processing. The GB / s-level data flow leads to high network transmission latency (often 1 second), becomes a bottleneck for real-time response, especially in remote water conservancy sites where network bandwidth transmission is insufficient.

[0018] In order to solve the above problems, the inventors found that the existing technology only uses amplitude or frequency band energy as a criterion in isolation, and lacks multi-dimensional feature fusion analysis. By studying the propagation law of vibration signals in the optical fiber space, it is found that abnormal signals have the characteristics of frequency band energy mutation and neighborhood space correlation; further analysis shows that dynamic adjustment of transmission strategy can be based on the hierarchical processing of signal importance, thereby optimizing resource allocation; based on this, it is proposed to combine frequency domain features with spatial topology to construct a criterion and design a hierarchical transmission mechanism.

[0019] Reference Figure 1-2 As shown, a distributed optical fiber sensing system for water conservancy project structural health monitoring includes:

[0020] The signal acquisition and preprocessing module is used to divide the distributed optical fiber into segments according to fixed lengths, collect the vibration signal of the center point of each optical fiber segment at a fixed period, and generate a sequence of sampling points with uniform time intervals;

[0021] For example, a distributed optical fiber sensor with a total length of 500m is laid out, and the optical fiber is divided into segments with a fixed length of 10m, resulting in 50 segments in total. The fixed period can be 5s, and 1000 points are collected from each center point in each period. The sampling interval is 5ms, and the sampling point sequence is ;

[0022] The feature extraction and criterion construction module is connected to the signal acquisition and preprocessing module. It is used to extract the amplitude of the vibration signal of each sampling point in the sampling point sequence, calculate the frequency band energy of the sliding window where it is located, and construct a two-level threshold criterion;

[0023] For example, a sliding window consisting of 10 sampling points can be taken (e.g. to , t is the current sampling point) to calculate the frequency band energy in real time. The window length is dynamically adapted to the water conservancy scenario. During the flood season (>100Hz), the window length is 5-15 sampling points, and during the dry season (<10Hz), the window length is 20-50 sampling points, ensuring feasibility and avoiding an overly wide range.

[0024] The spatial analysis and hotspot classification module is connected to the feature extraction and criterion construction module. It is used to perform neighborhood expansion and region merging on sampling points that meet the dual-level threshold criteria based on the spatial topological neighborhood of the fiber segment center point to determine candidate areas. It combines the frequency band energy threshold and spatial proximity to filter out non-candidate area sampling points and classify the sampling points into different hotspot categories.

[0025] The transmission strategy decision module is connected to the spatial analysis and hot spot division module. It is used to aggregate the sampling points within a fixed period according to the hot spot category, generate different types of transmission segments, and allocate transmission strategies according to the transmission segment type.

[0026] In one embodiment of the present invention, the spatial analysis and hotspot division module specifically includes:

[0027] The seed point generation and topology expansion unit is used to set the spatial topology neighborhood of the center point of each fiber segment based on the layout path of the distributed optical fiber, record the sampling points that meet the double-level threshold criteria as seed points, and perform three-order bidirectional expansion along the topological neighborhood, merge the areas where the number of consecutive seed points is greater than the dynamic threshold N to generate hot point candidate areas; where N is determined based on the fiber segment length and the characteristics of the water conservancy structure material, and meets , for example, take 5 for concrete dams and 3 for earth-rock dams;

[0028] It should be noted that the third-order bidirectional expansion along the topological neighborhood is: starting from a selected node on the fiber connection structure diagram, along the actual connection path of the fiber link, gradually find and include all nodes that can be reached through at most three fiber links; for example, taking a single fiber as an example, the neighborhood of the fiber node n is ; Merge the area where the number of consecutive seed points is greater than 5 as follows: All seed points in the neighborhood of fiber node n meet the double-level threshold criterion,

[0029] The abnormal interference filtering unit is used to identify and extract the interference coupling area and equipment noise area in the hot spot candidate area to generate an abnormal mask, and use the abnormal mask to eliminate the seed points in the hot spot candidate area to obtain the first-level hot spot;

[0030] The secondary hot spot capture unit is used to screen non-seed points based on the frequency band energy size. When the non-seed point is adjacent to the primary hot spot space, it is marked as a secondary hot spot.

[0031] Specifically, based on the spatial topological neighborhood of the center point of the optical fiber segment, seed points that meet the dual criteria of amplitude and frequency band energy are first screened. Continuous seed points are merged through third-order bidirectional expansion to form a candidate area. Then, the spatiotemporal gradient and discreteness are calculated within the candidate area, the power frequency harmonic components in the vibration signal spectrum are extracted, and the interference coupling area and equipment noise area are identified to generate anomaly masks. After removing the misjudged seed points, the first-level hot points are obtained. For sampling points that are not marked as seed points but have significantly exceeded the frequency band energy standard, if they are spatially adjacent to the first-level hot points, they are captured as secondary hot points. This process effectively distinguishes real structural damage signals from environmental interference signals through spatial correlation verification of signal propagation and a multi-level screening mechanism.

[0032] In one embodiment of the present invention, the following process is used to identify and extract the interference coupling area and the device noise area in the hot spot candidate area:

[0033] Calculate the spatiotemporal gradient of the amplitude of various sub-points in the hot spot candidate area, and select seed points whose spatiotemporal gradient discreteness is greater than the preset discreteness threshold to be included in the interference coupling area;

[0034] It should be noted that the spatiotemporal gradient reflects the degree of signal mutation at adjacent points in space and time. The discreteness threshold is preset to determine the critical value between environmental interference and real anomalies. The discreteness threshold must meet , and the fiber segment length , such as concrete structure, take 2;

[0035] Extract the power frequency harmonic components from the vibration signal spectrum of various seed points. If the 50 Hz power frequency harmonic component accounts for more than 40%, the seed point is included in the equipment noise area.

[0036] Merge the interference coupling area and the equipment noise area to generate an anomaly mask;

[0037] It should be noted that the spatial gradient is the amplitude difference between the seed point and the neighboring points, the time gradient is the amplitude difference between the seed point and the point at the same time in the previous cycle, the arc tangent value of the spatial gradient between the seed point and each neighboring point is recorded as the neighborhood spatial gradient angle, the average of all neighborhood spatial gradient angles is taken as the spatial gradient angle, and the standard deviation of the neighborhood spatial gradient angle is taken as the dispersion; when the time gradient is greater than 0, As the time gradient angle, when the time gradient is less than 0, As the time gradient angle; the spatiotemporal gradient is the weighted sum of the time gradient angle and the space gradient angle. For example, the weight ratio of the time gradient angle to the space gradient angle is ,The spatial gradient is used to identify the current vibration damage propagation direction, and the time gradient is used as an auxiliary to identify sudden vibration interference or continuous vibration damage;

[0038] It is further explained that the discreteness threshold is obtained by cross-validation of the coherence limit of vibration waves of different materials and the statistical inflection point of interference samples. Its mathematical basis is the standard deviation of uniform distribution in multiple neighborhood directions. times, indicating a significant cutoff, such as a concrete dam , earth-rock dam is ,The value of the direction of the time-space gradient is greater than the ,discreteness threshold, indicating that there is no dominant vector in the vibration ,direction within the topological neighborhood, which does not conform to the wavefront consistency ,characteristics of structural damage propagation;

[0039] Specifically, after generating candidate hotspot areas, we first perform a spatiotemporal gradient analysis on the amplitude changes of the seed points within the candidate areas, and screen out seed points with abnormal dispersion as interference coupling areas. At the same time, we perform spectral decomposition on the vibration signals of the seed points in the candidate areas, detect the proportion of power frequency harmonic components, and classify seed points that meet the noise characteristics of power equipment into the equipment noise area. By merging the two types of interference areas to form an anomaly mask, we can accurately eliminate pseudo-anomaly signal points generated by environmental coupling interference or equipment operation noise, and retain the anomaly areas that truly reflect the health status of the structure.

[0040] Through the above technical solution, the present application can accurately distinguish between the real abnormal vibration of the water conservancy project structure and the environmental coupling interference and equipment operation noise, avoid misjudging the vibration of the power equipment as a structural damage signal, and reduce the amount of invalid data transmission.

[0041] In one embodiment of the present invention, the frequency band range of the target frequency band is determined as follows:

[0042] Obtain the vibration signal time series of several historical cycles at each sampling point and associate them with environmental parameters;

[0043] Extract historical amplitude subsequences from the time series and cluster them into different working condition buckets based on environmental parameters;

[0044] For example, the vibration signal time series data of the past 30 days is collected, associated with environmental parameters such as water level and temperature, and clustered into five working condition buckets such as high water level-normal temperature;

[0045] Calculate the short-time Fourier transform of each working condition bucket data and generate frequency band weights based on the energy distribution entropy value;

[0046] Extract the spectrum of each vibration signal in the sampling point field matrix, calculate the spatial coherence coefficient and fuse the frequency band weights to obtain the frequency band score;

[0047] Adjacent spectrum points whose band score differences are within a preset tolerance range are merged to generate the frequency band range of the target frequency band.

[0048] Compared with existing technologies, this solution achieves dynamic frequency band division by integrating historical data clustering, frequency band energy entropy analysis and spatial coherence calculation; for example, when the dam encounters different water level conditions, the system can automatically adjust the target frequency band range to avoid the characteristic frequency band offset problem caused by changes in environmental parameters.

[0049] Through the above technical solution, the present application can accurately identify the characteristic frequency bands corresponding to structural abnormalities and reduce the probability of misjudgment caused by environmental noise; in the water conservancy project monitoring scenario, the solution can effectively distinguish between water flow impact vibration and structural damage vibration. For example, when monitoring water pipelines, dynamic frequency band selection can filter out the fixed frequency interference generated by pump operation, thereby improving the detection sensitivity of the characteristic frequency of crack expansion.

[0050] In one embodiment of the present invention, the determination of compliance with the dual-level threshold criteria is made with reference to the following process:

[0051] Extract the characteristic curve of the vibration signal spectrum of the sampling point and integrate it over the frequency band range of the target frequency band to obtain the frequency band energy;

[0052] Set an amplitude threshold and a frequency band energy threshold. If the amplitude of the sampling point is greater than or equal to the amplitude threshold and the frequency band energy is greater than or equal to the frequency band energy threshold, then the sampling point is determined to meet the dual-level threshold criterion.

[0053] The steps for setting the amplitude threshold are as follows: Taking the dam as an example, based on the 30-day continuous vibration signal time series data in its healthy state, first calculate the 99% quantile and maximum value of the amplitude distribution, and take 70% of the smaller value as the basic threshold. Then, it is adjusted according to the dam type: concrete dams float up 10%, earth-rock dams float down 15%, and temperature dynamic compensation is introduced in real-time monitoring ( Hours per rise Down 5%, Every time Increase ) and sensor drift calibration (drift Reverse compensation when measuring range), and reset periodically every 72 hours (reduced to 6 hours during flood season);

[0054] The steps for setting the frequency band energy threshold are as follows: extract the energy mean and standard deviation of each target frequency band under the dam no-load condition (such as late at night), and calculate the energy threshold according to the frequency band energy threshold. In principle, basic thresholds are set, and industry standard coefficients are applied to the characteristics of different frequency bands. For example, the seepage-sensitive frequency band (3-10Hz) is increased by 20%, the structural vibration frequency band (50-150Hz) is maintained at the baseline, and the high-frequency cavitation frequency band (>300Hz) is reduced by 20%. In the online stage, water level linkage correction is implemented (the frequency band energy threshold is reduced to 85% when the design water level is exceeded, and increased by the difference ratio at normal water levels). Periodic reset is also performed to ensure adaptability.

[0055] It should be noted that amplitude represents the instantaneous energy intensity of the signal and directly reflects the vibration / strain energy received by the fiber optic sensor. Its physical limitations include its inability to distinguish signal types and susceptibility to interference from environmental noise. Band energy represents the frequency distribution characteristics of the signal, and the energy concentration reflects the physical nature of the event. For example, high-frequency bands (>100Hz) may indicate crack propagation or brittle fracture, while low-frequency bands (<10Hz) may indicate dam settlement or foundation slip. The core auxiliary role of band energy in amplitude determination is to provide event type identification, noise removal, and physical mechanism interpretation capabilities, providing key contextual information for amplitude determination.

[0056] Compared with existing technologies, traditional methods usually only use a single amplitude threshold or isolated frequency band energy as a criterion, without considering the physical correlation of the vibration signal in terms of amplitude-frequency characteristics. For example, the operating noise of some equipment may have the characteristics of high amplitude but dispersed frequency band energy, while real structural damage signals often show high energy concentration in specific frequency bands. This scheme can accurately distinguish these signals and reduce the amount of invalid data transmission through the synergistic effect of two-level criteria.

[0057] In one embodiment of the present invention, the determination of the secondary hot spots is as follows:

[0058] Screen non-seed points whose amplitude is less than the amplitude threshold and whose frequency band energy is greater than K times the frequency band energy threshold; K is dynamically adjusted based on the type of water conservancy structure and the ambient noise level, and satisfies ;

[0059] Calculate the Euclidean distance from each non-seed point to the nearest first-level hot point;

[0060] When the Euclidean distance does not exceed the fixed lengths, the non-seed point is marked as a secondary hot point.

[0061] Through the above technical solution, the present application effectively captures the secondary heat points within its spatial propagation range while retaining the integrity of the main heat candidate area. This judgment mechanism can avoid the redundant data transmission caused by the misjudgment of isolated noise in traditional methods, and prevent the spatial correlation signals of real damage events from being missed.

[0062] In one embodiment of the present invention, the sampling points within a fixed period are aggregated according to the hot point category, as shown in the following process:

[0063] Within the same fixed period, the first sampling point marked as a primary or secondary hot point is used as the starting point of the segment, and subsequent hot points of the same level are included in the segment. The timestamp corresponding to each sampling point is synchronously recorded. When the hot point level switches, the current segment ends and a new segment starts, forming several primary hot segments and secondary hot segments.

[0064] Calculate the average amplitude of each medium heat segment. When the average amplitude is greater than M times the amplitude threshold, mark the medium heat segment as a virtual first-level heat segment; where M is based on the type of water conservancy structure and the signal stability dynamics, and satisfies , usually take ; Critical signals are identified through a secondary classification mechanism, separated from secondary hot spots based on amplitude statistics, and promoted to the second priority for transmitting the original signal to avoid the loss of key details.

[0065] In one embodiment of the present invention, the transmission strategy is allocated according to the transmission segment type as shown in the following process:

[0066] Obtain the original optical signal of the fiber segment center point corresponding to the timestamps of all sampling points in each level of heat segment, and record it as the original signal;

[0067] The first-level hot segment is given the first priority and the original signal is transmitted in real time;

[0068] The virtual first-level heat segment is used as the second priority to transmit the original signal at the end of the segment;

[0069] The secondary heat fragments are used as the third priority to transmit the compressed features of the original signal at the end of the cycle; the compressed features refer to a set of low-dimensional representations extracted from the original optical signal time series that can replace the original signal transmission. The low-dimensional representations include Rayleigh scattering intensity spectrum, phase difference matrix and photon arrival rate statistical histogram.

[0070] Count all sampling points within a statistical period, generate abnormal probability scores, and upload them regularly;

[0071] Compared with existing technologies, traditional systems adopt a unified transmission strategy for all data, which can neither distinguish the levels of abnormal signals nor lack a dynamic scheduling mechanism for transmission resources. This solution realizes gradient allocation of transmission resources through three-level priority division, giving priority to ensuring the real-time performance of key abnormal signals. At the same time, it uses compressed transmission to reduce the bandwidth occupation of secondary data. The abnormal probability scoring mechanism establishes a multi-period correlation evaluation framework to provide data support for long-term monitoring of structural health status.

[0072] In one embodiment of the present invention, the transmission strategy decision module includes a streaming segment processing mechanism: when the first sampling point of a first-level hot segment is included, transmission is immediately started. If the current transmission resources are insufficient, the second-priority transmission is suspended in sequence until sufficient resources are available. The transmission of the first-level hot segment adopts a block encapsulation method, with a fixed number of sampling points forming a data packet and an emergency identifier added;

[0073] Through the above technical solution, this application can prioritize the real-time transmission needs of the first-level hot fragments under limited bandwidth conditions, avoid delays or losses of key abnormal signals due to network resource competition, reduce the amount of data in a single packet by block encapsulation, and improve the transmission success rate. The emergency identifier ensures that the receiving end can identify and process high-priority data in a timely manner, thereby shortening the overall response time from abnormality detection to warning triggering.

[0074] In one embodiment of the present invention, the abnormal probability score is stored and regularly uploaded as follows:

[0075] Count the number of primary hot spots and secondary hot spots within the period, calculate the ratio of their weighted sum to the total number of sampling points, and record it as the abnormal probability score;

[0076] When the abnormal probability score value is greater than the preset score value for a consecutive preset number of cycles, the abnormal data upload is triggered at the end of the last cycle; the preset score value can be 5%;

[0077] Specifically, at the end of a fixed period, the system obtains the number of primary heat points and secondary heat points respectively through the built-in statistical unit, multiplies the two by the preset weight coefficient, accumulates them, and then divides them by the total number of sampling points in the period to obtain the abnormal probability score value; the score value is stored in the non-volatile memory, and the number of consecutive exceeding cycles is recorded through the cycle counter; when the number of consecutive monitoring cycles reaches the preset threshold and the last score is still higher than the preset critical value, the system automatically triggers the abnormal data compression module, and encapsulates the original signal and characteristic parameters of the relevant optical fiber segment into a data packet for uploading.

[0078] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0079] The above embodiments may be implemented in whole or in part through software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0080] It should be understood by those skilled in the art that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. Whether these functions are specifically performed in hardware or software depends on the specific application scenario and design constraints of the technical solution. In addition, those skilled in the art can use different methods to implement the described functions for each specific application scenario, and such implementation should be considered within the scope of this application.

[0081] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0082] The technical scope of the present invention is not limited to the contents of the above description. Those skilled in the art can make various deformations and modifications to the above embodiments without departing from the technical idea of ​​the present invention, and these deformations and modifications should all fall within the protection scope of the present invention.

Claims

1. A distributed optical fiber sensing system for structural health monitoring of water conservancy projects, characterized in that: The system comprises: The signal acquisition and preprocessing module is used to divide the distributed optical fiber into segments according to fixed lengths, collect the vibration signal of the center point of each optical fiber segment at a fixed period, and generate a sequence of sampling points with uniform time intervals; The feature extraction and criterion construction module is connected to the signal acquisition and preprocessing module. It is used to extract the amplitude of the vibration signal of each sampling point in the sampling point sequence, calculate the frequency band energy of the sliding window where it is located, and construct a two-level threshold criterion; The spatial analysis and hotspot classification module is connected to the feature extraction and criterion construction module. It is used to perform neighborhood expansion and region merging on sampling points that meet the dual-level threshold criteria based on the spatial topological neighborhood of the fiber segment center point to determine candidate areas. It combines the frequency band energy threshold and spatial proximity to filter out non-candidate area sampling points and classify the sampling points into different hotspot categories. The transmission strategy decision module is connected to the spatial analysis and hot spot division module. It is used to aggregate the sampling points within a fixed period according to the hot spot category, generate different types of transmission segments, and allocate transmission strategies according to the transmission segment type.

2. The system according to claim 1, wherein: The spatial analysis and hotspot division module specifically includes: The seed point generation and topology expansion unit is used to set the spatial topological neighborhood of the center point of each fiber segment based on the layout path of the distributed fiber. The sampling points that meet the double-level threshold criteria are recorded as seed points, and they are expanded in three orders along the topological neighborhood. The areas where the number of consecutive seed points is greater than the dynamic threshold N are merged to generate hot point candidate areas. The abnormal interference filtering unit is used to identify and extract the interference coupling area and equipment noise area in the hot spot candidate area to generate an abnormal mask, and use the abnormal mask to eliminate the seed points in the hot spot candidate area to obtain the first-level hot spot; The secondary hot spot capture unit is used to screen non-seed points based on the frequency band energy size. When the non-seed point is adjacent to the primary hot spot space, it is marked as a secondary hot spot.

3. The system according to claim 2, characterized in that The identification and extraction of interference coupling areas and device noise areas in the hotspot candidate areas refer to the following process: Calculate the spatiotemporal gradient of the amplitude of various sub-points in the hot spot candidate area, and select seed points whose spatiotemporal gradient discreteness is greater than the preset discreteness threshold to be included in the interference coupling area; Extract the power frequency harmonic components from the vibration signal spectrum of various seed points. If the 50 Hz power frequency harmonic component accounts for more than 40%, the seed point is included in the equipment noise area. The interference coupling area and the equipment noise area are merged to generate an anomaly mask.

4. The system according to claim 1, wherein: The process of determining the frequency band range of the target frequency band is as follows: Obtain the vibration signal time series of several historical cycles at each sampling point and associate them with environmental parameters; Extract historical amplitude subsequences from the time series and cluster them into different working condition buckets based on environmental parameters; Calculate the short-time Fourier transform of each working condition bucket data and generate frequency band weights based on the energy distribution entropy value; Extract the spectrum of each vibration signal in the sampling point field matrix, calculate the spatial coherence coefficient and fuse the frequency band weights to obtain the frequency band score; Adjacent spectrum points whose band score differences are within a preset tolerance range are merged to generate the frequency band range of the target frequency band.

5. The system according to claim 2, wherein: The determination of compliance with the dual-level threshold criteria is as follows: Extract the characteristic curve of the vibration signal spectrum of the sampling point and integrate it over the frequency band range of the target frequency band to obtain the frequency band energy; An amplitude threshold and a frequency band energy threshold are set. If the amplitude of the sampling point is greater than or equal to the amplitude threshold and the frequency band energy is greater than or equal to the frequency band energy threshold, the sampling point is determined to meet the dual-level threshold criterion.

6. The system according to claim 3, wherein: The determination of the secondary hot spots is as follows: Filter non-seed points whose amplitude is less than the amplitude threshold and whose band energy is greater than K times the band energy threshold; Calculate the Euclidean distance from each non-seed point to the nearest first-level hot point; When the Euclidean distance does not exceed the fixed lengths, the non-seed point is marked as a secondary hot point.

7. The system according to claim 2, wherein: The following process is used to aggregate the sampling points within a fixed period according to the hot point category: In the same fixed period, the first sampling point marked as a primary or secondary hot point is used as the starting point of the segment, and subsequent hot points of the same level are included in the segment. When the hot point level switches, the current segment ends and a new segment starts, forming several primary hot segments and secondary hot segments. The average amplitude of each medium-heat segment is calculated. When the average amplitude is greater than M times the amplitude threshold, the medium-heat segment is marked as a virtual first-level heat segment.

8. The system according to claim 7, characterized in that The allocation of transmission strategies according to transmission segment types is described in the following process: Obtain the original optical signal of the fiber segment center point corresponding to the timestamps of all sampling points in each level of heat segment, and record it as the original signal; The first-level hot segment is given the first priority and the original signal is transmitted in real time; The virtual first-level heat segment is used as the second priority to transmit the original signal at the end of the segment; The compressed features of the original signal are transmitted at the end of the period with the secondary heat fragment as the third priority; Count all sampling points within the statistical period, generate abnormal probability scores and upload them regularly.

9. The system according to claim 7, wherein: The transmission strategy decision module includes a streaming segment processing mechanism: when the first sampling point of the first-level hot segment is included, the transmission is started immediately. If the current transmission resources are insufficient, the second-priority transmission is suspended in sequence until sufficient resources are obtained. The transmission of the first-level hot segment adopts a block encapsulation method, with a fixed number of sampling points forming a data packet and an emergency identifier added.

10. The system according to claim 8, wherein: The abnormal probability score is stored and uploaded regularly, see the following process: Count the number of primary hot spots and secondary hot spots within the period, calculate the ratio of their weighted sum to the total number of sampling points, and record it as the abnormal probability score; When the abnormal probability score value is greater than the preset score value for a consecutive preset number of cycles, the abnormal data upload is triggered at the end of the last cycle.

Images