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

Abstract

The application discloses a distributed optical fiber sensing system for water conservancy project structure health monitoring, and relates to the technical field of water conservancy project safety monitoring.The application provides the following scheme, which comprises a signal acquisition and preprocessing module, is used for dividing the distributed optical fiber into segments according to a fixed length, acquires vibration signals of center points of each optical fiber segment at a fixed period, and generates a sampling point sequence with uniform time intervals; a feature extraction and criterion construction module is connected with the signal acquisition and preprocessing module and is used for each sampling point in the sampling point sequence.The distributed optical fiber sensing system for water conservancy project structure health monitoring provided by the application solves the problems of false signal flooding and transmission resource waste caused by multi-dimensional criterion isolation and lack of spatial correlation in the prior art through the construction of a double-level criterion based on the correlation of frequency domain features and spatial topology, the cooperative optimization of a dynamic hot point classification mechanism based on signal propagation characteristics and a hierarchical transmission strategy.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydraulic engineering safety monitoring, more particularly, the present application relates to a distributed optical fiber sensing system for hydraulic engineering structure health monitoring. BACKGROUND

[0002] In the field of hydraulic engineering structure health monitoring, distributed optical fiber sensing systems have been widely used in the monitoring of vibration, strain and other signals of structures such as dams, embankments and water pipelines due to their advantages of long distance, distribution, high sensitivity, etc. In the prior art, such systems usually divide the optical fiber sensing section by fixed length, collect vibration signals at a constant sampling period, and determine whether the structure has abnormalities by preset threshold values. At the same time, in order to ensure data integrity, the system adopts a unified transmission strategy to upload all the original signals or simply compressed feature data of the sampling points to the terminal processing unit without distinction.

[0003] Although the existing system introduces multi-parameter threshold values, the criteria of each dimension are isolated, without considering the physical correlation between parameters, and lacking real-time verification of the signal space propagation characteristics and the intersection of frequency domain features, resulting in excessive uploading of isolated or false high-heat signals occupying transmission bandwidth, hindering the real-time transmission of key abnormal signals, and ultimately causing delay in the early warning of the health monitoring system. Therefore, the present application provides a distributed optical fiber sensing system for hydraulic engineering structure health monitoring to solve this problem. SUMMARY

[0004] To solve the above technical problems, the present application provides a distributed optical fiber sensing system for hydraulic engineering structure health monitoring, which solves the problems raised in the background art.

[0005] To achieve the above purpose, the technical solution of the present application is as follows:

[0006] The present application provides a distributed optical fiber sensing system for hydraulic engineering structure health monitoring, which comprises:

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

[0008] A feature extraction and criterion construction module is connected to the signal acquisition and preprocessing module, and 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 in which the sampling point is located, and construct a double-level threshold criterion;

[0009] The space analysis and hot point division module is connected with the feature extraction and criterion construction module, is used for executing neighborhood expansion and region merging on the sampling points meeting the double-level threshold criterion based on the space topology neighborhood of the optical fiber segment center points to determine candidate regions, screening non-candidate region sampling points in combination with the frequency band energy threshold and the space adjacency, and dividing the sampling points into different hot point categories;

[0010] The transmission strategy decision module is connected with the space analysis and hot point division module, is used for aggregating the sampling points in a fixed period according to the hot point categories, generating different types of transmission segments, and distributing transmission strategies according to the transmission segment types;

[0011] The space analysis and hot point division module specifically comprises:

[0012] The seed point generation and topology expansion unit is used for setting the space topology neighborhood of each optical fiber segment center point based on the distribution path of the distributed optical fiber, recording the sampling points meeting the double-level threshold criterion as seed points, and performing three-order bidirectional expansion on the seed points along the topology neighborhood to generate hot point candidate regions by merging regions with a continuous seed point number greater than a dynamic threshold N;

[0013] The abnormal interference filtering unit is used for identifying and extracting interference coupling regions and device noise regions in the hot point candidate regions to generate an abnormal mask, and removing the seed points in the hot point candidate regions by using the abnormal mask to obtain primary hot points;

[0014] The secondary hot point capturing unit is used for screening non-seed points based on the frequency band energy size, and marking the non-seed points as secondary hot points when the non-seed points are spatially adjacent to the primary hot points;

[0015] The aggregation of the sampling points in a fixed period according to the hot point categories can refer to the following flow:

[0016] In the same fixed period, the first sampling point marked as a primary or secondary hot point is taken as the starting point of a segment, and subsequent hot points of the same level are included in the segment, when the hot point level is switched, the current segment is ended and a new segment is started, forming a plurality of primary hot segments and secondary hot segments;

[0017] The average amplitude of each hot segment is calculated, and when the average amplitude is greater than M times the amplitude threshold, the hot segment is marked as a virtual primary hot segment;

[0018] The transmission strategy distribution according to the transmission segment type can refer to the following flow:

[0019] The original optical signals of the optical fiber segment center points corresponding to the time stamps of all sampling points in each level hot segment are obtained, and are recorded as original signals;

[0020] The primary hot segment is taken as the first priority to transmit the original signal in real time;

[0021] transmitting the original signal at the end of the segment as a second priority of a virtual primary heat segment;

[0022] transmitting the compressed features of the original signal at the end of the cycle as a third priority of a secondary heat segment;

[0023] Counting the number of all sampling points in the statistical cycle, generating an abnormal probability score value and uploading periodically.

[0024] Compared with the prior art, the application has the following beneficial effects:

[0025] The distributed optical fiber sensing system for structural health monitoring of water conservancy projects provided by the application effectively solves the problems of false signal flooding and waste of transmission resources caused by the isolation of multi-dimensional criteria and the lack of spatial correlation in the prior art through the construction of a double-level criterion associated with the frequency domain features and the spatial topology, the dynamic heat point classification mechanism based on the signal propagation characteristics, and the cooperative optimization of the hierarchical transmission strategy. BRIEF DESCRIPTION OF DRAWINGS

[0026] The disclosure of the application will be described 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 application. Among them:

[0027] Fig. 1 The flowchart for performing the steps of the system in the application;

[0028] Fig. 2 The structural block diagram of the distributed optical fiber sensing system for structural health monitoring of water conservancy projects proposed by the application. DETAILED DESCRIPTION

[0029] It is easy to understand that according to the technical scheme of the application, those skilled in the art can propose a plurality of structure modes and implementation modes which can be replaced with each other without changing the essential spirit of the application. Therefore, the following specific embodiments and drawings are only exemplary descriptions of the technical scheme of the application, and should not be regarded as the whole or as a limitation or restriction on the technical scheme of the application.

[0030] In the prior art, the health and safety monitoring system of water conservancy projects requires a hazard perception and early warning capability with ultra-low delay (millisecond level) to support rapid emergency decision-making. Distributed optical fiber sensing is the core monitoring technology, which provides the advantage of continuous spatial perception, but the amount of original optical signal data is extremely large, often reaching GB level per second. The existing mainstream architecture relies on transmitting all original DFS data to a remote cloud / central server for processing. GB / s level data flow results in high network transmission delay (often 1 second), which becomes a bottleneck for real-time response, especially in remote water conservancy sites where network bandwidth transmission is insufficient.

[0031] To solve the above problems, the inventors found that the prior art uses amplitude or band energy as a criterion in isolation, lacks multi-dimensional feature fusion analysis, by studying the propagation law of vibration signals in the optical fiber space, it is found that the abnormal signal has the characteristics of frequency band energy mutation and neighborhood space correlation; further analysis shows that dynamically adjusting the transmission strategy can be processed according to the importance of the signal, thereby optimizing the resource allocation; based on this, it is proposed to combine the frequency domain features with the spatial topology to construct the criterion, and to design a hierarchical transmission mechanism.

[0032] Referring to Figs. 1-2 The distributed optical fiber sensing system for water conservancy engineering structure health monitoring comprises:

[0033] A signal acquisition and preprocessing module is configured to divide the distributed optical fiber into segments according to a fixed length, to acquire vibration signals of the center points of each optical fiber segment at a fixed period, and to generate a sequence of sampling points with uniform time intervals.

[0034] For example, a distributed optical fiber sensor with a total length of 500 m is laid out, the optical fiber is divided into segments according to a fixed length of 10 m, a total of 50 segments are obtained, the fixed period can be 5 s, 1000 points are acquired for each center point in each period, the sampling interval is 5 ms, and the sequence of sampling points is .

[0035] A feature extraction and criterion construction module is connected to the signal acquisition and preprocessing module, configured to extract the amplitude of the vibration signal of each sampling point in the sequence of sampling points, calculate the band energy of the sliding window in which the sampling point is located, and construct a double-level threshold criterion.

[0036] For example, a sliding window composed of every 10 sampling points (such as to , t is the current sampling point) can be used to calculate the frequency band energy in real time, and the window length dynamically adapts to the water conservancy scene, the window length is 5-15 sampling points in the flood season (> 100 Hz), and the window length is 20-50 sampling points in the dry season (< 10 Hz), to ensure that it can be implemented and avoid being too wide.

[0037] A spatial analysis and hot point division module is connected to the feature extraction and criterion construction module, configured to perform neighborhood expansion and region merging on the sampling points that meet the double-level threshold criterion based on the spatial topology neighborhood of the center points of the optical fiber segments to determine the candidate area, filter the non-candidate area sampling points by combining the band energy threshold and spatial adjacency, and divide the sampling points into different hot point categories.

[0038] A transmission strategy decision module is connected to the spatial analysis and hot point division module, configured to aggregate the sampling points in a fixed period according to the hot point categories, generate transmission segments of different types, and allocate transmission strategies according to the types of the transmission segments.

[0039] In an embodiment of the present application, the spatial analysis and hot spot division module specifically comprises:

[0040] A seed point generation and topology expansion unit is configured to set a spatial topology neighborhood of each fiber segment center point based on a distributed fiber layout path, record a sampling point meeting a double-level threshold criterion as a seed point, and perform a three-order bidirectional expansion on the topology neighborhood, and merge a region with a continuous seed point number greater than a dynamic threshold N to generate a hot spot candidate area; wherein N is determined based on a fiber segment length and a water conservancy structure material characteristic, and satisfies , for example, 5 for concrete and 3 for earth and rockfill dams;

[0041] It should be noted that the three-order bidirectional expansion along the topology neighborhood is a process of finding and including all nodes reachable through at most three fiber links from a selected node along the actual connection path of the fiber link on a fiber connection structure diagram; for example, for a single fiber, the neighborhood of fiber node n is ; and the merging of a region with a continuous seed point number greater than 5 is that all seed points in the neighborhood of fiber node n meet the double-level threshold criterion,

[0042] An abnormal interference filtering unit is configured to identify and extract an interference coupling area and a device noise area in the hot spot candidate area to generate an abnormal mask, and use the abnormal mask to remove seed points in the hot spot candidate area to obtain primary hot spots.

[0043] A secondary hot spot capturing unit is configured to filter non-seed points based on a frequency band energy size, and mark the non-seed points as secondary hot spots when the non-seed points are spatially adjacent to the primary hot spots.

[0044] Specifically, based on the spatial topology neighborhood of the fiber segment center point, first, seed points meeting amplitude and frequency band energy double criteria are filtered, continuous seed points are merged to form a candidate area through three-order bidirectional expansion, then, a time-space gradient and a dispersion are calculated in the candidate area, a power frequency harmonic component in a vibration signal spectrum is extracted, an interference coupling area and a device noise area are identified to generate an abnormal mask, and after removing misjudgment seed points, primary hot spots are obtained, for sampling points not marked as seed points but with a significantly over-standard frequency band energy, if the sampling points are spatially adjacent to the primary hot spots, the sampling points are captured as secondary hot spots, and the process verifies and multi-level filtering mechanism through spatial correlation of signal propagation, effectively distinguishing between real structure damage signals and environmental interference signals.

[0045] In an embodiment of the present application, the identification and extraction of the interference coupling area and the device noise area in the hot spot candidate area refer to the following flow:

[0046] The time-space gradient of each sub-point amplitude in the heat point candidate area is calculated, and the seed point with a dispersion greater than a preset dispersion threshold is selected into the interference coupling area;

[0047] It should be noted that the time-space gradient reflects the mutation degree of the signal at adjacent points in space and adjacent points in time, and the preset dispersion threshold determines the critical value of environmental interference and real anomaly. The dispersion threshold needs to satisfy , and the fiber segment length , such as 2 for a concrete structure;

[0048] The 50Hz power frequency harmonic component is extracted from the vibration signal spectrum of each sub-point. If the proportion of the 50Hz power frequency harmonic component is greater than 40%, the seed point is included in the equipment noise area;

[0049] The interference coupling area and the equipment noise area are merged to generate an anomaly mask;

[0050] It should be noted that the spatial gradient is the amplitude difference between the seed point and the neighborhood points, and the time gradient is the amplitude difference between the seed point and the same point in the previous period. The arctangent value of the spatial gradient of the seed point and each neighborhood point is called the neighborhood spatial gradient angle. The average value of all neighborhood spatial gradient angles is the spatial gradient angle, and the standard deviation of the neighborhood spatial gradient angle is the dispersion. When the time gradient is greater than 0, is taken as the time gradient angle, and when the time gradient is less than 0, is taken as the time gradient angle; the time-space gradient is the weighted sum of the time gradient angle and the spatial gradient angle. For example, the weight ratio of the time gradient angle and the spatial gradient angle is The spatial gradient is used to identify the propagation direction of the current vibration damage, and the time gradient is used as an auxiliary to identify sudden vibration interference or continuous vibration damage.

[0051] Further, the dispersion threshold is obtained by cross verification of the vibration wave coherence limit of different materials and the statistical inflection point of the interference sample. Its mathematical basis is times of the standard deviation of the uniform distribution of the multi-neighborhood direction, indicating a significant truncation. For example, the concrete dam is , the earth and rockfill dam is , and the time-space gradient direction value greater than the dispersion threshold indicates that there is no dominant vector in the topological neighborhood, which does not meet the wave front consistency feature of structure damage propagation.

[0052] Specifically, after generating the heat point candidate area, firstly, the amplitude variation of the seed point in the candidate area is analyzed by time-space gradient, and the seed point with abnormal dispersion is selected as the interference coupling area. At the same time, the vibration signal of the seed point in the candidate area is decomposed by frequency spectrum, the proportion of power frequency harmonic component is detected, and the seed point meeting the noise characteristics of the power equipment is classified into the equipment noise area. By merging the two types of interference areas to form an abnormal mask, the pseudo abnormal signal points generated by environmental coupling interference or equipment running noise can be accurately removed, and the abnormal area reflecting the true structure health state is retained.

[0053] Through the above technical scheme, the application can accurately distinguish the real abnormal vibration of the water conservancy project structure from the environmental coupling interference and the equipment running noise, avoid misjudging the power equipment vibration as the structure damage signal, and reduce the invalid data transmission amount.

[0054] In an embodiment of the application, the determination of the double-level threshold criterion is as follows:

[0055] The feature curve of the spectrum of the sampling point vibration signal is extracted, and the frequency band energy is obtained by integrating the feature curve in the frequency band range of the target frequency band.

[0056] The amplitude threshold and the frequency band energy threshold are set, and 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, it is determined that the sampling point meets the double-level threshold criterion.

[0057] The amplitude threshold setting step is: taking a dam as an example, based on the continuous 30-day vibration signal time series data under the health state, firstly, the 99th percentile and the maximum value of the amplitude distribution are calculated, and the smaller value of the two is taken as 70% of the basic threshold, then the threshold is corrected according to the dam type: the concrete dam is increased by 10%, the earth-rock dam is decreased by 15%, and the temperature dynamic compensation is introduced in real-time monitoring (when the temperature rises by 5%, when the temperature drops by 5%, increase ) and sensor drift calibration (drift amount reverse compensation when the range is

[0058] The frequency band energy threshold setting step is: under the no-load working condition of the dam (such as at night), the energy mean value and the standard deviation of each target frequency band are extracted, and the frequency band energy threshold is set as The principle sets a basic threshold value, applies industry specification coefficients for different frequency band characteristics, such as an increase of 20% on the seepage sensitive frequency band (3-10 Hz), maintenance of the benchmark on the structure vibration frequency band (50-150 Hz), and a decrease of 20% on the high-frequency cavitation frequency band (>300 Hz), and online stage correction through water level linkage (energy threshold value of the frequency band is reduced to 85% when the water level is higher than the design water level, and the value is increased by the difference proportion when the water level is normal), and periodic resetting to ensure adaptability;

[0059] It should be noted that the amplitude value represents the instantaneous energy intensity of the signal, directly reflecting the vibration / strain energy received by the optical fiber sensor, and its physical limitation is that it cannot distinguish the signal type and is easily disturbed by environmental noise. The frequency band energy represents the frequency distribution characteristics of the signal, and the energy concentration degree reflects the physical nature of the event, such as high frequency band (>100 Hz) which may represent crack propagation or brittle fracture, and low frequency band (<10 Hz) which may represent dam settlement or foundation slip. The core auxiliary role of frequency band energy in amplitude value determination is to provide event type identification, noise stripping and physical mechanism interpretation capability, and to supplement key context information for amplitude value determination;

[0060] Compared with the prior art, the traditional method usually only uses a single amplitude threshold value or isolated frequency band energy as a criterion, without considering the physical correlation of the vibration signal in amplitude-frequency characteristics. For example, some equipment operation noise may have high amplitude but dispersed frequency band energy characteristics, while the real structure damage signal often shows high energy aggregation in a specific frequency band. The present scheme can accurately distinguish these signals through the synergistic effect of the double-level criterion, reducing the amount of invalid data transmission.

[0061] In an embodiment of the present application, the frequency band range of the target frequency band is determined as follows:

[0062] Obtain the vibration signal time sequence of several historical periods of each sampling point, and associate the environmental parameters;

[0063] Extract the historical amplitude subsequence from the time sequence, and cluster it into different working condition barrels according to the environmental parameters;

[0064] For example, collect the vibration signal time sequence data of the past 30 days, associate the water level, temperature and other environmental parameters, and cluster them into 5 working condition barrels such as high water level-normal temperature;

[0065] Calculate the short-time Fourier transform for each working condition barrel data, and generate the frequency band weight based on the energy distribution entropy value;

[0066] Extract the frequency spectrum of each vibration signal in the sampling point field matrix, calculate the spatial coherence coefficient, and fuse the frequency band weight to obtain the frequency band score;

[0067] Merge adjacent frequency spectrum points with a difference in frequency band score within a preset tolerance range to generate the frequency band range of the target frequency band.

[0068] Compared with the prior art, the scheme realizes dynamic frequency band division by fusing historical data clustering, frequency band energy entropy analysis and spatial coherence calculation; for example, when the dam encounters different water level working conditions, the system can automatically adjust the target frequency band range to avoid the problem of characteristic frequency band deviation caused by environmental parameter changes.

[0069] Through the above technical scheme, the application can accurately identify the characteristic frequency band corresponding to the structural anomaly, and reduce the misjudgment probability caused by environmental noise; in the water conservancy engineering monitoring scene, the scheme can effectively distinguish water flow impact vibration and structural damage vibration; for example, in water pipeline monitoring, through dynamic frequency band selection, the fixed frequency interference generated by pump operation can be filtered, and the detection sensitivity of crack propagation characteristic frequency is improved.

[0070] In an embodiment of the present application, the determination of the secondary heat point is as follows:

[0071] Screening non-seed points with an amplitude less than the amplitude threshold and a frequency band energy greater than K times the frequency band energy threshold; wherein K is dynamically adjusted based on the type of water conservancy structure and the level of environmental noise, and satisfies ;

[0072] Calculate the Euclidean distance of each non-seed point to the nearest primary heat point;

[0073] When the Euclidean distance is not more than the fixed length, mark the non-seed point as a secondary heat point.

[0074] Through the above technical scheme, the application effectively captures the secondary heat points within the spatial propagation range of the primary heat candidate area while retaining the integrity of the primary heat candidate area. The determination mechanism can avoid the redundant data transmission caused by misjudgment of isolated noise in traditional methods, and at the same time prevent the spatial correlation signals of real damage events from being missed.

[0075] In an embodiment of the present application, the sampling points in a fixed period are aggregated according to the heat point category as follows:

[0076] In the same fixed period, the first sampling point marked as a primary or secondary heat point is taken as the starting point of the segment, and the subsequent heat points of the same level are included in the segment, and the time stamp corresponding to each sampling point is recorded synchronously, when the heat point level is switched, the current segment is ended and a new segment is started, forming a plurality of primary heat segments and secondary heat segments;

[0077] Calculate the average amplitude of each heat segment, when the average amplitude is greater than M times the amplitude threshold, mark the heat segment as a virtual primary heat segment; wherein M is dynamically based on the type of water conservancy structure and signal stability, and satisfies , usually The critical signal is identified through a secondary classification mechanism, is separated from the secondary hotness points based on amplitude statistics, is promoted to the second priority to transmit the original signal, and key details are avoided from being lost.

[0078] In one embodiment of the application, transmission strategies are allocated according to transmission segment types, as shown in the following flow:

[0079] The original optical signal of the fiber segmentation center point corresponding to the time stamp of all sampling points in each level of hotness segment is obtained, and is denoted as an original signal.

[0080] The primary hotness segment is taken as the first priority to transmit the original signal in real time.

[0081] The virtual primary hotness segment is taken as the second priority to transmit the original signal at the end of the segment.

[0082] The secondary hotness segment is taken as the third priority to transmit the compressed features of the original signal at the end of the period; the compressed features refer to a low-dimensional representation set extracted from the original optical signal time sequence, which can replace the transmission of the original signal, and the low-dimensional representation includes a Rayleigh scattering intensity spectrum, a phase difference matrix, and a photon arrival rate statistical histogram.

[0083] The number of all sampling points in the period is counted, an abnormal probability score value is generated, and is uploaded regularly.

[0084] Compared with the prior art, the traditional system adopts a unified transmission strategy for all data, cannot distinguish the levels of abnormal signals, and lacks a dynamic transmission resource scheduling mechanism; the present application realizes gradient allocation of transmission resources through three-level priority division, preferentially guarantees the real-time performance of key abnormal signals, reduces the bandwidth occupation of secondary data by using compressed transmission, and establishes a multi-period correlation evaluation framework through an abnormal probability scoring mechanism, thereby providing data support for long-term monitoring of the structural health state.

[0085] In one embodiment of the application, the transmission strategy decision module comprises a streaming segment processing mechanism: when the first sampling point of the primary hotness segment is included, transmission is immediately started; if the current transmission resource is insufficient, the transmission of the second priority is suspended in sequence until sufficient resources are obtained, the transmission of the primary hotness segment adopts a block packaging method, a fixed number of sampling points are used to form a data packet, and an emergency identifier is added.

[0086] Through the above technical solution, the application can preferentially guarantee the real-time transmission demand of the primary hotness segment under limited bandwidth conditions, avoid the delay or loss of key abnormal signals caused by network resource competition, reduce the amount of single packet data in a block packaging manner, improve the transmission success rate, ensure that the receiving end timely identifies and processes high-priority data through the emergency identifier, and thereby shorten the overall response time from abnormal detection to early warning triggering.

[0087] In one embodiment of the present application, the anomaly probability score is stored and periodically uploaded, see the following flow:

[0088] The number of primary hot points and the number of secondary hot points in the period are counted, and the ratio of the weighted sum of the two to the total number of sampling points is calculated, denoted as the anomaly probability score value;

[0089] When the anomaly probability score value is greater than the preset score value in a continuous preset number of periods, the abnormal data upload is triggered at the end of the last period; the preset score value can be 5%;

[0090] Specifically, at the end of a fixed period, the system obtains the number of primary hot points and secondary hot points through the built-in statistical unit, multiplies the two by the preset weight coefficient respectively, and then divides the total number of sampling points in the period to obtain the anomaly probability score value; the score value is stored in the non-volatile memory, and the number of continuous over-limit periods is recorded by the period counter; when the number of continuous monitoring periods reaches the preset threshold and the last score is still higher than the preset threshold, the system automatically triggers the abnormal data compression module to package the original signal and characteristic parameters of the related optical fiber segment into a data packet for uploading.

[0091] The above formulas are dimensionless values, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the latest real situation, and the preset parameters in the formula are set by the person skilled in the art according to the actual situation.

[0092] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized in the form of a computer program product wholly or partially.

[0093] Those skilled in the art should understand that the modules and algorithm steps of each example described in connection with the embodiments disclosed herein can be realized by electronic hardware, computer software or a combination of the two. Whether the functions are specifically implemented in hardware or software depends on the specific application scenarios and design constraints of the technical solutions. In addition, those skilled in the art can use different methods to realize the described functions for each specific application scenario, and such implementation shall be considered within the scope of the present application.

[0094] In addition, each functional module in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0095] The technical scope of the present application is not limited to the above description, and those skilled in the art can make various modifications and changes to the above embodiments without departing from the technical idea of the present application, and these modifications and changes should all belong to the protection scope of the present application.

Claims

1. A distributed optical fiber sensing system for structural health monitoring of hydraulic structures, characterized in that, The system comprises: a signal acquisition and preprocessing module, configured to divide the distributed optical fiber segments according to a fixed length, to acquire vibration signals of the center points of the optical fiber segments at a fixed period, and to generate a sequence of sampling points with uniform time intervals; a feature extraction and criterion construction module, connected to the signal acquisition and preprocessing module, configured to extract the amplitude of the vibration signal of each sampling point in the sequence of sampling points, to calculate the frequency band energy of the sliding window in which the sampling point is located, and to construct a double-level threshold criterion; a spatial analysis and hot point division module, connected to the feature extraction and criterion construction module, configured to perform neighborhood expansion and region merging on the sampling points meeting the double-level threshold criterion based on the spatial topological neighborhood of the center points of the optical fiber segments to determine candidate regions, to screen non-candidate region sampling points based on the frequency band energy threshold and spatial adjacency, and to divide the sampling points into different hot point categories; a transmission strategy decision module, connected to the spatial analysis and hot point division module, configured to aggregate the sampling points in a fixed period according to the hot point categories, to generate transmission segments of different types, and to assign transmission strategies according to the transmission segment types; the spatial analysis and hot point division module specifically comprises: a seed point generation and topological expansion unit, configured to set the spatial topological neighborhood of the center points of the optical fiber segments based on the layout path of the distributed optical fiber, to mark the sampling points meeting the double-level threshold criterion as seed points, and to perform three-order bidirectional expansion on the seed points along the topological neighborhood to merge regions with a number of continuous seed points greater than a dynamic threshold N to generate hot point candidate regions; an abnormal interference filtering unit, configured to identify and extract interference coupling regions and equipment noise regions in the hot point candidate regions to generate an abnormal mask, and to remove the seed points in the hot point candidate regions using the abnormal mask to obtain primary hot points; a secondary hot point capturing unit, configured to screen non-seed points based on the size of the frequency band energy, and to mark the non-seed points as secondary hot points when the non-seed points are spatially adjacent to the primary hot points; the aggregation of the sampling points in a fixed period according to the hot point categories can be seen from the following process: in the same fixed period, the first sampling point marked as a primary or secondary hot point is taken as the starting point of a segment, and subsequent hot points of the same level are included in the segment, when the level of the hot point changes, the current segment is ended and a new segment is started, forming a plurality of primary and secondary hot segments; the average amplitude of each hot segment is calculated, and when the average amplitude is greater than M times the amplitude threshold, the hot segment is marked as a virtual primary hot segment; the assignment of transmission strategies according to the transmission segment types can be seen from the following process: the original optical signals of the center points of the optical fiber segments corresponding to the time stamps of all sampling points in each hot segment are obtained and are marked as original signals; the primary hot segments are taken as the first priority to transmit the original signals in real time; the virtual primary hot segments are taken as the second priority to transmit the original signals at the end of the segment; the secondary hot segments are taken as the third priority to transmit the compressed features of the original signals at the end of the period; the number of all sampling points in the period is counted to generate an abnormal probability score value and to upload the value periodically.

2. The system of claim 1, wherein, the identification and extraction of interference coupling regions and equipment noise regions in the hot point candidate regions can be seen from the following process: The time-space gradient of each sub-point amplitude in the heat point candidate area is calculated, and seed points with a dispersion greater than a preset dispersion threshold are selected into the interference coupling area; The 50Hz power frequency harmonic component is extracted from the vibration signal spectrum of each sub-point, and if the proportion of the 50Hz power frequency harmonic component is greater 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 abnormality mask.

3. The system of claim 1, wherein, The sampling points meeting the double-level threshold criterion are recorded as seed points, as follows: The feature curve of the sampling point vibration signal spectrum is extracted, and the frequency band energy is obtained by integrating the feature curve in the target frequency band range; The amplitude threshold and the frequency band energy threshold are set, and 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, it is determined that the sampling point meets the double-level threshold criterion.

4. The system of claim 3, wherein, The frequency band range of the target frequency band is determined, as follows: The vibration signal time sequence of each sampling point in several historical periods is obtained, and the environmental parameters are associated; The historical amplitude sub-sequence is extracted from the time sequence, and is clustered into different working condition barrels according to the environmental parameters; The short-time Fourier transform is calculated for each working condition barrel data, and the frequency band weight is generated based on the energy distribution entropy value; The spectrum of each vibration signal in the sampling point field matrix is extracted, the spatial coherence coefficient is calculated, and the frequency band score is obtained by fusing the frequency band weight; The adjacent frequency spectrum points with a frequency band score difference within a preset tolerance range are merged to generate the frequency band range of the target frequency band.

5. The system of claim 3, wherein, The determination of the secondary heat point is as follows: Non-seed points with an amplitude less than the amplitude threshold and a frequency band energy greater than K times the frequency band energy threshold are screened; The Euclidean distance of each non-seed point to the nearest primary heat point is calculated; When the Euclidean distance is not more than the several fixed lengths, the non-seed point is marked as a secondary heat point.

6. The system of claim 1, wherein, The transmission strategy decision module includes a streaming segment processing mechanism: when the first sampling point of the primary heat segment is included, the transmission is immediately started, if the current transmission resource is insufficient, the second priority transmission is suspended in turn until sufficient resources are obtained, the transmission of the primary heat segment uses a block packaging method, a fixed number of sampling points are used to form a data packet, and an emergency identifier is added.

7. The system of claim 1, wherein, The generation of the abnormal probability score value and the periodic uploading are as follows: The number of primary heat points and the number of secondary heat points in the period are counted, the weighted sum of the two is calculated, and the ratio of the weighted sum to the total number of sampling points is recorded as the abnormal probability score value; When the abnormal probability score value is greater than the preset score value in a continuous preset number of periods, the abnormal data uploading is triggered at the end of the last period.

Images