Distributed sensor based method and system for subsea fiber optic communication and vibration monitoring

Abstract

This invention relates to the fields of optical fiber communication and sensing technology, and discloses a method and system for submarine optical fiber communication and vibration monitoring based on distributed sensors. The method includes: simultaneously conducting forward communication via submarine optical fiber as a distributed optical fiber sensing medium for reverse vibration monitoring based on optical time-domain reflectometry (OTDR); acquiring real-time sensing data through multiple distributed sensor nodes set along the optical fiber; acquiring real-time vibration signals during reverse vibration monitoring; matching the real-time sensing data with historical data; generating confidence scores for the real-time sensing data based on the matching results; constructing and solving a vibration monitoring model based on sparse dictionary matching of vibration events based on the real-time vibration signals; obtaining the event type corresponding to the real-time vibration signals based on the solution results; and using confidence scores to improve the event matching and search efficiency during vibration monitoring. This invention can perform vibration monitoring while ensuring communication, improving the accuracy, positioning precision, and efficiency of vibration monitoring.

Description

Technical Field

[0001] This invention relates to the fields of optical fiber communication and sensing technology, and in particular to a method and system for submarine optical fiber communication and vibration monitoring based on distributed sensors. Background Technology

[0002] Submarine fiber optic communication networks, as a key component of global information infrastructure, not only carry the vast majority of international data transmission tasks, but also possess enormous application potential in marine environmental monitoring, resource exploration, and geological disaster early warning due to their wide geographical distribution and sensitivity to environmental disturbances. Among these, technologies utilizing optical fibers themselves as sensing media, particularly distributed vibration sensing (DVS) or distributed acoustic sensing (DAS) based on optical time domain reflectance (OTDR) technology, are research hotspots.

[0003] In submarine fiber optic vibration monitoring, OTDR-based technology is commonly used. This technology detects and locates vibration events along the fiber by analyzing the time-domain or frequency-domain changes of Rayleigh, Brillouin, or Raman scattered light in the fiber. However, this technology has some problems in practical applications, mainly including:

[0004] (1) Low event recognition accuracy. The seabed environment is complex. Ocean currents, tides, marine biological activities, seabed geological activities and the slight disturbances of the optical cable itself may generate signals similar to vibration events. OTDR has difficulty effectively distinguishing real threat events from environmental noise, resulting in a high false alarm rate and low event recognition accuracy.

[0005] (2) Limited event identification and classification capabilities. Simple OTDR signals mainly reflect the presence, intensity, and approximate location of vibrations, but have limited ability to accurately identify and classify the types of vibration events (such as ship navigation, submarine pipeline leaks, illegal intrusions, earthquakes, etc.). Different events may produce similar vibration characteristics, making it difficult to effectively distinguish them.

[0006] (3) Low positioning accuracy. The positioning accuracy of OTDR is affected by various factors such as pulse width and sampling rate. Improving positioning accuracy often requires sacrificing a certain detection distance or increasing system complexity, resulting in limited spatial information and low positioning accuracy.

[0007] (4) Low data processing and analysis efficiency. Long-distance submarine optical fibers will generate massive amounts of monitoring data, but existing methods cannot meet the real-time requirements for processing massive amounts of data, mining and identifying information, resulting in low monitoring efficiency. Summary of the Invention

[0008] Therefore, the technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a method and system for submarine optical fiber communication and vibration monitoring based on distributed sensors, which can perform vibration monitoring while ensuring communication, thereby improving the accuracy, positioning precision and efficiency of vibration monitoring.

[0009] To address the aforementioned technical problems, this invention provides a method for submarine fiber optic communication and vibration monitoring based on distributed sensors, comprising:

[0010] Using submarine optical fibers to carry forward communication signals, the submarine optical fibers are used as distributed optical fiber sensing media for reverse vibration monitoring based on optical time domain reflection technology.

[0011] Multiple distributed sensor nodes are set along the length of the submarine optical fiber. The distributed sensor nodes acquire real-time sensing data and acquire the corresponding real-time vibration signal of the submarine optical fiber at the location of the distributed sensor node during the reverse vibration monitoring process.

[0012] The real-time sensing data is matched with historical data to assess the degree of matching or the probability of occurrence of event types in the real-time sensing data and the historical data. The historical data includes sensing data of distributed sensor nodes under different known environmental conditions or known event types. The confidence level of the real-time sensing data is generated based on the matching results.

[0013] A vibration monitoring model based on sparse dictionary matching of vibration events is constructed and solved according to the real-time vibration signal. The event type corresponding to the real-time vibration signal is obtained according to the solution result of the vibration monitoring model. The confidence level is used to improve the event matching and search efficiency during vibration monitoring in the process of solving the vibration monitoring model.

[0014] Furthermore, the distributed sensor node includes a temperature sensing unit and an acceleration sensing unit. The temperature sensing unit collects real-time temperature data of the surrounding environment, and the acceleration sensing unit collects real-time vibration acceleration data in the X-axis, Y-axis, and Z-axis directions.

[0015] The real-time sensing data is fused with the real-time temperature data and the real-time vibration acceleration data, specifically as follows:

[0016] ,

[0017] in, Let be the real-time sensing data of the j-th distributed sensor node at time t. This indicates the location of the j-th distributed sensor node. Let be the average real-time temperature data of the j-th distributed sensor node at time t within the time window. Let be the variance of the real-time temperature data of the j-th distributed sensor node at time t within the time window. Let be the root mean square value of the real-time vibration acceleration data of the j-th distributed sensor node along the X-axis at time t within the time window. Let be the root mean square value of the real-time vibration acceleration data of the j-th distributed sensor node along the Y-axis at time t within the time window. Let be the root mean square value of the real-time vibration acceleration data of the j-th distributed sensor node along the Z-axis at time t within the time window. Main frequency, , For Fourier transform, This indicates taking the frequency point with the highest energy in the spectrum. Let be the real-time vibration acceleration data of the j-th distributed sensor node at time t. For energy, , Representing vectors Length, The time window length is T, where T represents the transpose operation.

[0018] Furthermore, the real-time sensing data is matched with historical data, specifically as follows:

[0019] When the kth known historical event type occurs, observe The probability is denoted as ;

[0020] The confidence level of the real-time sensing data generated based on the matching results is:

[0021] ,

[0022] in, This represents the k-th known historical event type. The confidence level of real-time sensor data under the probabilistic model. Indicates at time t, Real-time sensing data indication of the location of distributed sensor nodes The probability of occurrence for The prior probability is M, where M is the number of known historical event types.

[0023] Furthermore, the real-time sensing data is matched with historical data, specifically as follows:

[0024] Let the k-th known historical event type be ? ,use The corresponding historical data is used to train the SVM model, and the normal vector parameters of the hyperplane of the SVM model after training are obtained as follows: Bias term is ;

[0025] Calculate real-time sensor data and The decision function is:

[0026] ,

[0027] in, For real-time sensing data and Decision function, representation and The matching results It is a Gaussian kernel function;

[0028] The confidence level of the real-time sensing data is generated based on the matching results.

[0029] Furthermore, the confidence level for generating the real-time sensing data based on the matching results is:

[0030] ,

[0031] in, To match the confidence level of real-time sensor data under the model, Indicates at time t, Real-time sensing data indication of the location of distributed sensor nodes The likelihood of it happening, , These are preset parameters.

[0032] Furthermore, a vibration monitoring model based on sparse dictionary matching of vibration events is constructed and solved based on the real-time vibration signal. The event type corresponding to the real-time vibration signal is obtained based on the solution result of the vibration monitoring model, specifically:

[0033] The real-time vibration signal at time t along the length l of the coastal optical fiber is divided to obtain a vibration signal segment, denoted as . , N is the dimension of the vibration signal segment. This indicates the location of the j-th distributed sensor node corresponding to the real-time vibration signal;

[0034] Construct a sparse dictionary of vibration events, denoted as D. Each column of atoms in D represents the representation of the vibration signal corresponding to a certain event type on the OTDR signal. K is the number of atoms in the sparse dictionary of vibrational events;

[0035] The optimization objective for constructing the vibration monitoring model is:

[0036] ,or

[0037] ;

[0038] in, Let be the sparse coefficient vector to be solved. It is the L0 paradigm. It is an L2 paradigm. It is the reconstruction error tolerance. It is a regularization parameter. It is in L1 normal form;

[0039] Solving the optimization objective yields the optimal sparse coefficient vector, denoted as . ,according to The atoms of the sparse dictionary of events corresponding to non-zero elements are used to obtain the corresponding event types.

[0040] Furthermore, using the aforementioned confidence level to improve event matching and search efficiency during vibration monitoring includes:

[0041] Set a confidence threshold and perform sparse dictionary matching only on real-time vibration signals with confidence levels exceeding the confidence threshold;

[0042] When performing sparse dictionary matching, the real-time vibration signals are prioritized based on their confidence level, with high-confidence real-time vibration signals being processed first.

[0043] If the current real-time vibration signal confidence indicates one or more specific event types, then during sparse dictionary matching, a search is prioritized or performed only on a subset of dictionaries related to these specific event types.

[0044] Furthermore, using the confidence level to improve event matching and search efficiency during vibration monitoring also includes introducing the confidence level as a weight when solving the optimization objective. Specifically, the optimization objective is solved using the orthogonal matching pursuit method, and when selecting an atom in the m-th iteration, the current residual is denoted as... The selected atom index is:

[0045] ,

[0046] in, The atom index to be selected when selecting atoms in the m-th iteration. For the i-th column of the sparse dictionary of vibrational events, for The weight, and Corresponding event type associations This represents the vector dot product operation.

[0047] Furthermore, The calculation method is as follows:

[0048] ,

[0049] in, Indicates and Corresponding or similar event types, For time t, The real-time sensing data corresponding to the distributed sensor nodes at the location and the historical data The degree of matching or the probability of occurrence.

[0050] This invention also provides a submarine fiber optic communication and vibration monitoring system based on distributed sensors, including submarine fiber optic cables, multiple distributed sensor nodes, and a processing unit.

[0051] One or both ends of the submarine optical fiber are equipped with a forward communication device and an OTDR device. The forward communication device carries forward communication signals through the submarine optical fiber. With the cooperation of the OTDR device, the submarine optical fiber serves as a distributed optical fiber sensing medium for reverse vibration monitoring based on optical time domain reflection technology. During the reverse vibration monitoring process, the corresponding real-time vibration signal of the submarine optical fiber at the location of the distributed sensor node is acquired.

[0052] The distributed sensor nodes are set along the length of the submarine optical fiber to acquire real-time sensing data.

[0053] The processing unit matches the real-time sensing data with historical data to assess the degree of matching or the probability of occurrence of event types in the real-time sensing data and the historical data. The historical data includes typical sensing data characteristics of distributed sensor nodes under different known environmental conditions or known event types. Based on the matching results, the processing unit generates the confidence score of the real-time sensing data. Based on the real-time vibration signal, a vibration monitoring model based on sparse dictionary matching of vibration events is constructed and solved. The process of solving the vibration monitoring model uses the confidence score to improve the event matching and search efficiency during vibration monitoring. Based on the solution result of the vibration monitoring model, the event type corresponding to the real-time vibration signal is obtained.

[0054] Compared with the prior art, the above-described technical solution of the present invention has the following advantages:

[0055] This invention utilizes submarine optical fibers as a distributed optical fiber sensing medium for reverse vibration monitoring based on optical time-domain reflectometry, achieving vibration monitoring while ensuring communication. By acquiring local real-time sensing data through distributed sensor nodes and matching it with historical data to generate confidence scores, environmental noise interference can be effectively filtered out, improving the accuracy of identifying specific event types. The historical data includes sensing data under different known environmental conditions or known event types, enhancing event classification capabilities. Multiple distributed sensor nodes provide richer spatial information, improving event location accuracy. Confidence scores are used in solving the vibration monitoring model to improve event matching and search efficiency during vibration monitoring. Attached Figure Description

[0056] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein:

[0057] Figure 1 This is a flowchart of a method in a preferred embodiment of the present invention.

[0058] Figure 2 This is a schematic diagram of the system structure in a preferred embodiment of the present invention.

[0059] Explanation of reference numerals in the accompanying drawings: 1. Submarine fiber optic cable; 2. Forward communication equipment; 3. OTDR device; 4. Distributed sensor node; 5. Processing unit; 6. Connector box. Detailed Implementation

[0060] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0061] Reference Figure 1 As shown, this invention discloses a method for submarine fiber optic communication and vibration monitoring based on distributed sensors, comprising the following steps:

[0062] S1: Using submarine optical fiber to carry forward communication signals is the basic function of submarine optical fiber, which is realized by forward communication equipment.

[0063] S2: Using submarine optical fiber as a distributed optical fiber sensing medium for reverse vibration monitoring based on optical time domain reflectometry (OTDR) technology; in this embodiment, an OTDR device is used to inject probe light pulses into the submarine optical fiber, and the backscattered Rayleigh light caused by external disturbances such as vibration and strain in the submarine optical fiber is received to obtain the vibration signal of the submarine optical fiber. The OTDR device used can be a coherent optical time domain reflectometer or a phase-sensitive optical time domain reflectometer (φ-OTDR).

[0064] S3: Multiple distributed sensor nodes are set along the length of the seabed optical fiber. The distributed sensor nodes acquire real-time sensing data and acquire the corresponding real-time vibration signal of the seabed optical fiber at the location of the distributed sensor nodes during the reverse vibration monitoring process.

[0065] In this embodiment, the distributed sensor nodes can be encapsulated in a pressure-resistant and waterproof housing and fixed near the outside of the seabed optical fiber by anchoring or other methods. Multiple deployment points are set up every L_node kilometers along the length of the seabed optical fiber. The value of L_node is adjusted according to the actual situation. At each deployment point, at least one distributed sensor node is dispersed around the seabed optical fiber in the transverse direction to obtain richer spatial information.

[0066] The distributed sensor node includes a temperature sensing unit and an acceleration sensing unit. The temperature sensing unit (which can be a high-precision thermistor or PT100) collects real-time temperature data of its environment, and the real-time temperature data of the j-th distributed sensor node at time t is denoted as... , This represents the location of the j-th distributed sensor node. The accelerometer unit (which can be a triaxial MEMS accelerometer) collects real-time vibration acceleration data in the X, Y, and Z axes. The real-time vibration acceleration data of the j-th distributed sensor node at time t is denoted as... , , Let be the real-time vibration acceleration data of the j-th distributed sensor node along the X-axis at time t. Let be the real-time vibration acceleration data of the j-th distributed sensor node along the Y-axis at time t. This represents the real-time vibration acceleration data of the j-th distributed sensor node along the Z-axis at time t.

[0067] Real-time sensor data is fused with real-time temperature data and real-time vibration acceleration data, specifically as follows:

[0068] ,

[0069] in, Let be the real-time sensing data of the j-th distributed sensor node at time t. This indicates the location of the j-th distributed sensor node. Let be the average real-time temperature data of the j-th distributed sensor node at time t within the time window. Let be the variance of the real-time temperature data of the j-th distributed sensor node at time t within the time window. Let be the root mean square value of the real-time vibration acceleration data of the j-th distributed sensor node along the X-axis at time t within the time window. Let be the root mean square value of the real-time vibration acceleration data of the j-th distributed sensor node along the Y-axis at time t within the time window. Let be the root mean square value of the real-time vibration acceleration data of the j-th distributed sensor node along the Z-axis at time t within the time window. Main frequency, , For Fourier transform, This indicates taking the frequency point with the highest energy in the spectrum. Let be the real-time vibration acceleration data of the j-th distributed sensor node at time t. For energy, , Representing vectors Length, The time window length is T, where T represents the transpose operation.

[0070] S4: Match real-time sensing data with historical data to assess the degree of matching or probability of occurrence of event types in real-time sensing data and historical data. Historical data includes typical sensing data characteristics of distributed sensor nodes under different known environmental conditions or known event types (such as ship navigation, minor seabed earthquakes, pipeline leak simulation, normal background noise, etc.). Generate confidence of real-time sensing data based on the matching results.

[0071] In this embodiment, historical data can be represented as a probability model, specifically:

[0072] When the kth known historical event type occurs, observe The probability is denoted as ;

[0073] The confidence level of the real-time sensing data generated based on the matching results is:

[0074] ,

[0075] in, This represents the k-th known historical event type. The confidence level of real-time sensor data under the probabilistic model. Indicates at time t, Real-time sensing data indication of the location of distributed sensor nodes The probability of occurrence for The prior probability is M, where M is the number of known historical event types.

[0076] In this embodiment, historical data can also be represented as a matching model, specifically:

[0077] Let the k-th known historical event type be ? ,use The corresponding historical data is used to train the SVM model, and the normal vector parameters of the hyperplane of the SVM model after training are obtained as follows: Bias term is ;

[0078] Calculate real-time sensor data and The decision function is:

[0079] ,

[0080] in, For real-time sensing data and Decision function, representation and The matching results It is a Gaussian kernel function.

[0081] The confidence level for generating real-time sensor data based on the matching results is:

[0082] ,

[0083] in, To match the confidence level of real-time sensor data under the model, Indicates at time t, Real-time sensing data indication of the location of distributed sensor nodes The likelihood of it happening, , These are preset parameters. , The value is determined through methods such as cross-validation.

[0084] S5: Construct and solve a vibration monitoring model based on sparse dictionary matching of vibration events according to real-time vibration signals. Obtain the event type corresponding to the real-time vibration signal based on the solution results of the vibration monitoring model. Use confidence level to improve the event matching and search efficiency during vibration monitoring in the process of solving the vibration monitoring model.

[0085] S5-1: Divide the real-time vibration signal at time t along the length l of the coastal optical fiber to obtain the vibration signal segment, denoted as... , N is the dimension of the vibration signal segment. This indicates the location of the j-th distributed sensor node corresponding to the real-time vibration signal.

[0086] S5-2: Construct a sparse dictionary of vibrational events, denoted as D, where each column of atoms in D is denoted as... This represents the form of vibration signal corresponding to a certain event type on the OTDR signal (such as the vibration spectrum of a specific ship, seismic waveform, etc.). K is the number of atoms in the sparse dictionary of vibrational events.

[0087] For vibration signal segment The goal of sparse representation is to find a sparse coefficient vector. , , making and Therefore, the optimization objective for constructing the vibration monitoring model is:

[0088] ,or

[0089] ;

[0090] in, Let be the sparse coefficient vector to be solved. It is the L0 paradigm. It is an L2 paradigm. It is the reconstruction error tolerance. It is a regularization parameter. It follows the L1 paradigm. In this embodiment, the method for solving the optimization objective can be the Orthogonal Matching Pursuit (OMP) method, the Basis Pursuit (BP) method, or other similar methods.

[0091] S5-3: Solving the optimization objective yields the optimal sparse coefficient vector, denoted as... ,according to The atoms in the sparse dictionary of events corresponding to non-zero elements are used to obtain the corresponding event types. Furthermore, based on the location of the distributed sensor nodes at the real-time vibration signal and the real-time sensing data, the location and intensity of the corresponding event types can also be obtained, thereby outputting alarms or records.

[0092] Improving event matching and search efficiency in vibration monitoring using confidence levels specifically includes:

[0093] (1) Filter vibration signals based on confidence level: Set a confidence level threshold and perform sparse dictionary matching only on real-time vibration signals with confidence levels exceeding the confidence level threshold in order to avoid calculating a large number of irrelevant signal segments.

[0094] (2) Set processing priority according to confidence level: When performing sparse dictionary matching, prioritize the real-time vibration signal according to the confidence level of the real-time vibration signal and process the real-time vibration signal with high confidence level first.

[0095] (3) When solving the optimization objective, confidence is introduced as a weight. Specifically, the orthogonal matching pursuit method is used to solve the optimization objective. When selecting an atom in the m-th iteration, the current residual is denoted as... The selected atom index is:

[0096] ,

[0097] in, The atom index to be selected when selecting atoms in the m-th iteration. For the i-th column of the sparse dictionary of vibrational events, for The weight, and Corresponding event type associations This represents the vector dot product operation.

[0098] If the distributed sensor nodes are at the current location right confidence level (i.e.) or If the value is higher, then for vibration events represented in the sparse dictionary D... weight of atoms The corresponding increase. For example, The calculation method can be:

[0099] ,

[0100] in, Indicates and Corresponding or similar event types, For time t, The real-time sensing data corresponding to the distributed sensor nodes at the location and the historical data The degree of matching or the probability of occurrence.

[0101] The calculation method and or Same, that is Indicates at time t, Real-time sensing data indication of the location of distributed sensor nodes The probability or likelihood of it occurring.

[0102] (4) Narrow the search range of the sparse dictionary of vibration events based on confidence: If the current real-time vibration signal confidence strongly indicates one or more specific event types, then when matching sparse dictionaries, priority is given to searching or only searches are performed in the dictionary subsets related to these specific event types.

[0103] Through the above four methods, confidence effectively guides the matching process of sparse dictionaries, reduces the amount of computation, and speeds up the search. At the same time, by focusing on more relevant signals and dictionary atoms, the accuracy of matching can also be improved.

[0104] This invention also discloses a submarine fiber optic communication and vibration monitoring system based on distributed sensors, such as... Figure 2 As shown, it includes a submarine optical fiber 1, multiple distributed sensor nodes 4, a junction box 6, and a processing unit 5.

[0105] One or both ends of the submarine optical fiber 1 are equipped with a forward communication device 2 (which may be an optical transceiver module) and an OTDR device 3. The forward communication device 2 carries forward communication signals through the submarine optical fiber 1. With the cooperation of the OTDR device 3, the submarine optical fiber 1 acts as a distributed optical fiber sensing medium to perform reverse vibration monitoring based on optical time domain reflection technology. During the reverse vibration monitoring process, the corresponding real-time vibration signal of the submarine optical fiber 1 at the location of the distributed sensor node 4 is obtained.

[0106] Distributed sensor nodes 4 are arranged along the length of the seabed optical fiber 1 and laterally dispersed at specific locations. Each distributed sensor node 4 includes a temperature sensing unit, an acceleration sensing unit, a node microprocessor unit 5, and a node communication unit to acquire real-time sensing data. The node microprocessor unit 5 (which can be an MCU or FPGA) preprocesses the acquired real-time temperature data and real-time vibration acceleration data (e.g., filtering, feature extraction). The node communication unit is connected to the processing unit 5 via a wired link (such as a CAN, RS485, or other wired interface) or a wireless link (such as a cable integrated into the optical fiber or an underwater acoustic communication module) to transmit the temperature data from each distributed sensor node 4. and acceleration data The signal is sent to processing unit 5. Distributed sensor node 4 is connected to submarine fiber optic cable 1 via junction box 6.

[0107] Processing unit 5 matches real-time sensing data with historical data to assess the degree of matching or probability of occurrence of event types in the real-time sensing data and historical data. Historical data includes typical sensing data characteristics of distributed sensor nodes 4 under different known environmental conditions or known event types. Based on the matching results, the confidence level of the real-time sensing data is generated. A vibration monitoring model based on sparse dictionary matching of vibration events is constructed and solved based on the real-time vibration signal. The confidence level is used to improve the event matching and search efficiency during vibration monitoring during the process of solving the vibration monitoring model. The event type corresponding to the real-time vibration signal is obtained based on the solution result of the vibration monitoring model.

[0108] In this embodiment, the processing unit 5 can be a high-performance computer or server cluster located in a shore-based monitoring center. The processing unit 5 is equipped with a data receiving module, a historical database, a matching and confidence generation module, a sparse dictionary matching module, and an event discrimination and output module.

[0109] The data receiving module is used to receive real-time vibration signals acquired during the reverse vibration monitoring process and real-time sensing data sent by the node communication unit.

[0110] The historical database stores the sensing data of distributed sensor node 4 under different known environmental conditions or known event types.

[0111] The matching and confidence generation module matches real-time sensing data with historical data to assess the degree of matching or the probability of occurrence of event types in the real-time sensing data and historical data. The historical data includes typical sensing data characteristics of distributed sensor node 4 under different known environmental conditions or known event types. The confidence of the real-time sensing data is generated based on the matching results.

[0112] The sparse dictionary matching module constructs and solves a vibration monitoring model based on sparse dictionary matching of vibration events according to real-time vibration signals. The process of solving the vibration monitoring model uses confidence to improve the efficiency of event matching and search during vibration monitoring.

[0113] The event discrimination and output module obtains the event type corresponding to the real-time vibration signal based on the solution results of the vibration monitoring model.

[0114] The present invention also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a method for submarine fiber optic communication and vibration monitoring based on distributed sensors.

[0115] The present invention also discloses a device including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a method for submarine fiber optic communication and vibration monitoring based on distributed sensors.

[0116] Compared with the prior art, the advantages of the present invention are:

[0117] (1) Achieving dynamic sensing integration. By using submarine optical fiber as a distributed optical fiber sensing medium for reverse vibration monitoring based on optical time domain reflection technology, vibration monitoring is achieved while ensuring communication. Without affecting the normal communication function of submarine optical fiber, dynamic environmental perception and vibration monitoring along the optical fiber are realized, thereby improving the comprehensive utilization value of submarine optical cables.

[0118] (2) Improve the accuracy of event recognition. By acquiring local real-time sensing data (temperature, acceleration data) through distributed sensor nodes and matching it with historical data to generate confidence scores, environmental noise interference can be effectively filtered out, OTDR data can be used to judge real events, the false alarm rate can be significantly reduced, and the recognition accuracy of specific event types can be improved.

[0119] (3) Enhance event classification capabilities. Historical data includes sensor data under different known environmental conditions or known event types. Combining sensor data and OTDR data can more effectively classify vibration events and improve event classification capabilities.

[0120] (4) Improve event location accuracy. More spatial information can be obtained by using multiple distributed sensor nodes that are horizontally dispersed, which can help the OTDR system to locate the vibration source more accurately and improve the event location accuracy.

[0121] (5) Improve data processing efficiency. Confidence is used in the process of solving the vibration monitoring model. Confidence can guide the sparse dictionary matching process, giving priority to signal regions with high confidence or dictionary atoms related to high confidence event types, thereby reducing unnecessary calculations and improving the efficiency of event matching and searching.

[0122] (6) Enhance system flexibility and scalability. Distributed sensor nodes can be flexibly deployed according to monitoring needs, and the system is easy to expand.

[0123] To further demonstrate the beneficial effects of this invention, a simulation experiment was conducted in this embodiment to simulate oil pipeline monitoring along a 100-kilometer seabed optical fiber. A distributed sensor node (a total of 101 nodes) was deployed every 1 kilometer along the seabed optical fiber. Each node included a PT100 temperature sensor and a sensor with a sensitivity of [missing information]. The device is a three-axis MEMS accelerometer. The sensor node is powered and communicates via RS485 through a spare copper wire in an optical fiber, and the processing unit is located in the shore-based control room.

[0124] Historical data is built in the laboratory or on actual pipe sections to simulate different types of events, including:

[0125] The first type of known historical event (denoted as...) : Normal background (stable water flow, noise from normal equipment operation).

[0126] The second type of known historical event (denoted as...) ): Small leaks (characteristics: temperature may drop slightly, acceleration vibrations at a specific frequency).

[0127] The third type of known historical event (denoted as...) Third-party intrusion (such as anchor smashing, characterized by: violent, short-term high-frequency acceleration vibration, with no significant temperature change).

[0128] The fourth type of known historical event (denoted as...) : Ship approaching (characteristics: low-frequency, continuous acceleration vibration, no significant temperature change).

[0129] Simulate temperature and acceleration data from various sensor nodes under the four event types described above, and extract features (such as energy, kurtosis, and kurtosis in the 0-10Hz, 10-50Hz, and 50-200Hz frequency bands of the acceleration signal; mean temperature and rate of change). Use this labeled feature data to train a multi-class Gaussian mixture model (GMM) or an SVM classifier, using historical data. The model outputs the confidence level of a given sensor feature vector belonging to each event type.

[0130] For the four events mentioned above, simulated OTDR (such as φ-OTDR) vibration signals are used to extract typical time-frequency characteristics of the φ-OTDR signal under different events (such as spectral blocks after using Short Time Fourier Transform (STFT) or wavelet packet decomposition coefficients) as dictionary atoms to construct the event sparse dictionary D. During normal operation of the simulation system, the OTDR device continuously monitors the entire optical cable, and the sensor nodes report temperature and acceleration data once per second.

[0131] Next, two scenarios were simulated: a suspected small leak incident and the elimination of a false alarm.

[0132] In a suspected minor leak incident, sensor node #51, located at kilometer 50.5 of the optical cable, reported the following data: temperature decreased by 0.5°C compared to the previous average, and the Z-axis acceleration exhibited a continuous, weak vibration peak between 15-25 Hz. The matching and confidence generation module of the processing unit processed the real-time characteristics of node #51. Matching with historical data. GMM model output: , , , This indicates that the node's data highly matches the characteristics of a small leak (…). The OTDR system also detected weak abnormal vibration signals in the 50-51 km range. The sparse dictionary matching module received a high confidence score from node #51. ) information. Therefore, in the context of... When performing sparse matching, priority will be given to using the same as... The relevant dictionary atoms are assigned higher weights to the matching results of these atoms. The matching results show that the φ-OTDR signal has a good match with the atoms related to small leaks in the dictionary, and the sparsity coefficient is significant. The event discrimination and output module jointly judges that the probability of a small leak event occurring at 50.6 km is very high, and issues an early warning.

[0133] During false alarm elimination, the OTDR system detected a sudden vibration signal at 70 km, consistent with a third-party intrusion detected in the dictionary. A certain atom in the sample shows some similarity. However, the temperature and acceleration data reported by sensor node #70 at 70 kilometers, after matching with historical models, show that they are more consistent with the normal background. ) or the ship is close to ( The low-frequency vibrations of ) are harmful to third-party intrusion. ) confidence level Extremely low (e.g., less than 0.1). The processing unit considers that although the OTDR signal is somewhat suspicious, the local sensor data does not support a severe impact event. Therefore, this OTDR signal is judged as a normal environmental disturbance caused by a distant vessel, or noise fluctuations within the OTDR system itself, thus avoiding a false alarm.

[0134] The simulation experiments above demonstrate that this invention, by introducing local high-confidence information provided by distributed sensor nodes, effectively guides the analysis of global OTDR data, significantly improving the classification ability, accuracy, and processing efficiency of event recognition.

[0135] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0136] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0137] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0138] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0139] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for submarine fiber optic communication and vibration monitoring based on distributed sensors, characterized in that, include: Using submarine optical fibers to carry forward communication signals, the submarine optical fibers are used as distributed optical fiber sensing media for reverse vibration monitoring based on optical time domain reflection technology. Multiple distributed sensor nodes are set along the length of the submarine optical fiber. The distributed sensor nodes acquire real-time sensing data and acquire the corresponding real-time vibration signal of the submarine optical fiber at the location of the distributed sensor node during the reverse vibration monitoring process. The real-time sensing data is matched with historical data to assess the degree of matching or the probability of occurrence of event types in the real-time sensing data and the historical data. The historical data includes sensing data of distributed sensor nodes under different known environmental conditions or known event types. The confidence level of the real-time sensing data is generated based on the matching results. A vibration monitoring model based on sparse dictionary matching of vibration events is constructed and solved based on the real-time vibration signal. The event type corresponding to the real-time vibration signal is obtained based on the solution result of the vibration monitoring model. The confidence level is used to improve the event matching and search efficiency during vibration monitoring in the process of solving the vibration monitoring model. When constructing and solving a vibration monitoring model based on sparse dictionary matching of vibration events using the real-time vibration signals, an optimization objective for the vibration monitoring model is established. The process of solving the vibration monitoring model utilizes the confidence level to improve the event matching and search efficiency during vibration monitoring. This includes introducing the confidence level as a weight when solving the optimization objective. Specifically, the optimization objective is solved using the orthogonal matching pursuit method. When selecting an atom in the m-th iteration, the current residual is denoted as r. m-1 The selected atom index is: Among them, idx m d is the atom index selected when choosing an atom in the m-th iteration. i For the i-th column of the sparse dictionary of vibrational events, w i For d i The weight, w i With d i The corresponding event type is associated, and <> represents the vector inner product operation.

2. The method for submarine fiber optic communication and vibration monitoring based on distributed sensors according to claim 1, characterized in that: The distributed sensor node includes a temperature sensing unit and an acceleration sensing unit. The temperature sensing unit collects real-time temperature data of the surrounding environment, and the acceleration sensing unit collects real-time vibration acceleration data in the X-axis, Y-axis and Z-axis directions. The real-time sensing data is fused with the real-time temperature data and the real-time vibration acceleration data, specifically as follows: Where, x real (l j Let (t) be the real-time sensing data of the j-th distributed sensor node at time t, and l j This indicates the location of the j-th distributed sensor node. Let be the average real-time temperature data of the j-th distributed sensor node at time t within the time window, var(T(l) j Let A be the variance of the real-time temperature data of the j-th distributed sensor node at time t within the time window. rms,x (l j Let A be the root mean square value of the real-time vibration acceleration data of the j-th distributed sensor node along the X-axis at time t within the time window. rms,y (l j Let A be the root mean square value of the real-time vibration acceleration data of the j-th distributed sensor node along the Y-axis at time t within the time window. rms,z (l j Let f be the root mean square value of the real-time vibration acceleration data of the j-th distributed sensor node along the Z-axis at time t within the time window. p (l j ) is the main frequency, For Fourier transform, argmax f This represents the frequency point with the highest energy in the spectrum, A(l) j E(t) represents the real-time vibration acceleration data of the j-th distributed sensor node at time t. A (l j E represents energy. A (l j )=∫Δt||A(l j ,t)|| 2 dt,||A(l j ,t)|| represents vector A(l j The length of ,t), where Δt is the length of the time window, and T represents the transpose operation.

3. The method for submarine fiber optic communication and vibration monitoring based on distributed sensors according to claim 2, characterized in that: The real-time sensing data is matched with historical data, specifically as follows: When the kth known historical event type occurs, observe x. real (l j The probability of x,t) is denoted as P(x,t). real (l j ,t)|E k ); The confidence level of the real-time sensing data generated based on the matching results is: Among them, E k Let C(l) represent the k-th known historical event type. j ,t,E k C(l) represents the confidence level of real-time sensor data under the probabilistic model. j ,t,E k ) indicates that at time t, l j The real-time sensing data indication E corresponding to the location of the distributed sensor nodes k The probability of occurrence, P(E) k ) is E k The prior probability is M, where M is the number of known historical event types.

4. The method for submarine fiber optic communication and vibration monitoring based on distributed sensors according to claim 2, characterized in that: The real-time sensing data is matched with historical data, specifically as follows: Let E be the type of the k-th known historical event. k Using E k The corresponding historical data is used to train the SVM model, and the normal vector parameters of the hyperplane of the SVM model after training are obtained as follows: w k, the bias term is b k ; Calculate real-time sensor data and E k The decision function is: Among them, f k (x real (l j ,t)) represents real-time sensor data and E k The decision function, represented by E k and x real (l j The matching result of ,t), where φ() is the Gaussian kernel function; The confidence level of the real-time sensing data is generated based on the matching results.

5. The method for submarine fiber optic communication and vibration monitoring based on distributed sensors according to claim 4, characterized in that: The confidence level for generating the real-time sensing data based on the matching result is: Among them, C(l) j ,t,E k )′ represents the confidence level of real-time sensing data under the matching model, C(l)′ is the confidence level of the data. j ,t,E k )′ represents at time t, l j The real-time sensing data indication E corresponding to the location of the distributed sensor nodes k The probability of occurrence, α k β k These are preset parameters.

6. The method for submarine fiber optic communication and vibration monitoring based on distributed sensors according to any one of claims 1-5, characterized in that: A vibration monitoring model based on sparse dictionary matching of vibration events is constructed and solved according to the real-time vibration signal. The event type corresponding to the real-time vibration signal is obtained from the solution result of the vibration monitoring model, specifically: The real-time vibration signal at time t along the length l of the coastal fiber is divided to obtain a vibration signal segment, denoted as sOTDR(l). j ,t), N is the dimension of the vibration signal segment, l j This indicates the location of the j-th distributed sensor node corresponding to the real-time vibration signal; Construct a sparse dictionary of vibration events, denoted as D. Each column of atoms in D represents the representation of the vibration signal corresponding to a certain event type on the OTDR signal. K is the number of atoms in the sparse dictionary of vibrational events; The optimization objective for constructing the vibration monitoring model is: or Where α is the sparse coefficient vector to be solved, ||||0 is the L0 normal form, ||||2 is the L2 normal form, ε is the reconstruction error tolerance, λ is the regularization parameter, and ||||1 is the L1 normal form; Solving the optimization objective yields the optimal sparse coefficient vector, denoted as . according to The atoms of the sparse dictionary of events corresponding to non-zero elements are used to obtain the corresponding event types.

7. The method for submarine fiber optic communication and vibration monitoring based on distributed sensors according to claim 6, characterized in that: Improving event matching and search efficiency in vibration monitoring using the aforementioned confidence level includes: Set a confidence threshold and perform sparse dictionary matching only on real-time vibration signals with confidence levels exceeding the confidence threshold; When performing sparse dictionary matching, the real-time vibration signals are prioritized based on their confidence level, with high-confidence real-time vibration signals being processed first. If the current real-time vibration signal confidence indicates one or more specific event types, then during sparse dictionary matching, a search is prioritized or performed only on a subset of dictionaries related to these specific event types.

8. The method for submarine fiber optic communication and vibration monitoring based on distributed sensors according to claim 1, characterized in that: w i The calculation method is as follows: w i =C(l j ,t,E type (d i ))”, Among them, E type (d i ) indicates the relationship with d i Corresponding or similar event types, C(l j ,t,E type (d i "))" represents l at time t j The real-time sensing data corresponding to the distributed sensor nodes at the location and the E in the historical data type (d i The degree of matching or the probability of occurrence of a corresponding event.

9. A submarine fiber optic communication and vibration monitoring system based on distributed sensors, characterized in that: Includes submarine optical fiber, multiple distributed sensor nodes, and processing units. One or both ends of the submarine optical fiber are equipped with a forward communication device and an OTDR device. The forward communication device carries forward communication signals through the submarine optical fiber. With the cooperation of the OTDR device, the submarine optical fiber serves as a distributed optical fiber sensing medium for reverse vibration monitoring based on optical time domain reflection technology. During the reverse vibration monitoring process, the corresponding real-time vibration signal of the submarine optical fiber at the location of the distributed sensor node is acquired. The distributed sensor nodes are set along the length of the submarine optical fiber to acquire real-time sensing data. The processing unit matches the real-time sensing data with historical data to assess the degree of matching or the probability of occurrence of event types in the real-time sensing data and the historical data. The historical data includes typical sensing data characteristics of distributed sensor nodes under different known environmental conditions or known event types. Based on the matching results, the processing unit generates the confidence score of the real-time sensing data. Based on the real-time vibration signal, a vibration monitoring model based on sparse dictionary matching of vibration events is constructed and solved. The confidence score is used to improve the event matching and search efficiency during vibration monitoring during the process of solving the vibration monitoring model. Based on the solution result of the vibration monitoring model, the event type corresponding to the real-time vibration signal is obtained. When constructing and solving a vibration monitoring model based on sparse dictionary matching of vibration events using the real-time vibration signals, an optimization objective for the vibration monitoring model is established. The process of solving the vibration monitoring model utilizes the confidence level to improve the event matching and search efficiency during vibration monitoring. This includes introducing the confidence level as a weight when solving the optimization objective. Specifically, the optimization objective is solved using the orthogonal matching pursuit method. When selecting an atom in the m-th iteration, the current residual is denoted as r. m-1 The selected atom index is: Among them, idx m d is the atom index selected when choosing an atom in the m-th iteration. i For the i-th column of the sparse dictionary of vibrational events, w i For d i The weight, w i With d i The corresponding event type is associated, and <> represents the vector inner product operation.

Images