Health monitoring system and method for water diversion engineering optical fiber network

Abstract

The invention relates to the technical field of optical fiber fault detection, in particular to a health monitoring system and method for a water diversion engineering optical fiber network, and the system comprises a sensing module which is used for synchronously collecting optical fiber environment parameters through an edge calculation node and a multi-source sensor array; the transmission network construction module is used for carrying out short-distance networking and constructing a backbone transmission network by adopting an industrial grade PON (Passive Optical Network) optical fiber network; the cloud management platform is used for preprocessing the data acquired by the transmission network construction module; inputting the preprocessed data into a preset D-S evidence theory fusion model to obtain an early warning trigger signal; and the fault classification module is also used for carrying out fault positioning and classification through the trained fault classification model after the edge node detects an abnormal signal, and outputting a fault position and a fault type. According to the invention, the problems of insufficient real-time performance and incapability of dealing with sudden faults in the prior art can be solved; the positioning precision and the fault identification capability are limited; the long-distance deployment cost is high; and the management complexity is high.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of optical fiber fault detection, and particularly relates to a health monitoring system and method for optical fiber network of water diversion project. BACKGROUND

[0002] In large-scale water diversion projects, optical cable channels are the core infrastructure supporting intelligent operation of the project, and undertake key tasks such as transmission of water diversion scheduling instructions, monitoring of equipment status, and video image return. Due to the fact that the project line is usually tens to hundreds of kilometers long and passes through complex geological areas (such as tunnels, rivers, and high-intensity earthquake zones), the optical cable faces multiple environmental threats: geological disasters: soil settlement, rock layer fracture, etc. cause mechanical stress concentration of the optical cable, leading to fracture or signal attenuation; climate factors: extreme temperature changes (-40℃ ~ +70℃) cause aging of the optical cable sheath material, and humidity erosion (above 95% RH) causes failure of the optical fiber connector.

[0003] Traditional monitoring methods rely on manual inspection or single sensor detection, which has the following problems: poor real-time performance: unable to monitor sudden failures all day long; low positioning accuracy: difficult to quickly locate the fault point; limited coverage: long-distance projects require deployment of a large number of devices, which is costly and complex to manage. Human interference: construction excavation, theft, etc. directly damage the physical structure of the optical cable. If the above risks are not monitored in a timely manner, it may lead to communication interruption, and further cause safety accidents such as loss of control of water diversion equipment and delay in disaster warning. Therefore, building an optical cable state monitoring system covering the entire line is an urgent need to ensure the safe operation of the water diversion project.

[0004] The existing technologies in the field of optical cable monitoring at present mainly fall into three categories, and their core architectures and technical features are as follows:

[0005] The first category: optical fiber physical property detection technology based on OTDR; technical principle: through optical time domain reflectometer (OTDR), pulse light is transmitted to the optical cable, and backscattered light signals are received, and the location of the optical fiber fracture point or the abnormal area of loss is located according to the signal attenuation curve, and the positioning accuracy is theoretically 1-5 meters (depending on the pulse width). Typical application: mainly used for periodic detection (such as once a day) of optical fiber link, suitable for static loss monitoring.

[0006] The second category: environmental monitoring technology based on single sensor; vibration monitoring system: acceleration sensors or optical fiber vibration sensors are deployed along the optical cable, human damage behaviors such as excavation and knocking are identified by analyzing the vibration signal spectrum, and the positioning accuracy is about 100-200 meters (dependent on the sensor spacing). Temperature and humidity monitoring network: distributed temperature and humidity sensors (such as DHT11) are used for networking, data is transmitted through RS485 or LoRa, and the environmental parameters in the optical cable joint box are monitored, and the oxidation problem of the connector caused by excessive humidity is warned.

[0007] The third type: manual inspection combined with offline detection; manual inspection: regularly (such as once a week) dispatching operation and maintenance personnel to patrol along the cable route, and detecting obvious fault points through visual inspection and portable OTDR tester. Offline analysis: collecting detection data to the management platform, and relying on manual experience to judge the fault type, lacking real-time performance.

[0008] The core problems and shortcomings of the prior art are:

[0009] 1. Lack of real-time performance, unable to respond to sudden failures: OTDR detection needs to be manually triggered or executed at regular intervals (several hours to several days), and cannot capture instantaneous breakage events (such as sudden breakage caused by geological disasters); a single vibration sensor is only sensitive to a specific type of interference (such as mechanical vibration), and has no monitoring capability for temperature sudden change, slow stress accumulation and other progressive failures (accounting for more than 60%).

[0010] 2. Limited positioning accuracy and fault recognition capability: OTDR has a "blind area" problem (near-end blind area 5-10 meters, event blind area 1-2 meters), and cannot accurately identify the fault location in a short-distance dense joint scene (such as one joint every 50 meters in a tunnel); a single sensor relies on threshold judgment (such as vibration amplitude exceeding 5g to alarm), and is easily disturbed by environmental noise (such as wind and rain vibration), with a false alarm rate of up to 15%-20%, leading to frequent misjudgment by operation and maintenance personnel.

[0011] 3. High deployment cost for long distance and high management complexity: the traditional scheme needs to independently deploy a sensor network for each monitoring function (strain, vibration, temperature and humidity), with high equipment cost per kilometer; multiple system data are independently stored (such as vibration data in platform A and OTDR data in platform B), forming a "data island", lacking cross-dimension correlation analysis capability, and the fault diagnosis relies on manual comprehensive judgment, with low efficiency. SUMMARY

[0012] One of the purposes of the present application is to provide a health monitoring system for water diversion project optical fiber network, solving the above problems.

[0013] In order to achieve the above purpose, a health monitoring system for water diversion project optical fiber network is provided, comprising:

[0014] The perception module is used for synchronously collecting optical fiber environment parameters through the edge computing node and the multi-source sensor array; the multi-source sensor array comprises a strain sensor, a vibration sensor, a temperature and humidity sensor and a tension sensor;

[0015] The transmission network construction module is used for constructing a star network through LoRa between adjacent edge computing nodes for short-distance networking; an industrial PON optical fiber network is adopted, so that the edge nodes are connected to an optical line terminal (OLT) through an optical splitter, and a point-to-multipoint backbone transmission network is constructed;

[0016] The cloud management platform is configured to preprocess data obtained by the transmission network construction module to obtain vibration signals, strain data and temperature and humidity data, input the preprocessed data into a preset D-S evidence theory fusion model, obtain a pre-warning trigger signal indicating whether a fault pre-warning is triggered, define the vibration signals, the strain data and the temperature and humidity data as three evidence bodies in the D-S evidence theory fusion model, and perform fusion calculation and confidence calculation according to a BPA function of a single evidence body, and when a final confidence reaches a set threshold, generate the pre-warning trigger signal, and is further configured to perform fault positioning and classification by using a trained fault classification model after the edge node detects an abnormal signal, and output a fault position and a fault type.

[0017] Further, the vibration signals are obtained by removing environmental noise by using a 50-500H fiber fault detection band-pass filter, and a plurality of frequency components are extracted as characteristic values by using fast Fourier transform; the strain data are obtained by filtering high-frequency interference by using a sliding average method, and a strain change rate per minute is calculated as a characteristic value; the temperature and humidity data are obtained by polynomial fitting, and a deviation between a real-time value and a historical average value is calculated as a characteristic value after a trend item is removed; and the strain data are obtained by filtering high-frequency interference by using a sliding average method, and a strain change rate per minute is calculated as a characteristic value.

[0018] Further, the cloud management platform comprises a fault positioning module.

[0019] The fault positioning module is configured to trigger two-end OTDR equipment to synchronously collect optical signals after the edge node detects an abnormal signal, and obtain backscattering curves S1(x) and S2(x); perform wavelet denoising on the curves, identify loss mutation points x1 and x2, and extract loss gradient features; and establish a loss gradient-physical distance mapping relationship by using an SVM regression model, and a calculation formula is as follows:

[0020]

[0021] Wherein, c is the speed of light, t1 and t2 are two-end signal transmission times, and n is the refractive index of the optical fiber.

[0022] Further, the cloud management platform comprises a fault classification module.

[0023] The fault classification module is configured to collect a fault sample library, and construct a fault classification model based on a BP neural network structure; and is further configured to input an abnormal signal detected by the edge node into the trained fault classification model, and output a fault type.

[0024] Further, the edge computing node adopts a lightweight structure design, including a processor, a LoRa / PON dual-communication module architecture, a solar cell panel and a lithium battery.

[0025] Further, the cloud management platform further comprises:

[0026] The visualization module is used for loading the fault points into the GIS map in correspondence with the fault type output by the fault classification module and displaying the cable state in real time based on the GIS map, and checking the historical data curve within a set time through the fault points.

[0027] The second object of the present application is to provide a health monitoring method for a water diversion engineering optical fiber network, comprising:

[0028] The sensing step is to synchronously collect optical fiber environment parameters through the edge computing node and the multi-source sensor array; the multi-source sensor array comprises a strain sensor, a vibration sensor, a temperature and humidity sensor and a tension sensor;

[0029] The transmission network construction step is to form a star network through LoRa between adjacent edge computing nodes for short-distance networking; an industrial-grade PON optical fiber network is adopted to enable the edge nodes to access an optical line terminal (OLT) through an optical splitter to construct a point-to-multipoint backbone transmission network;

[0030] The health monitoring step is to pre-process the data obtained by the transmission network construction module to obtain vibration signals, strain data and temperature and humidity data; the pre-processed data is input into a preset D-S evidence theory fusion model to obtain an early warning trigger signal for triggering a fault early warning, wherein the vibration signals, the strain data and the temperature and humidity data are defined as three evidence bodies in the D-S evidence theory fusion model, and fusion calculation and confidence calculation are performed according to the BPA function of a single evidence body; when the final confidence reaches a set threshold, the early warning trigger signal is generated; the health monitoring step is also used for performing fault positioning and classification through a trained fault classification model after the edge node detects an abnormal signal, and outputting a fault position and a fault type.

[0031] Further, the vibration signals are obtained by removing environmental noise through a 50-500H optical fiber fault detection band-pass filter and extracting a plurality of frequency components as characteristic values through fast Fourier transform; the strain data are obtained by filtering high-frequency interference through a moving average method and calculating a strain change rate per minute as a characteristic value; the temperature and humidity data are obtained by performing polynomial fitting, calculating the deviation of a real-time value from a historical mean value as a characteristic value after removing the trend term; and the strain data are obtained by filtering high-frequency interference through a moving average method and calculating a strain change rate per minute as a characteristic value.

[0032] Further, the health monitoring step specifically comprises the following steps:

[0033] Fault positioning step: after the edge node detects an abnormal signal, trigger the two-end OTDR equipment to synchronously collect optical signals, obtain backscattering curves S1(x), S2(x); wavelet denoising is performed on the curves, loss mutation points x1, x2 are identified, and loss gradient features are extracted; a loss gradient-physical distance mapping relationship is established using a SVM regression model, and the calculation formula is:

[0034]

[0035] Wherein, c is the speed of light, t1, t2 are the signal transmission times of the two ends respectively, and n is the refractive index of the optical fiber.

[0036] Further, the health monitoring step specifically comprises the following steps:

[0037] Fault classification step: collect fault sample library, and construct fault classification model based on BP neural network structure; input the abnormal signal detected by the edge node into the trained fault classification model, and output the fault type.

[0038] Principle and advantage:

[0039] 1. Hardware dimension: multi-dimensional perception and lightweight deployment

[0040] 1) Multi-source sensor array integration technology

[0041] The optical fiber strain sensor, vibration acceleration sensor, temperature and humidity composite sensor and tension sensor are physically integrated to build a three-dimensional perception system containing 12 core monitoring parameters (strain, vibration frequency, temperature and humidity, tension, etc.), covering three major types of fault inducements, i.e. mechanical stress of optical cable, environmental erosion and human interference.

[0042] Technical effect: Compared with the traditional single sensor scheme, the integrity of the monitoring parameters is improved by 300%, and 8 typical fault modes (such as strain mutation caused by soil settlement and connector oxidation caused by high temperature and humidity) can be identified.

[0043] 2) Light weight design of edge computing node

[0044] Adopting processor + LoRa / PON dual communication module architecture, integrating Modbus / TCP protocol conversion function, realizing 500H optical fiber fault detection real-time data acquisition and local preprocessing, and supporting solar + lithium battery hybrid power supply.

[0045] Technical effect: The coverage radius of a single node reaches 2km, the deployment density of sensors is reduced by 30% compared with the traditional scheme, the equipment cost per kilometer is reduced from 50,000 yuan to less than 35,000 yuan, and the problem of high cost of long-distance engineering deployment is solved.

[0046] 2. Algorithm dimension: intelligent fusion and accurate diagnosis

[0047] 1) D-S evidence theory multi-source data fusion algorithm

[0048] Define the differentiated weights of vibration signals (ω1), strain data (ω2), and temperature and humidity (ω3), construct a fusion model containing 12 basic probability assignment (BPA) rules, and trigger an early warning when the confidence reaches a threshold.

[0049] Technical effects: The fault recognition accuracy is improved from 80% of the traditional single sensor to 98.7%, the false positive rate is reduced from 18% to 1.3%, and the misjudgment problem caused by environmental noise interference is solved.

[0050] 2) Double-end OTDR-SVM regression positioning algorithm

[0051] Through double-end synchronous acquisition of optical signals, an SVM regression model is used to establish a "loss gradient-physical distance" mapping relationship.

[0052] Technical effects: The positioning accuracy is improved from the traditional single-end OTDR of 100 meters to ≤50 meters, the near-end blind area is reduced from 10 meters to 2 meters, and the rapid and accurate positioning of the fault point is realized. BRIEF DESCRIPTION OF DRAWINGS

[0053] Figure 1 It is an architecture diagram of a health monitoring system for a water diversion project optical fiber network according to an embodiment of the present application;

[0054] Figure 2 It is a hardware schematic diagram of an edge computing node;

[0055] Figure 3 It is a multi-source data fusion flowchart;

[0056] Figure 4 It is a fault positioning flowchart;

[0057] Figure 5 It is a BP neural network structure diagram. DETAILED DESCRIPTION

[0058] The following will be further described in detail through specific embodiments:

[0059] EMBODIMENT

[0060] A health monitoring system for a water diversion project optical fiber network, substantially as shown in Figure 1 , Figure 2 , Figure 3 , comprising:

[0061] The perception module is used for synchronously collecting optical fiber environment parameters through the edge computing node and the multi-source sensor array; the multi-source sensor array includes a strain sensor, a vibration sensor, a temperature and humidity sensor, and a tension sensor, the multi-source sensor array is a composite sensor group, and one group is arranged every 50 meters, and important nodes are encrypted to 20 meters.

[0062] The strain sensor is pasted on the surface of the cable armor layer to monitor the axial stress; in the embodiment, the strain sensor is a fiber optic strain sensor (model: FBG-ST01); the Bragg fiber grating principle is adopted, the strain sensor is pasted on the surface of the cable armor layer, and the axial strain (precision ±0.001με) is monitored; the strain sensor is rigidly connected with the cable through 3M epoxy resin glue.

[0063] The vibration sensor is fixed on the cable support and arranged next to the strain sensor; in the embodiment, the vibration sensor is a vibration acceleration sensor (model: ADXL375); the range is ±16g, a built-in 200H fiber optic fault detection low-pass filter is arranged, the vibration sensor is fixed on the cable support through a magnetic clamp, and the distance between the vibration sensor and the strain sensor is ≤10cm.

[0064] The temperature and humidity sensor is a temperature and humidity composite sensor (model: SHT30); the temperature and humidity sensor integrates a temperature (precision ±0.3℃) and humidity (precision ±2% RH) sensing module, is packaged in an IP68 protective shell, and is connected with the edge computing node through an RS485 bus.

[0065] The tension sensor measures the axial tension of the cable; in the embodiment, the tension sensor (model: HBM U2A) is connected in series between the cable fixing fitting and the tower, measures the axial tension of the cable (precision 0.1% FS), and outputs a 4-20mA analog signal.

[0066] The edge computing node adopts a lightweight structure design, includes an ARM Cortex-A72 processor (main frequency 1.8GHz), integrates 2GB DDR4 memory and 16GB eMMC storage, supports Python / C++ algorithm deployment, a LoRa / PON double communication module architecture (simultaneously configures a LoRa wireless module (frequency band 470MHz, transmission power 20dBm, coverage radius 2km) and an industrial PON optical fiber module (supports a 1.25Gbps rate, 200km non-repeater transmission)), is connected with the multi-source sensor array through an RJ45 network port, and a power supply system; the power supply system includes a solar cell panel (10W) and a lithium battery (12V / 20Ah), is also provided with an energy management chip, a sleep current is <10μA, and supports a wide temperature operation of-40℃ to +70℃.

[0067] Transmission network building module: used for constructing a star network by LoRa between adjacent edge computing nodes for short-distance networking; a single gateway can access 200 sensor nodes to realize low-power wide-area connection (power consumption reduced by 80% compared with traditional WiFi). An industrial PON optical fiber network is adopted to enable edge nodes to access an optical line terminal (OLT) through an optical splitter (1:32) to build a point-to-multipoint backbone transmission network; IP data encapsulation and QoS (quality of service) management are supported to ensure real-time transmission of monitoring data.

[0068] Cloud management platform: including a data center, an AI analysis engine and a visualization module.

[0069] Data center: based on a Hadoop distributed architecture, containing an HDFS storage cluster (single cluster capacity ≥ 10 PB) and a Kafka message queue, supporting concurrent writing of 100,000 data per second, and a storage period ≥ 5 years.

[0070] AI analysis engine: integrating TensorFlow / PyTorch framework, deploying D-S evidence theory fusion model, fault classification model based on SVM regression positioning algorithm and BP neural network classifier, supporting online training of D-S evidence theory fusion model and fault classification model (weight parameters are automatically updated once per hour).

[0071] Visualization module: the visualization module is used for loading fault points corresponding to the fault type output by the fault classification module into a GIS map to display the cable state in real time based on a GIS map, and viewing historical data curves within a set time through the fault point. In the embodiment, a 0.1-meter-precision GIS map is developed based on GIS to display the cable state in real time (green-normal, yellow-prewarning, red-fault), and the historical data curves (strain, vibration, temperature and humidity, etc.) within 30 days can be viewed by clicking the fault point.

[0072] AI analysis engine: used for preprocessing the data obtained by the transmission network building module to obtain vibration signals, strain data and temperature and humidity data; inputting the preprocessed data into a preset D-S evidence theory fusion model to obtain an early warning trigger signal for triggering a fault early warning, wherein the vibration signals, the strain data and the temperature and humidity data are defined as three evidence bodies in the D-S evidence theory fusion model, and fusion calculation and confidence calculation are performed according to the BPA function of a single evidence body, and when the final confidence reaches a set threshold, the early warning trigger signal is generated; and used for performing fault positioning and classification through a trained fault classification model after an abnormal signal is detected at an edge node, and outputting a fault position and a fault type.

[0073] Wherein the pretreatment: the vibration signal: using 50-500H fiber fault detection band pass filter to remove environmental noise, extracting several frequency components (50H fiber fault detection, 100H fiber fault detection, 200H fiber fault detection, etc.) as characteristic values by fast Fourier transform; the strain data: using the moving average method (window size 100 points) to filter out high frequency interference, calculating the strain change rate (Δε / min) as characteristic value every minute; the temperature and humidity data: polynomial fitting (3 times polynomial), calculating the deviation (ΔT / ΔRH) of real-time value and historical mean value as characteristic value after removing the trend term.

[0074] The calculation and analysis of the D-S evidence theory fusion model includes the following modules:

[0075] The basic parameter analysis module: used for analyzing the basic parameters; the basic parameters include:

[0076] Fault type set: θ={F1,F1,...,F8}; including mechanical fracture, stress caused by soil settlement, human excavation damage, optical cable sheath aging, connector oxidation, wind vibration fatigue, temperature sudden change damage and loose fastener caused tension anomaly;

[0077] Evidence body: vibration signal E1, strain data E2 and temperature and humidity data E3;

[0078] Single evidence body BPA function: m i (F j ), represents the support degree of the ith evidence body to the jth fault;

[0079] Uncertainty:

[0080] Preliminary fusion calculation module: used for selecting any 2 evidence bodies from the vibration signal E1, strain data E2 and temperature and humidity data E3, and calculating the first joint support sum according to the joint support sum calculation formula, and calculating the first conflict coefficient of the 2 evidence bodies, reflecting the contradiction degree between the 2 evidence bodies; calculating the intermediate support degree according to the first joint support sum and the first conflict coefficient; the joint support sum calculation formula is as follows:

[0081] S xy (F i )=m x (F j )×m y (F j )+m x (F j )×m y (θ)+m x (θ)×m y (F j ), xy∈i

[0082]

[0083] Among them, S xy (F i ) represents the sum of the first joint support, m xy (F i K represents the intermediate support. xy The first conflict coefficient;

[0084] The secondary fusion calculation module is used to fuse the intermediate support with the remaining evidence, calculate the second joint support sum according to the joint support sum calculation formula, and calculate the second conflict coefficient to reflect the degree of contradiction among the three pieces of evidence; and calculate the final confidence level based on the second joint support sum and the second conflict coefficient.

[0085] S xyz (F i ) = m xy (F j )×m z (F j )+m xy (F j )×m z (θ)+m xy (θ)×m z (F j ), xyz∈i

[0086]

[0087] Among them, S xyz (F i ) represents the sum of the second joint support, m xyz (F i K represents the final confidence level. xyz This is the second conflict coefficient.

[0088] The AI ​​analysis engine specifically includes a fault location module and a fault classification module.

[0089] Fault location module: such as Figure 4 As shown, this is used to trigger the synchronous acquisition of optical signals by the two OTDR devices after an abnormal signal is detected at the edge node, obtaining backscattering curves S1(x) and S2(x); wavelet denoising (db4 wavelet, 3-level decomposition) is performed on the curves to identify loss abrupt change points x1 and x2, and loss gradient features are extracted; the loss gradient-physical distance mapping relationship is established using an SVM regression model, and the calculation formula is as follows:

[0090]

[0091] Wherein, c is the speed of light, t1, t2 are the signal transmission time of both ends respectively, n is the refractive index of optical fiber (1.467). The positioning accuracy is less than or equal to 50 meters (50% higher than the traditional single-end OTDR).

[0092] The fault classification module is used for collecting a fault sample library and constructing a fault classification model based on a BP neural network structure, as shown in the formula (1) ; and is also used for inputting the abnormal signal detected by the edge node into the trained fault classification model to output the fault type. Figure 5

[0093] In the embodiment, a fault sample library containing 8 types of faults (≥1000 samples per type) is established; the input parameters corresponding to the abnormal signal include 12 characteristics such as strain peak value, vibration main frequency, temperature gradient and humidity threshold; and the fault sample library is shown in the following table:

[0094]

[0095] The BP neural network structure: 12 neurons in the input layer, 2 layers (30 / 15 neurons each) in the hidden layer, 8 neurons (corresponding to the fault type) in the output layer, using Adam optimizer (learning rate 0.001), and training until the classification accuracy is greater than or equal to 98%.

[0096] A health monitoring method for a water diversion engineering optical fiber network, comprising the following steps:

[0097] The sensing step: synchronously collecting optical fiber environment parameters by the edge computing node and the multi-source sensor array; the multi-source sensor array includes strain sensors, vibration sensors, temperature and humidity sensors and tension sensors;

[0098] The transmission network construction step: adjacent edge computing nodes are connected through LoRa to form a star network for short-distance networking; an industrial-grade PON optical fiber network is used to enable the edge nodes to access an optical line terminal (OLT) through an optical splitter to construct a point-to-multipoint backbone transmission network;

[0099] The health monitoring step: pre-processing the data obtained by the transmission network construction module to obtain vibration signals, strain data and temperature and humidity data; inputting the pre-processed data into a preset D-S evidence theory fusion model to obtain a pre-alarm trigger signal for triggering a fault pre-alarm, wherein the vibration signals, strain data and temperature and humidity data are defined as three evidence bodies in the D-S evidence theory fusion model, and fusion calculation and confidence calculation are performed according to the BPA function of a single evidence body; when the final confidence reaches a set threshold, the pre-alarm trigger signal is generated; the trained fault classification model is also used for fault positioning and classification after the edge node detects an abnormal signal, and outputs the fault position and fault type.

[0100] ​The calculation and analysis of the D-S evidence theory fusion model includes the following steps:

[0101] The basic parameter analysis step: analyzing the basic parameters; the basic parameters include:

[0102] The fault type set: θ = {F1, F1,..., F8};

[0103] The evidence body: vibration signal E1, strain data E2 and temperature and humidity data E3;

[0104] The BPA function of a single evidence body: m i (F j ) represents the support degree of the ith evidence body to the jth fault type;

[0105] Uncertainty:

[0106] The preliminary fusion calculation step: selecting any two evidence bodies from the vibration signal E1, strain data E2 and temperature and humidity data E3, and calculating the first joint support sum according to the joint support sum calculation formula; and calculating the first conflict coefficient of the two evidence bodies, reflecting the contradiction degree between the two evidence bodies; calculating the intermediate support degree according to the first joint support sum and the first conflict coefficient; the joint support sum calculation formula is as follows:

[0107] S xy (F i ) = m x (F j ) × m y (F j ) + m x (F j ) × m y (θ) + m x (θ) × m y (F j ), xy ∈ i

[0108]

[0109] Wherein, S xy (F i ) is the first joint support sum, m xy (F i ) is the intermediate support degree, and K xy is the first conflict coefficient;

[0110] Secondary fusion calculation step: the intermediate support degree is fused with the remaining evidence bodies, the second joint support sum is calculated according to the joint support sum calculation formula, and the second conflict coefficient is calculated to reflect the contradiction degree among the three evidence bodies; the final confidence is calculated according to the second joint support sum and the second conflict coefficient;

[0111] S xyz (F i )=m xy (F j )×m z (F j )+m xy (F j )×m z (θ)+m xy (θ)×m z (F j )

[0112]

[0113] Wherein, S xyz (F i ) is the second joint support sum, m xyz (F i ) is the final confidence, and K xyz is the second conflict coefficient.

[0114] The health monitoring step specifically comprises the following steps:

[0115] Fault positioning step: after the edge node detects an abnormal signal, the two-end OTDR equipment is triggered to synchronously collect optical signals, and backward scattering curves S1(x) and S2(x) are obtained; wavelet denoising is performed on the curves, loss mutation points x1 and x2 are identified, and loss gradient features are extracted; a loss gradient-physical distance mapping relationship is established by using an SVM regression model, and the calculation formula is:

[0116]

[0117] Wherein, c is the speed of light, t1 and t2 are the signal transmission times of the two ends respectively, and n is the refractive index of the optical fiber.

[0118] Fault classification step: a fault sample library is collected, and a fault classification model is constructed based on a BP neural network structure; the abnormal signal detected by the edge node is input into the trained fault classification model, and the fault type is output.

[0119] The above-mentioned are only embodiments of the present application, and the common knowledge of the specific structure and characteristics in the scheme is not described too much, the ordinary skilled in the art knows all the ordinary technical knowledge in the field of the present application before the application date or the priority date, can know all the prior art in the field, and has the ability to apply the conventional experimental means before the date, the ordinary skilled in the art can improve and implement the present scheme under the inspiration given by the present application, and some typical known structure or known method should not become the obstacle for the ordinary skilled in the art to implement the present application. It should be pointed out that for those skilled in the art, without departing from the structure of the present application, a number of modifications and improvements can be made, which should also be regarded as the protection scope of the present application, which will not affect the effect and practicality of the patent. The protection scope of the present application should be subject to the content of its claims, and the specific implementation mode and the like in the specification can be used to explain the content of the claims.

Claims

1. A health monitoring system for fiber optic networks used in water diversion projects, characterized in that, include: Sensing module: used to synchronously acquire fiber optic environmental parameters through edge computing nodes and a multi-source sensor array; the multi-source sensor array includes strain sensors, vibration sensors, temperature and humidity sensors, and tensile sensors; Transmission network construction module: used to connect adjacent edge computing nodes into a star network via LoRa for short-distance networking; An industrial-grade PON fiber optic network is adopted, enabling edge nodes to access the optical line terminal (OLT) via optical splitters, thus constructing a point-to-multipoint backbone transmission network. Cloud management platform: used to preprocess the data acquired by the transmission network construction module to obtain vibration signals, strain data and temperature and humidity data; The preprocessed data is input into a preset DS evidence theory fusion model to obtain a warning trigger signal for whether a fault warning is triggered. In the DS evidence theory fusion model, vibration signal, strain data and temperature and humidity data are defined as three evidence bodies, and fusion calculation and confidence calculation are performed according to the BPA function of a single evidence body. When the final confidence reaches a set threshold, a warning trigger signal is generated. It is also used to locate and classify faults through a trained fault classification model after an abnormal signal is detected at the edge node, and output the fault location and fault type.

2. A health monitoring system for managing water engineering optical fiber networks according to claim 1, characterized in that: The computational analysis of the DS evidence theory fusion model includes the following modules: Basic Parameter Analysis Module: Used to analyze basic parameters; the basic parameters include: Fault type set: θ = {F1, F1, ..., F8}; Evidence includes: vibration signal E1, strain data E2, and temperature and humidity data E3; Single evidence body BPA function: m i (F j ), represents the support degree of the ith evidence body to the jth fault. Uncertainty: The preliminary fusion calculation module is used to randomly select two pieces of evidence from vibration signal E1, strain data E2, and temperature and humidity data E3, and calculate the first joint support sum according to the joint support sum calculation formula; it also calculates the first conflict coefficient of the two pieces of evidence, reflecting the degree of contradiction between the two pieces of evidence; and calculates the intermediate support based on the first joint support sum and the first conflict coefficient. The joint support sum calculation formula is as follows: S xy (F i )=m x (F j )×m y (F j )+m x (F j )×m y (θ)+m x (θ)×m y (F j ), xy∈i wherein S xy (F i ) is the first joint support sum, m xy (F i ) is the intermediate support, K xy is the first conflict coefficient; The secondary fusion calculation module is used to fuse the intermediate support with the remaining evidence, calculate the second joint support sum according to the joint support sum calculation formula, and calculate the second conflict coefficient to reflect the degree of contradiction among the three pieces of evidence; and calculate the final confidence level based on the second joint support sum and the second conflict coefficient. S xyz (F i ) = m xy (F j ) x m z (F j ) + m xy (F j ) x m z (θ) + m xy (θ) x m z (F j ) where S xyz (F i ) is the second joint support sum, m xyz (F i ) is the final confidence, K xyz is the second conflict coefficient.

3. A health monitoring system for managing an engineered optical fiber network according to claim 2, wherein: The cloud management platform includes a fault location module: The fault location module is used to trigger the synchronous acquisition of optical signals by the two OTDR devices after an abnormal signal is detected at the edge node, and to obtain the backscattering curves S1(x) and S2(x); to perform wavelet denoising on the curves, identify the loss abrupt change points x1 and x2, and extract the loss gradient features; and to establish the loss gradient-physical distance mapping relationship using an SVM regression model, with the calculation formula as follows: Where c is the speed of light, t1 and t2 are the signal transmission times at both ends, and n is the refractive index of the optical fiber.

4. A health monitoring system for managing an engineered optical fiber network according to claim 3, wherein: The cloud management platform includes a fault classification module: The fault classification module is configured to collect a fault sample library and construct a fault classification model based on a BP neural network structure, and input an abnormal signal detected by the edge node into the trained fault classification model to output a fault type.

5. A health monitoring system for managing water engineering optical fiber networks according to claim 1, characterized in that: The edge computing node adopts a lightweight structure design, and comprises a processor, a LoRa / PON dual-communication module architecture, a solar panel and a lithium battery.

6. A health monitoring system for managing water engineering optical fiber networks according to claim 4, characterized in that: The cloud management platform further comprises: The visualization module is configured to load a fault point corresponding to the fault type output by the fault classification module into a GIS map to display a cable state in real time according to the GIS map, and view historical data curves within a set time through the fault point.

7. A method for health monitoring of a water engineering optical fiber network, characterized by: The method comprises the following steps: The sensing step comprises synchronously collecting fiber environment parameters by the edge computing node and a multi-source sensor array; The multi-source sensor array comprises a strain sensor, a vibration sensor, a temperature and humidity sensor and a tension sensor; The transmission network construction step comprises constructing a star network by LoRa between adjacent edge computing nodes for short-distance networking; The industrial-grade PON optical fiber network is adopted to enable the edge node to access an optical line terminal (OLT) through an optical splitter to construct a point-to-multipoint backbone transmission network; The health monitoring step comprises pre-processing data obtained by the transmission network construction module to obtain vibration signals, strain data and temperature and humidity data; The pre-processed data is input into a preset D-S evidence theory fusion model to obtain a pre-warning trigger signal for triggering a fault pre-warning, wherein the vibration signals, the strain data and the temperature and humidity data are defined as three evidence bodies in the D-S evidence theory fusion model, and a BPA function of a single evidence body is used for fusion calculation and confidence calculation, and the pre-warning trigger signal is generated when the final confidence reaches a set threshold. The D-S evidence theory fusion model is further configured to perform fault positioning and classification through the trained fault classification model after detecting an abnormal signal by the edge node, and output a fault position and a fault type.

8. A method for health monitoring of a water engineering optical fiber network according to claim 7, characterized in that: The calculation and analysis of the D-S evidence theory fusion model comprises the following steps: The basic parameter analysis step comprises analyzing basic parameters, wherein the basic parameters comprise: A fault type set: θ={F1,F1,...,F8}; Evidence bodies: vibration signals E1, strain data E2 and temperature and humidity data E3; Single evidence body BPA function: m i (F j ), represents the support degree of the ith evidence body to the jth fault. Uncertainty: The preliminary fusion calculation step comprises selecting two evidence bodies from the vibration signals E1, the strain data E2 and the temperature and humidity data E3, calculating a first joint support sum according to a joint support sum calculation formula, calculating a first conflict coefficient of the two evidence bodies to reflect a contradiction degree between the two evidence bodies, and calculating an intermediate support degree according to the first joint support sum and the first conflict coefficient; and the joint support sum calculation formula is as follows: S xy (F i )=m x (F j )×m y (F j )+m x (F j )×m y (θ)+m x (θ)×m y (F j ), xy∈i wherein S xy (F i ) is the first joint support sum, m xy (F i ) is the intermediate support, K xy is the first conflict coefficient; The secondary fusion calculation step comprises fusing the intermediate support degree with the remaining evidence bodies, calculating a second joint support sum according to the joint support sum calculation formula, calculating a second conflict coefficient to reflect a contradiction degree between the three evidence bodies, and calculating a final confidence degree according to the second joint support sum and the second conflict coefficient. S xyz (F i ) = m xy (F j ) x m z (F j ) + m xy (F j ) x m z (θ) + m xy (θ) x m z (F j ) where S xyz (F i ) is the second joint support sum, m xyz (F i ) is the final confidence, K xyz is the second conflict coefficient.

9. A method for health monitoring of a water engineering optical fiber network according to claim 8, characterized in that: The health monitoring step specifically comprises the following steps: The fault positioning step comprises the following steps: after the edge node detects an abnormal signal, triggering two-end OTDR equipment to synchronously collect optical signals, obtaining backscattering curves S1(x), S2(x); performing wavelet denoising on the curves, identifying loss mutation points x1, x2, and extracting loss gradient features; and establishing a loss gradient-physical distance mapping relationship by using a SVM regression model, and the calculation formula is as follows: Wherein, c is the speed of light, t1 and t2 are the signal transmission times of two ends respectively, and n is the refractive index of the optical fiber.

10. A method for health monitoring of a water engineering optical fiber network according to claim 9, characterized in that: The health monitoring step specifically comprises the following steps: The fault classification step comprises the following steps: collecting a fault sample library, and constructing a fault classification model based on a BP neural network structure; inputting the abnormal signal detected by the edge node into the trained fault classification model, and outputting a fault type.

Images