Water supply pipeline burst and leakage monitoring and early warning method using existing communication optical cable

Abstract

The present application relates to a kind of water supply pipeline burst and leakage monitoring and early warning method using existing communication cable, comprising the following steps: S1, selecting existing communication cable;S2, the sound vibration signal of communication cable is collected by distributed optical fiber acoustic wave sensing equipment, and the correspondence between cable passage number and geographic position coordinates is established;S3, the sound vibration signal of the pipeline along the line to be monitored is collected by distributed optical fiber acoustic wave sensing equipment, and preliminary evaluation index is set to screen;S4, the sound vibration signal at the suspected burst or leakage event pipe section is sample segmented, and the feature vector of each single-channel data after segmentation is extracted;S5, the extracted feature vector is input into pre-trained machine learning model for classification identification;S6, if classification identification result meets early warning strategy, then send early warning.The present application uses existing communication cable, without additional cable laying construction, realizes the combination of preliminary evaluation index and early warning strategy to reduce false alarm rate under the premise of ensuring identification accuracy.

Description

Technical Field

[0001] This invention relates to the field of water supply pipeline safety monitoring, specifically to a method for monitoring and early warning of water supply pipeline bursts and leaks using existing communication optical cables. Background Technology

[0002] Water supply pipelines are an important urban infrastructure and are closely related to daily life. Pipe bursts and leaks not only waste water resources, but may also affect nearby infrastructure and the surrounding environment.

[0003] With technological advancements, current methods for diagnosing leaks in water supply pipelines include negative pressure wave methods, acoustic emission methods, fiber optic sensing, fluid pressure methods, and acoustic wave methods. Among these, distributed fiber optic sensing technology offers the advantage of continuously distributed spatial and temporal measurement information. Furthermore, fiber optics, as sensing elements, possess advantages such as high sensitivity, resistance to electromagnetic interference, light weight, small size, high positioning accuracy, and minimal susceptibility to intrusive media. Distributed Acoustic Sensing (DAS) technology uses acoustic waves as the sensitive signal, identifying abnormal events in the surrounding environment by collecting acoustic vibration signals along the fiber optic axis, demonstrating promising application prospects in pipeline leak monitoring.

[0004] Currently, some work has applied distributed fiber optic acoustic sensing technology to pipeline leak detection: Existing technology, such as patent document CN116805061A, discloses a method for determining leakage events based on fiber optic sensing. This technology is applied to water supply pipelines within pipe racks, utilizing laid sensing optical cables for data acquisition. For non-pipe welded areas, the sensing optical cable is laid parallel to the water supply pipeline on the pipe rack floor; at pipe welded areas, the optical cable is laid on a cable fixing frame below the water collection device, employing a multi-point fixing method. Leakage events are determined using the optical cable sensing signals collected by an optical time domain reflectometer distributed fiber optic sensing system.

[0005] Existing technologies, such as the patent document with publication number CN119934447A, disclose a water pipeline safety monitoring system and monitoring method based on fiber optic vibration and underwater acoustic sensing. The distributed fiber optic acoustic vibration sensing technology uses sensing optical fibers tightly installed on the outer wall of the pipeline to collect vibration data, and finally displays the identification results on the pipeline safety monitoring platform software.

[0006] There is currently some work on using distributed fiber optic acoustic vibration data for leak identification: Existing technologies, such as patent document with publication number CN119123343A, disclose a method for identifying leaks in distributed optical fiber sensing pipelines. This method denoises the time-series signal of DAS and then converts it into a spectrogram, and then identifies leaks using an improved MobileNetV3-large network model.

[0007] Existing technologies, such as patent document CN118940169A, disclose a method and device for pipeline leak detection based on the joint input of distributed optical fiber sensing signals and features into an FC-ANN network. This method involves filtering and preprocessing the acquired optical fiber signals and extracting time-domain and frequency-domain features. The preprocessed optical fiber signals and the extracted features are then combined to construct a joint sequence, which is input into a fully connected artificial neural network model for pipeline leak detection. Both of these methods primarily improve upon distributed optical fiber acoustic vibration data processing methods, but do not specify the sensing system, applicable pipeline types, or optical cables.

[0008] In practical applications, real-time monitoring is mainly achieved by laying separate sensing optical cables combined with distributed fiber optic acoustic sensing technology. However, implementing this method faces many practical difficulties: First, laying sensing optical cables requires large-scale civil engineering construction. Due to the dense underground pipelines, high traffic pressure, strict construction approval, and the fact that many water supply pipelines have been built for many years and the surrounding environment is complex, there is a lack of space for excavation and cable laying. Therefore, in old urban areas or areas with dense pipeline networks, it is almost impossible to lay sensing optical cables alone. This seriously limits the application scenarios of distributed fiber optic acoustic sensing technology.

[0009] Secondly, laying individual optical sensing cables is costly and time-consuming, and can easily disrupt existing traffic and residents' lives, making large-scale application difficult. Furthermore, the monitoring range of optical sensing cables is typically limited by the laying distance and cable length, making it difficult to achieve city-level, long-distance pipeline network coverage.

[0010] In summary, the optical cables required for distributed fiber optic acoustic sensing technology pose a significant challenge in practical applications. Summary of the Invention

[0011] To address the aforementioned technical problems, this invention provides a method for monitoring and early warning of water supply pipeline bursts and leaks using existing communication optical cables. This method utilizes existing communication optical cables to monitor water supply pipeline bursts and leaks, solving the problem of difficult cable laying in practical applications of existing distributed optical fiber sensing technology.

[0012] The technical objective of this invention is achieved through the following technical solution: A method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables, the method comprising: S1. Select an existing communication optical cable in the area along the water supply pipeline to be monitored. The selected communication optical cable is located near the water supply pipeline to be monitored and is in an idle state. S2. Acoustic and vibration signals along the communication optical cable are collected through distributed optical fiber acoustic wave sensing equipment, and the optical cable channel number of the communication optical cable is associated with the geographical coordinates. S3. Collect acoustic and vibration signals along the water supply pipeline to be monitored using distributed optical fiber acoustic wave sensing equipment, establish preliminary evaluation indicators to screen data in the acoustic and vibration signals that are suspected of having pipe bursts or leaks, and if a pipe burst or leak is suspected, proceed to step S4. S4. Perform sample segmentation on the acoustic and vibration signals along the monitored water supply pipeline suspected of having a pipe burst or leak, and extract the feature vector of each single-channel data after segmentation. S5. Input the extracted feature vectors into the pre-trained machine learning model for classification and recognition; S6. If the classification and identification results meet the early warning strategy of time continuity or repetition within a time period, an early warning will be issued.

[0013] Further, in step S1, a geographic information map of the water supply pipeline to be monitored and a geographic information map of the communication optical cables along the water supply pipeline to be monitored are obtained, and an idle communication optical cable near the water supply pipeline to be monitored is selected by comparing the geographic information maps.

[0014] Furthermore, step S2 also includes checking the quality of the sensing signal of the selected communication optical cable. If the conditions are met, the next step can be carried out.

[0015] Furthermore, when establishing preliminary evaluation indicators, the power spectral density is first calculated for the acoustic and vibration signals of each single channel, and then the absolute value of the integral of the power spectral density of the key frequency bands is extracted and used as the preliminary evaluation indicator.

[0016] Furthermore, data with a sampling frequency of 1000Hz or higher for acoustic and vibration signals, and data that meet the threshold conditions for preliminary evaluation indicators, are considered as data indicating suspected pipe rupture or leakage.

[0017] Furthermore, after screening based on preliminary evaluation indicators, the data suspected of pipe bursts or leaks are divided into several single-channel data of equal duration through sample segmentation, and feature vectors are obtained from each single-channel data.

[0018] Furthermore, the features extracted during feature extraction include the first two orders of wavelet scattering network features and linear prediction features.

[0019] Furthermore, in step S6, if the classification and identification results identify a pipe burst or leakage event within a continuous set time period, or if more than a set proportion of data within a set time period are identified as pipe burst or leakage events, an early warning is issued.

[0020] Furthermore, the machine learning model is a machine learning model for classification, including but not limited to extreme gradient boosting tree models, random forest models, support vector machine models, or convolutional neural network models.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention utilizes the idle optical fibers in existing communication optical cables that are widely deployed in cities as the sensing medium, eliminating the need for additional optical cable laying construction. This avoids the impact of civil construction on traffic, residents' lives, and existing underground pipelines, and solves the problem that distributed optical fiber sensing technology cannot be applied in scenarios such as old urban areas and densely networked pipeline areas, which is conducive to expanding the practical application scenarios of the technology.

[0022] 2. This invention relies on existing urban communication optical cables, which are typically laid continuously for tens of kilometers. Monitoring can be carried out using their long, idle optical fibers. Thus, a single monitoring system can achieve continuous, long-distance real-time monitoring over tens of kilometers within a city, significantly improving the coverage compared to existing technologies and effectively solving the problem of short monitoring distance in existing technologies.

[0023] 3. Long-distance, long-term monitoring generates massive amounts of distributed fiber optic acoustic and vibration data. Performing high-precision calculations and judgments on all data would significantly prolong the identification of pipe bursts or leaks, requiring substantial computing power. This invention combines preliminary evaluation indicators with early warning strategies. The preliminary evaluation indicators enable the extraction of suspected leak signals from massive amounts of data, significantly reducing the amount of data required for fine processing. Since pipe bursts or leaks typically have temporal continuity or short-term repetition, while random noise interference does not possess this characteristic, the early warning strategy confirms and filters the initially screened data in the time dimension. This reduces the number of false alarms caused by occasional interference without sacrificing the accuracy of pipe burst or leak identification, thus lowering the false alarm rate. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the hardware connection in the monitoring method of the present invention.

[0025] Figure 2 This is a schematic diagram of the process of the water supply pipeline burst and leakage monitoring and early warning method using communication optical cables according to the present invention. Detailed Implementation

[0026] The technical solution of the present invention will be further described below with reference to specific embodiments: A method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables, such as... Figure 1 As shown, at the hardware level, selected communication optical cables are connected to a distributed fiber optic acoustic wave sensing device. This device is then connected to an industrial control computer via a network. The industrial control computer processes the acquired data, including preliminary evaluation of the data using initial screening indicators, sample segmentation, feature extraction, inputting the data into a machine learning model for classification and judgment, determining early warning strategies, and issuing early warnings. Preferably, the industrial control computer can also be connected to a cloud management platform to achieve cloud-based sharing and storage of early warnings.

[0027] like Figure 2 As shown, the method includes: S1. Obtain the geographic information map of the water supply pipeline to be monitored and the geographic information map of the communication optical cable in the area where the water supply pipeline is located. Select the communication optical cable that is located near the water supply pipeline to be monitored and is idle by comparing the geographic information maps, such as the one located on the same side of the road and with the same direction.

[0028] S2. Acoustic and vibration signals along the communication optical cable are collected through distributed optical fiber acoustic wave sensing equipment, and the optical cable channel number of the communication optical cable is associated with the geographical coordinates. Preferably, the method also includes collecting the acoustic and vibration signals generated by the selected communication optical cable and evaluating the attenuation coefficient, strain resolution, spatial resolution, correlation coefficient, and other indicators of the collected sensor signals. For example, if the attenuation coefficient < 1, strain resolution < 20 nε, spatial resolution > 10 m, and correlation coefficient > 0.4, it indicates that the sensor signal quality of the communication optical cable meets the requirements, and if it meets the requirements, the communication optical cable can be used for subsequent monitoring.

[0029] S3. Collect acoustic and vibration signals along the water supply pipeline to be monitored using distributed optical fiber acoustic wave sensing equipment. Set preliminary evaluation indicators to screen data in the vibration signals that are suspected of having pipe bursts or leaks. If a pipe burst or leak is suspected, proceed to step S4 to reduce the amount of data through preliminary evaluation indicators. More specifically, when establishing preliminary evaluation indicators, the power spectral density of the single-channel time-domain data of the acoustic vibration signal is continuously calculated for each time period according to the set time period, and the absolute value of the integral of the power spectral density curve of the key focus frequency band is selected as the preliminary evaluation indicator.

[0030] In a specific implementation, taking a 3.6km long communication optical cable as an example, monitoring is carried out along the water supply pipeline along the optical cable. Environmental noise and burst or leakage signals are collected by distributed fiber optic acoustic sensing equipment. The sampling frequency is greater than or equal to 1000Hz, which can meet the requirements of collecting signals covering the frequency band below 500Hz. In a specific implementation, the sampling frequency is preferably 2000Hz, and the spatial sampling interval and gauge length are 4m. The burst event of the water supply pipeline is simulated by opening fire hydrants.

[0031] In a specific implementation process, the power spectral density of single-channel time-domain data is calculated every 60 seconds, and the absolute value of the integral of the power spectral density in the concentrated frequency band of the leakage signal is extracted as a preliminary evaluation index. For example, the threshold of the evaluation index is set to 1.2. When the preliminary evaluation index is <1.2, the data is considered as data suspected of pipe bursting or leakage, thus achieving preliminary data screening; when the preliminary evaluation index is ≥1.2, it is considered that there is no pipe bursting or leakage, and no further judgment is required.

[0032] S4. Segment the acoustic and vibration signals suspected of pipe bursting or leakage, and extract the feature vector of each single-channel data after segmentation. During segmentation, the data after initial screening in step S3 is divided into 10s single-channel data. Feature extraction is performed on each 10s single-channel data after segmentation. The extracted features include the first two-order features of the wavelet scattering network and the linear prediction features. By combining the first two-order features of the wavelet scattering network with the linear prediction features, a stable, robust, and interpretable feature vector can be formed.

[0033] S5. The extracted feature vectors are input into a pre-trained machine learning model for classification and recognition. This machine learning model includes, but is not limited to, extreme gradient boosting tree models, random forest models, support vector models, or convolutional neural network models. Taking the extreme gradient boosting tree model as an example, it is pre-trained using field test data and then transferred to the scenario of this application to achieve target domain transfer learning. The training dataset is obtained by collecting data from simulated pipe burst events in test pipe sections and then segmenting the samples. The ambient noise data consists of all channel samples, while the burst or leakage event data consists of five channels before and after the burst location. Sample weights are set according to the number of samples during training to balance the impact of differences in sample quantity.

[0034] S6. If the classification and identification results meet the early warning strategy of time continuity or repetition within a time period, an early warning will be issued.

[0035] More specifically, if the classification results identify a leak or pipe burst event within a continuous set time period, or if more than a set proportion of data within a set time period are identified as leaks or pipe burst events, an early warning will be issued.

[0036] In a specific implementation process, a time period of 60 seconds is set. If data of the category of pipe burst or leakage event appears in two consecutive time periods, it indicates that there is a pipe burst or leakage, and an early warning is issued. In another specific implementation process, a 60-second time period is set. If more than 50% of the data within a time period is identified as a pipe burst or leak, it indicates that a pipe burst or leak has occurred and an early warning needs to be issued.

[0037] This embodiment is merely a further explanation of the present invention and is not intended to limit the present invention. Those skilled in the art can make non-inventive modifications to this embodiment as needed after reading this specification, but such modifications are protected by patent law as long as they are within the scope of the claims of the present invention.

Claims

1. A method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables, characterized in that, The method includes: S1. Select an existing communication optical cable along the water supply pipeline to be monitored. The selected communication optical cable is located near the water supply pipeline to be monitored and is in an idle state. S2. Acoustic and vibration signals along the communication optical cable are collected through distributed optical fiber acoustic wave sensing equipment, and the optical cable channel number of the communication optical cable is associated with the geographical coordinates. S3. Collect acoustic and vibration signals along the water supply pipeline to be monitored using distributed optical fiber acoustic wave sensing equipment, establish preliminary evaluation indicators to screen pipe sections suspected of having pipe bursts or leaks in the acoustic and vibration signals, and if a pipe burst or leak is suspected, proceed to step S4. S4. Perform sample segmentation on the acoustic and vibration signals along the monitored water supply pipeline suspected of having a pipe burst or leak, and extract the feature vector of each single-channel data after segmentation. S5. Input the extracted feature vectors into the pre-trained machine learning model for classification and recognition; S6. If the classification and identification results meet the early warning strategy of time continuity or repetition within a time period, an early warning will be issued.

2. The method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables according to claim 1, wherein in step S1, a geographic information map of the water supply pipeline to be monitored and a geographic information map of the communication optical cables along the water supply pipeline to be monitored are obtained, and an idle communication optical cable located near the water supply pipeline to be monitored is selected by comparing the geographic information maps.

3. The method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables according to claim 2, characterized in that, Step S2 also includes checking the quality of the sensing signal of the selected communication optical cable. If the conditions are met, the next step can be carried out.

4. The method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables according to claim 1, characterized in that, When establishing preliminary evaluation indicators, the power spectral density is first calculated for the acoustic and vibration signals of each single channel. Then, the absolute value of the integral of the power spectral density of the key frequency bands is extracted and used as the preliminary evaluation indicator.

5. A method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables, as described in claim 4, is characterized in that... Data with a sampling frequency of 1000Hz or higher and whose preliminary evaluation indicators meet the threshold conditions are considered as data indicating suspected pipe rupture or leakage.

6. The method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables according to claim 5, characterized in that, After screening based on preliminary evaluation indicators, the data suspected of pipe bursts or leaks are divided into several single-channel data of equal duration through sample segmentation. Feature vectors are then extracted from each single-channel data.

7. A method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables, as described in claim 6, is characterized in that... The features extracted during feature extraction include the first two orders of wavelet scattering network features and linear prediction features.

8. A method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables, as described in claim 1, is characterized in that... In step S6, if the classification and identification results identify a pipe burst or leakage event within a continuous set time period, or if more than a set proportion of data within a set time period are identified as pipe burst or leakage events, an early warning is issued.

9. A method for monitoring and early warning of pipe bursts and leaks in water supply pipelines using existing communication optical cables, as described in claim 1, is characterized in that... The machine learning model is a classification machine learning model, including extreme gradient boosting tree model, random forest model, support vector machine model or convolutional neural network model.

Images