Water supply pipeline leakage noise monitoring and early warning method and system based on correlation principle

Abstract

The invention discloses a water supply pipeline leakage noise monitoring and early warning method and system based on related principles, a plurality of audition monitoring devices are arranged at a plurality of preset nodes of a water supply network, and each audition monitoring device comprises a data acquisition module, a wireless communication module and a power supply module; a noise signal of a water supply pipeline is acquired in real time through a data acquisition module, and the acquired noise signal is transmitted to a data processing center through a wireless communication module; the data processing center preprocesses the received noise signals collected by the audition monitoring devices to obtain preprocessed noise signals and time-frequency images; based on a correlation principle, calculating a cross correlation coefficient between the preprocessed noise signals acquired by different listening monitoring devices to obtain a cross correlation matrix; and according to a preset leakage noise template, matching with the cross-correlation matrix, judging whether leakage exists or not, and estimating a leakage position. The detection efficiency is improved, the water supply pipe network is monitored in real time, and the accuracy of leakage judgment is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of leakage detection, and in particular to a water supply pipeline leakage noise monitoring and early warning method and system based on relevant principles. Background Art

[0002] With the acceleration of urbanization, the construction and management of water supply pipe networks are becoming increasingly important. Pipeline leakage not only affects water supply efficiency, but may also lead to water resource waste and environmental pollution. Therefore, it is particularly important to detect and locate leakage points of water supply pipes in a timely and accurate manner.

[0003] Traditional leakage detection methods usually rely on workers holding listening rods to conduct inspections one by one, that is, manually inspecting along the pipeline and judging whether there is leakage by listening to the leakage noise. Although this method is effective to a certain extent, it has obvious shortcomings.

[0004] First, this manual detection method is inefficient. Since urban pipelines are usually widely distributed and cover a large area, the use of handheld listening rods requires a lot of manpower and time, especially when the pipeline length is several kilometers or even longer. A complete detection takes several days or even longer. In addition, the leakage point is usually not easy to identify, and workers need to listen carefully in various harsh environments, which increases the difficulty of operation.

[0005] Secondly, this method relies on the experience of workers, and the detection results are highly subjective. Different workers have different auditory sensitivity, experience, and skill levels, which often lead to inconsistency and inaccuracy in leakage detection results. Especially in noisy urban environments, background noise can easily interfere with judgment, increasing the probability of missed detection and false alarms.

[0006] Furthermore, manual inspection is difficult to achieve all-weather monitoring, and the inspection time can only be selected during the late night period, because less water is used during this period and the pressure is high. Leakage noise detection can only be carried out in a quiet environment without external noise. The operating time is not conducive to the workers' work and rest.

[0007] Therefore, the manual detection method in the prior art has many shortcomings such as low efficiency, low accuracy, poor real-time performance and poor safety. Summary of the invention

[0008] In order to overcome the shortcomings of the prior art, the purpose of this application is to provide a water supply pipeline leakage noise monitoring and early warning method and system based on relevant principles, which is mainly used to solve the defects of the manual detection method in the prior art such as low efficiency, low accuracy, poor real-time performance and poor safety.

[0009] To solve the above problems, the technical solutions adopted in this application are as follows:

[0010] In a first aspect of the present application, a water supply pipeline leakage noise monitoring and early warning method based on relevant principles is provided, comprising the following steps:

[0011] S10, arranging a plurality of listening monitoring devices at a plurality of preset nodes of the water supply network, each of the listening monitoring devices comprising a data acquisition module, a wireless communication module and a power supply module;

[0012] S20, collecting noise signals of the water supply pipeline in real time through the data acquisition module, and transmitting the collected noise signals to the data processing center through the wireless communication module;

[0013] S30, at the data processing center, preprocessing the noise signals collected by each listening monitoring device to obtain the preprocessed noise signals and time-frequency images;

[0014] S40, based on the correlation principle, calculating the mutual correlation coefficients between the pre-processed noise signals collected by different listening monitoring devices to obtain a mutual correlation matrix;

[0015] S50, matching the preset leakage noise template with the cross-correlation matrix to determine whether there is leakage and estimate the leakage position.

[0016] The second aspect of the present application provides a system for the water supply pipeline leakage noise monitoring and early warning method based on the above-mentioned related principles, including:

[0017] A plurality of listening monitoring devices, each of which is configured to be arranged near a preset node of a water supply network, and each of which includes a data acquisition module, a wireless communication module and a power supply module. The data acquisition module is used to collect noise signals of the water supply pipeline in real time, the wireless communication module is used to transmit the collected noise signals to a data processing center, and the power supply module is used to supply power to the data acquisition module and the wireless communication module.

[0018] The data processing center is used to pre-process the noise signals collected by each listening monitoring device to obtain the pre-processed noise signals and time-frequency images;

[0019] The correlation principle analysis module is used to calculate the mutual correlation coefficients between the pre-processed noise signals collected by different listening monitoring devices based on the correlation principle, obtain the mutual correlation matrix, match it with the mutual correlation matrix according to the preset leakage noise template, determine whether there is leakage, and estimate the leakage position.

[0020] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0021] 1. By deploying multiple listening monitoring devices, comprehensive coverage of specific areas of the water supply network can be achieved, and a large number of noise signals can be collected in a short time, which significantly improves the efficiency of leakage detection, reduces the time and labor costs required for manual inspections, improves detection efficiency, and can achieve real-time monitoring of the water supply network, obtain pipeline status information in a timely manner, and quickly respond to potential leakage risks;

[0022] 2. Using the data collected by multiple monitoring devices for cross-correlation analysis, combined with the preset leakage noise template, can effectively improve the accuracy of leakage judgment, reduce the probability of missed detection and false alarm, ensure that leakage problems can be discovered and located in time, and enhance detection accuracy;

[0023] 3. Through automated data collection and analysis, the reliance on manual inspection is reduced, and the inspection errors caused by human factors are reduced; at the same time, the monitoring equipment can continue to work under various environmental conditions, which improves the reliability of inspection and reduces reliance on manual labor.

[0024] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A schematic diagram of a flow chart of a water supply pipeline leakage noise monitoring and early warning method based on relevant principles provided in Example 1;

[0026] Figure 2 A flow chart of a water supply pipeline leakage noise monitoring and early warning method based on relevant principles provided in Example 2.

[0027] Figure 3 A flow chart of a water supply pipeline leakage noise monitoring and early warning method based on relevant principles provided in Example 3.

[0028] Figure 4 A flow chart of a water supply pipeline leakage noise monitoring and early warning method based on relevant principles provided in Example 4.

[0029] Figure 5 A schematic diagram of the framework of a water supply pipeline leakage noise monitoring and early warning system based on relevant principles provided in Example 5.

[0030] Figure 6 A schematic diagram of the framework of a water supply pipeline leakage noise monitoring and early warning system based on relevant principles provided in Example 6. DETAILED DESCRIPTION

[0031] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them.

[0032] Embodiment 1:

[0033] like Figure 1 As shown, this embodiment 1 provides a water supply pipeline leakage noise monitoring and early warning method based on relevant principles, including the following steps:

[0034] S10, arranging a plurality of listening monitoring devices at a plurality of preset nodes of the water supply network, each listening monitoring device comprising a data acquisition module, a wireless communication module and a power supply module; the power supply module provides power to each module;

[0035] Since water supply pipelines are generally buried underground and are intricate, maintenance units need to regularly inspect and maintain various areas. Therefore, certain preset nodes will be formed based on the maintenance tasks. For example, this time it is for a certain industrial area, and next time it will be for a certain residential area. Therefore, this preset node is generated based on actual maintenance needs. After determining the preset node based on experience of where leakage is prone to occur, multiple listening monitoring devices are arranged around the preset node to enable it to automatically collect noise signals at potential leakage points.

[0036] S20, collecting noise signals of the water supply pipeline in real time through the data acquisition module, and transmitting the collected noise signals to the data processing center through the wireless communication module;

[0037] The data acquisition module is responsible for real-time acquisition of noise signals in the pipeline at a certain sampling rate. The acquired signals reflect the dynamics of water flow and any possible abnormalities, such as leakage. These signals can effectively reduce the interference of environmental noise after being filtered. The wireless communication module transmits the acquired data to the data processing center in real time, ensuring the timeliness and integrity of the data.

[0038] S30, at the data processing center, preprocessing the noise signals collected by each listening monitoring device to obtain the preprocessed noise signals and time-frequency images;

[0039] Preprocessing can remove background noise and unnecessary interference in the signal and enhance the useful components of the signal. Through time-frequency transformation (such as short-time Fourier transform), the time and frequency information of the signal can be separated, which is helpful for feature extraction and pattern recognition in subsequent analysis.

[0040] S40, based on the correlation principle, calculating the mutual correlation coefficients between the pre-processed noise signals collected by different listening monitoring devices to obtain a mutual correlation matrix;

[0041] By calculating the cross-correlation coefficients of signals between different devices, the correlation between signals can be evaluated. This process can reveal the similarity of noise signals collected by sensors at different locations. Especially when there is a leakage, the noise characteristics generated by the leakage often show obvious correlation in the monitoring signals at different locations. The resulting cross-correlation matrix provides a quantitative basis for determining leakage.

[0042] S50, matching the preset leakage noise template with the cross-correlation matrix to determine whether there is leakage and estimate the leakage position;

[0043] By matching with the preset leakage noise template, specific patterns in the data can be identified, which correspond to the known leakage noise characteristics. The matching process uses the correlation data provided by the cross-correlation matrix to improve the accuracy of the judgment. Once the leakage is identified, the specific location of the leakage can be estimated in combination with the location data, providing a scientific basis for subsequent maintenance and repair work.

[0044] Embodiment 2:

[0045] like Figure 2 As shown, this embodiment 2 provides a water supply pipeline leakage noise monitoring and early warning method based on relevant principles, including the following steps:

[0046] S100, arranging a plurality of listening monitoring devices at a plurality of preset nodes of the water supply network, each listening monitoring device comprising a data compression module, an acceleration sensor array, a data acquisition module, a wireless communication module and a power supply module; the acceleration sensor array comprises a plurality of acceleration sensors distributed in a circular array for collecting vibration signals in different directions; the power supply module provides power to each module;

[0047] It should be noted that the data acquisition module is a sound sensor, which is used to collect the noise generated after water flows through the water pipe. After water flows through the water pipe, there are two states: leakage and no leakage. However, the acceleration sensor is used to collect the vibration signal generated by the water pipe itself when water flows through, which reflects the vibration of the water pipe in the two states of leakage and no leakage.

[0048] The circular layout of the acceleration sensor array can collect vibration signals in multiple directions at the same time. This design enables the sensor to capture vibration information from different directions, especially when a pipeline leaks. The vibration signal may propagate along the pipeline and produce different characteristics at different locations. Phase difference analysis can be used to determine the direction and location of the vibration source; the circular array can more comprehensively cover the environment around the pipeline and reduce blind spots.

[0049] S200, collecting noise signals of the water supply pipeline in real time through the data acquisition module, collecting vibration signals of the water supply pipeline in real time through the acceleration sensor array, and adaptively compressing the collected noise signals and vibration signals through the data compression module, the compression ratio is automatically adjusted according to the signal complexity and the wireless communication bandwidth, and then the compressed noise signals and vibration signals are transmitted to the data processing center through the wireless communication module;

[0050] It should be noted that the data compression module performs time-frequency analysis on the collected noise signal and vibration signal, and optionally performs short-time Fourier transform (STFT); calculates the entropy value or spectral density of the signal as a quantitative indicator of complexity. The higher the complexity, the greater the amount of information contained in the signal, and a lower compression ratio is required to retain important information; monitors the currently available wireless communication bandwidth in real time, and estimates the current transmission rate and delay using network detection technology. The lower the bandwidth, the higher the compression ratio is required to ensure that the data can be transmitted in a timely manner; uses an audio compression algorithm (such as MP3) for noise signals and a wavelet compression algorithm for vibration signals; this is in consideration of real-time leakage data collection underground, and the need to solve the problems of weak signals and large data volumes. Otherwise, under the conditions of long-term and large data volume detection, data loss may easily occur during periods of poor signals, affecting the accuracy of monitoring and early warning.

[0051] S300, in the data processing center, adaptively decompress and preprocess the noise signals and vibration signals collected and compressed by various listening monitoring devices, the preprocessing including adaptive filtering, wavelet noise reduction, normalization and time-frequency transformation operations to remove noise and interference, wherein the normalization and time-frequency transformation can transform the original signal into a form more suitable for analysis, so that important features can be highlighted, and the preprocessed noise signal, vibration signal and time-frequency image are obtained; the preprocessing enhances the effective information in the signal and improves the accuracy of subsequent analysis, and the abnormal change pattern in the signal, such as the characteristic frequency of leakage, can be more intuitively observed through the time-frequency image;

[0052] S400, based on the correlation principle, calculating the mutual correlation coefficients and mutual correlation functions between the pre-processed noise signals collected by different listening monitoring devices to obtain a mutual correlation matrix and a mutual correlation function matrix; at the same time, calculating the autocorrelation coefficients and autocorrelation functions of multiple vibration signals collected by the acceleration sensor array inside each listening monitoring device to obtain an autocorrelation matrix and an autocorrelation function matrix;

[0053] It should be noted that the analysis based on the correlation principle can effectively extract the relationship between the signals collected by different sensors:

[0054] 1. Calculation of cross-correlation coefficient and cross-correlation function

[0055] 1.1 Cross-correlation function R xy (τ) describes the similarity between two signals x(t) and y(t) and is defined as:

[0056]

[0057] where τ is the time delay.

[0058] 1.2 Calculation of the mutual correlation coefficient

[0059] Mutual correlation coefficient C xy It can be obtained by normalizing the cross-correlation function, which is defined as:

[0060]

[0061] Where R xx (0) and R yy (0) is the value of the autocorrelation function of signals x and y at zero delay.

[0062] 1.3 Calculation steps

[0063] Signal pair: select a pair of noise signals x i (t) and x j (t), where i and j are different listening monitoring devices;

[0064] Calculate the cross-correlation function: Use the above cross-correlation function definition to calculate R xy (τ);

[0065] Calculate the mutual correlation coefficient: Calculate the mutual correlation coefficient C according to the normalized formula xy ;

[0066] Constructing the cross-correlation matrix: Perform the above calculations on the noise signals of all devices to form a cross-correlation matrix C, where C[i][j] represents the cross-correlation coefficient between device i and device j.

[0067] 2. Calculation of autocorrelation coefficient and autocorrelation function

[0068] 2.1 Definition of autocorrelation function

[0069] Autocorrelation function R xx (τ) describes the similarity of the signal x(t) to itself and is defined as:

[0070]

[0071] 2.2 Calculation of autocorrelation coefficient

[0072] Autocorrelation coefficient C xx It can be obtained by normalizing the autocorrelation function, which is usually defined as:

[0073]

[0074] 2.3 Calculation steps

[0075] Signal Pairs: Select multiple vibration signals within each device k (t), where k represents different acceleration sensors.

[0076] Calculate the autocorrelation function: Use the autocorrelation function to define and calculate R yy (τ).

[0077] Calculate the autocorrelation coefficient: Calculate the autocorrelation coefficient C according to the normalization formula yy .

[0078] Constructing the autocorrelation matrix: Perform the above calculations on the vibration signals of all acceleration sensors to form an autocorrelation matrix C, where C[k][l] represents the autocorrelation coefficient between sensor k and sensor l.

[0079] Among them, the cross-correlation matrix reflects the similarity of noise signals between different listening monitoring devices, and the autocorrelation matrix reflects the similarity between vibration signals collected by acceleration sensors inside the same device. These matrices can be used for subsequent data analysis, feature extraction, and pattern recognition to help determine pipeline status, locate water leaks, etc.

[0080] S500. Based on the preset leakage noise vibration template, multi-feature fusion matching is performed with the cross-correlation matrix, cross-correlation function matrix, autocorrelation matrix and autocorrelation function matrix to determine whether there is leakage, and the leakage location, leakage degree and leakage type are estimated by combining the cross-correlation coefficient, cross-correlation function, autocorrelation coefficient and autocorrelation function.

[0081] Among them, the leakage noise vibration template includes noise feature template, vibration feature template and cross-correlation feature template. The main parameters of the noise feature template include spectrum feature, energy distribution and time domain feature. The main parameters of the vibration feature template include acceleration amplitude, frequency feature and vibration intensity. The main parameters of the cross-correlation feature template include correlation coefficient distribution, correlation function waveform and time delay feature. Then, the feature distance algorithm of Euclidean distance is used to match the real-time collected cross-correlation matrix, cross-correlation function matrix, autocorrelation matrix and autocorrelation function matrix with the preset template for multi-dimensional feature matching. The presence of leakage is judged according to the matching distance and confidence threshold. Based on the time difference, amplitude change and spectrum feature of the cross-correlation coefficient, cross-correlation function, autocorrelation coefficient and autocorrelation function, the triangulation positioning, machine learning classification and feature scoring techniques are comprehensively used to accurately estimate the location, severity and type of leakage. This method can achieve high-precision leakage detection and evaluation by comprehensively analyzing multi-dimensional features such as noise, vibration and cross-correlation.

[0082] As an implementation mode, in step S10 or S100, for a three-way pipe or a complex node with multiple branch pipes, at least one listening monitoring device is arranged on each branch to form a triangular array;

[0083] In step S40 or S400, in addition to calculating the cross-correlation coefficients and cross-correlation functions between different listening monitoring devices, the cross-power spectrum density and phase difference between two of the listening monitoring devices on the multiple branches are also calculated;

[0084] In step S50 or S500, the main frequency component of the leakage noise is determined by the peak frequency of the cross-power spectrum density, the propagation direction and flow characteristics of the leakage noise in the three-way pipe are determined by the sign and size of the phase difference, and the positioning information of the leakage position is quantified in combination with the water supply network topology and hydraulic model.

[0085] It should be noted that by forming a triangular array monitoring network at the nodes of a complex pipe network, and using the cross-correlation, cross-power spectrum and phase difference analysis of multi-sensor signals, the propagation characteristics of leakage noise between different branches can be fully captured. Its working principle is based on the propagation law of sound waves in the pipeline. The main frequency components of leakage noise are identified by the cross-power spectrum density, and the noise propagation direction is inferred by the phase difference. Combined with the pipe network topology and hydraulic model constraints, accurate three-dimensional positioning of the leakage position is achieved. This multi-dimensional information fusion method significantly improves the accuracy and reliability of leakage detection in complex pipe network environments, while enhancing the system's anti-interference ability and adaptability.

[0086] Further, in step S10 or S100, at least two listening monitoring devices are arranged on each branch of the three-way pipe, the number of listening monitoring devices arranged on each branch increases in proportion to the length and diameter of the branch, and the distance between adjacent listening monitoring devices does not exceed 10 meters; another listening monitoring device is arranged at the connection of the three-way pipe to form a triangular monitoring array; wherein the configuration formula for the number of listening monitoring devices arranged on each branch is:

[0087] N=max(2,[L / (10×D)])

[0088] Where N is the number of devices, L is the length of the branch pipeline (m), and D is the pipe diameter (m).

[0089] As the larger the pipe diameter, the longer the leakage noise propagation distance, the smaller the required equipment density. For pipe diameters below DN200, one is arranged every 10 meters; for pipe diameters of DN200-400, one is arranged every 15 meters; for pipe diameters of DN400-600, one is arranged every 20 meters; for pipe diameters above DN600, one is arranged every 25 meters.

[0090] For example, for a DN300 tee pipe:

[0091] The main pipe is 30 meters long, with three devices arranged every 15 meters;

[0092] The two branch pipes are 20 meters long each, with two devices arranged every 15 meters;

[0093] One device is arranged at the three-way connection, and a total of 3+2+2+1=8 devices form a monitoring array.

[0094] This arrangement ensures that adjacent devices can receive leakage signals of sufficient strength, reduces the impact of environmental noise interference, reduces positioning deviations caused by multipath effects, improves the accuracy of leakage point azimuth judgment, and achieves reasonable control of monitoring costs.

[0095] As an implementation method, the listening monitoring devices on the same branch are strongly correlated and combined, and the mutual correlation coefficient and the mutual correlation function between them are calculated. The propagation direction and speed of the leakage noise on the branch are determined by the extreme value of the mutual correlation coefficient and the time delay corresponding to the extreme value;

[0096] The listening monitoring devices on different branches are strongly correlated and combined, and the cross-power spectrum density and phase difference between them are calculated. The main propagation path of the leakage noise in the three-way pipe is determined by the peak frequency of the cross-power spectrum density, and the transmission direction and attenuation law of the leakage noise between adjacent branches are determined by the sign and size of the phase difference.

[0097] Combined with the pipe material and wall thickness parameters, the distance and azimuth between the leakage point and the center of the tee are estimated. Combined with the water supply network topology and hydraulic model, the location information of the leakage position is quantified.

[0098] Among them, the strong correlation combination is to combine the listening monitoring devices with highly similar signal characteristics on the same branch or different branches. Because there are multiple listening monitoring devices on the same branch or different branches, each listening monitoring device is far away from the leakage point. When a leakage point appears, the distance or response of each listening monitoring device to it will be different. At this time, the two with closer distance and highly similar signal characteristics are strongly correlated and combined for the next step of analysis. This can eliminate the influence of signal noise and interference; optionally, the number of listening monitoring devices in the combination is two;

[0099] The extreme values ​​of the cross-correlation coefficient indicate the maximum similarity of the signals, while the time delays corresponding to the extreme values ​​reflect the time difference in the signal propagation between the two devices. By analyzing these time delays, and knowing the fixed distance between the devices, the system can calculate the propagation speed of the leakage noise within that branch. In addition, the sign of the time delay (positive or negative) helps determine the propagation direction - that is, whether the noise is moving toward or away from a specific sensor - thereby accurately locating the location of the leak along the branch;

[0100] In order to determine the main propagation path of the leakage noise in the three-way pipeline, the present embodiment involves calculating the cross-power spectrum density between the monitoring devices located on different branches. The peak frequency identified in the cross-power spectrum density represents the main frequency component of the leakage noise, which is related to the main path of the noise propagation in the pipeline network. In addition, by checking the sign and magnitude of the phase difference between the signals from different branches, the system can determine the transmission direction of the leakage noise between adjacent branches. The positive or negative phase difference indicates the direction of energy flow, while the magnitude reflects the degree of attenuation of the signal when it propagates in the pipeline;

[0101] Combining cross-correlation and cross-power spectral density analysis, the system estimates the distance and azimuth of the leak relative to the center of the tee. By utilizing known parameters such as pipe material properties and wall thickness, as well as calculated propagation speed and direction, this embodiment applies triangulation to determine the exact location of the leak. Time delay and phase difference provide spatial information, which when combined with the topology of the pipeline can calculate the coordinates of the leak. The azimuth is derived from the relative position of the sensors and the inferred propagation direction, providing an accurate estimate of the distance and direction between the leak and the tee.

[0102] As an implementation manner, in the leakage judgment process of step S50, the following is further included:

[0103] The water supply network is modeled in different regions, and the water supply network is divided into several monitoring zones. Each zone contains multiple listening monitoring devices to form a local monitoring network. The entire water supply network is divided into several monitoring zones according to factors such as geographical location, pipeline structure and hydraulic characteristics. Through regional modeling, the correlation of signals in the local area is fully utilized to improve the accuracy of leakage detection.

[0104] In each monitoring zone, the Gaussian Process Regression (GPR) method is used to combine the mutual correlation coefficient and autocorrelation coefficient to construct the spatial interpolation model and propagation inversion model of leakage noise, and estimate the leakage noise intensity distribution at any position in the zone; the collected noise signal is preprocessed to extract characteristic parameters such as the mutual correlation coefficient and autocorrelation coefficient, and the GPR model is trained using the monitoring equipment data at known locations to learn the spatial distribution law of leakage noise intensity. After the model is trained, the spatial interpolation model and propagation inversion model of leakage noise are constructed to estimate the leakage noise intensity at any position in the zone; Gaussian process regression is applied to the spatial estimation of leakage noise intensity, making full use of the spatial correlation between signals and improving the prediction accuracy of the leakage location;

[0105] The sound line method is used to simulate the propagation path and attenuation law of leakage noise in the pipeline network, and the spatial coordinates and source intensity of the leakage point are inverted by the least squares method. According to the material, diameter, wall thickness and internal medium parameters of the pipeline, a sound wave propagation model in the pipeline is established, and the sound line method is used to simulate the propagation path of the sound wave from the leakage point to each monitoring device, including reflection, refraction and other phenomena. Considering the attenuation of sound waves due to factors such as distance, medium absorption and pipeline damping during the propagation process, the attenuation on each path is calculated, and then the least squares method is used to combine the simulated propagation model with the actual measurement data to invert the location and source intensity of the leakage point. Combining the sound line method with the least squares inversion can more accurately simulate the propagation characteristics of leakage noise in a complex pipeline network and realize the positioning of the leakage point.

[0106] The positioning results of each partition are summarized and globally optimized to obtain the leakage distribution map and risk level map of the entire water supply network; the coordinate system and data format of each partition are unified, all positioning results are integrated, and the weighted least squares method is used to optimize the overall data to eliminate possible errors and contradictions. According to factors such as leakage location, source intensity and network importance, risk levels are formulated and risk level maps are generated. Through global optimization, a comprehensive assessment of the leakage situation of the entire water supply network is achieved, and a global leakage distribution map and risk level are provided, providing a scientific basis for the maintenance and management of the water supply network.

[0107] Embodiment 3:

[0108] Combination Figure 3 In this embodiment 3, a water supply pipeline leakage noise monitoring and early warning method based on relevant principles is provided. On the basis of embodiments 1 and 2, it is considered that the short-distance area in the water supply network, that is, the pipe section close to the road, is easily disturbed by the vibration noise generated by the passing of vehicles. These vehicle noises will affect the accuracy of leakage detection and need to be specially processed and eliminated. Specifically, it includes:

[0109] Identify the short-cut area in the water supply network, and arrange a linear monitoring array consisting of multiple vibration sensors in the short-cut area. The number of vibration sensors is not less than 3, and the arrangement interval is 5-10 meters. Multiple vibration sensors use wireless signal self-organizing network; multiple vibration sensors are arranged linearly to capture the propagation characteristics of vibration waves generated by vehicle movement. The spacing of 5-10 meters can ensure sufficient spatial resolution. No less than 3 sensors can realize accurate estimation of vehicle motion parameters;

[0110] Noise collection and feature extraction are performed on typical vehicle models in the near-road area to build a vehicle noise template library. The template parameters include vehicle model, speed, and road condition, which contains noise characteristics under different vehicle models, speeds, and road conditions. Through cluster analysis and statistical modeling, the spectrum distribution law of vehicle noise under different working conditions is characterized and the spectrum distribution law of vehicle noise is mastered;

[0111] According to the real-time collected vehicle vibration signal, the vehicle position, speed and acceleration parameters are estimated, and then combined with the vehicle noise template, a reference noise signal matching the current working condition is adaptively generated as the input of the adaptive filter, the filter weight is updated in real time, the notch frequency and order are dynamically adjusted, and the vehicle interference noise in the noise signal and time-frequency image is eliminated;

[0112] It should be noted that the vibration signals collected by the linear monitoring array are used for time delay estimation and wavefront analysis, the Doppler effect principle is applied to calculate the vehicle speed, the acceleration is estimated by the speed change in the continuous time window, and the relative position of the vehicle is estimated by the triangulation positioning method combined with the array geometric configuration;

[0113] The noise template library is retrieved based on the estimated vehicle parameters (position, speed, acceleration), the most similar noise template is selected using a fuzzy matching algorithm, and the selected template is interpolated and scaled to match the current operating conditions;

[0114] Based on the matched noise template, the vehicle motion parameters are considered, and the acoustic propagation model is applied to simulate the process of noise propagating from the vehicle to the monitoring point. The Doppler effect, attenuation and other factors are superimposed to generate a time-varying reference noise signal.

[0115] The generated reference noise signal is used as the filter input, and the actually collected signal is used as the expected output. The filter weight coefficient is updated through iterative optimization, and the filtered signal is subjected to real-time spectrum analysis to identify the main frequency components of the residual vehicle noise. The center frequency and bandwidth of the notch filter are dynamically adjusted according to the identification results, and the order of the notch filter is adaptively adjusted to balance the filtering effect and computational complexity. The adaptive filtering and dynamic notch are applied in series to the original signal, and the noise elimination effect is evaluated in both the time domain and the frequency domain.

[0116] This multi-step, closed-loop feedback approach allows precise identification and effective elimination of vehicle noise, thereby improving the accuracy of leak detection.

[0117] Embodiment 4:

[0118] Combination Figure 4 In this embodiment 3, a water supply pipeline leakage noise monitoring and early warning method based on relevant principles is provided. On the basis of embodiments 1 and 2, in order to effectively identify and eliminate various background noises in a complex environment, thereby improving the accuracy of leakage detection, in the preprocessing process, it also includes:

[0119] The background noise template library constructed includes pipeline vibration noise, valve switch noise, water pump noise and other pipe network equipment noise, as well as bird calls, dog barking and human voice noise. Large sample collection and data enhancement technology are used to generate noise samples. Each noise has its own unique characteristics.

[0120] Noise classification adopts a hybrid architecture based on deep convolutional neural network and long short-term memory network (CNN-LSTM). CNN is used to extract local features and spectral features of noise signals, and LSTM is used to model the temporal association and contextual information of noise signals. Through end-to-end joint training, the weight parameters of CNN and LSTM are optimized to distinguish the noise category.

[0121] Using sliding window and probability voting mechanism, a continuous noise signal is segmented into multiple scales and overlaps, and then sent to CNN-LSTM model for classification and probability estimation. The classification results of each segment are voted and merged to obtain the confidence distribution of noise type. The sound signal is divided into small segments, each of which is analyzed by CNN-LSTM model, and the most likely noise type is determined by voting, and the credibility of each noise type is evaluated.

[0122] Based on this, the filtering and noise reduction strategy is adaptively selected to complete the identification and removal of background noise; by accurately judging the current type of noise, the most appropriate noise reduction solution is selected to specifically remove the noise that interferes with leakage detection.

[0123] By building a comprehensive background noise template library and combining it with advanced deep learning technology (CNN-LSTM hybrid architecture), accurate noise classification is achieved. The sliding window and probabilistic voting mechanism are used to further improve the robustness and reliability of classification.

[0124] Embodiment 5:

[0125] like Figure 5 As shown, this embodiment 5 provides a system for monitoring and early warning of water supply pipeline leakage noise based on the relevant principles as in the above-mentioned embodiment 1, 3 or 4, including:

[0126] There are multiple listening monitoring devices, each of which is configured to be arranged near a preset node of the water supply network. Each listening monitoring device includes a data acquisition module, a wireless communication module and a power supply module. The data acquisition module is used to collect noise signals from the water supply pipeline in real time. The wireless communication module is used to transmit the collected noise signals to a data processing center. The power supply module is used to supply power to the data acquisition module and the wireless communication module.

[0127] The data processing center is used to pre-process the noise signals collected by each listening monitoring device to obtain the pre-processed noise signals and time-frequency images;

[0128] The correlation principle analysis module is used to calculate the mutual correlation coefficients between the pre-processed noise signals collected by different listening monitoring devices based on the correlation principle, obtain the mutual correlation matrix, match it with the mutual correlation matrix according to the preset leakage noise template, determine whether there is leakage, and estimate the leakage position.

[0129] Embodiment 6:

[0130] like Figure 6 As shown, this embodiment 6 provides a system for the water supply pipeline leakage noise monitoring and early warning method based on the relevant principles as in the above-mentioned embodiment 2, 3 or 4. This embodiment 6 is a further limited optimization based on embodiment 5:

[0131] The listening monitoring device also includes a data compression module and an acceleration sensor array. The acceleration sensor array includes multiple acceleration sensors distributed in a circular array and is used to collect vibration signals in different directions. The data compression module is used to adaptively compress the collected noise signals.

[0132] In summary, the present application provides a water supply pipeline leakage noise monitoring and early warning method and system based on relevant principles, which has at least the following beneficial effects compared with the prior art:

[0133] 1. By constructing multi-dimensional matrices such as the cross-correlation matrix and the autocorrelation matrix, accurate correlation analysis of pipe network noise signals can be achieved. This technology breaks through the limitations of traditional single signal analysis and can fully capture the propagation characteristics of leakage noise by calculating the signal correlation between different monitoring devices and between sensors within the same device. This multi-dimensional correlation analysis method significantly improves the accuracy and reliability of leakage detection, can effectively identify tiny leakage signals, and accurately locate the leakage location;

[0134] 2. Innovatively proposed a noise classification architecture and adaptive filtering strategy based on CNN-LSTM. By building a comprehensive background noise template library and using deep learning technology to identify noise types, combined with sliding windows and probabilistic voting mechanisms, it achieves accurate identification and effective suppression of various types of background noise in complex environments. This intelligent noise processing technology can significantly improve the anti-interference ability of the leakage detection system in complex environments and ensure the accuracy of the detection results.

[0135] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0136] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for those of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application intends to include these modifications and variations.

Claims

1. A water supply pipeline leakage noise monitoring and early warning method based on correlation principle, characterized in that: The following steps are involved: S10, arranging a plurality of listening monitoring devices at a plurality of preset nodes of the water supply network, each of the listening monitoring devices comprising a data acquisition module, a wireless communication module and a power supply module; S20, collecting noise signals of the water supply pipeline in real time through the data acquisition module, and transmitting the collected noise signals to the data processing center through the wireless communication module; S30, at the data processing center, preprocessing the noise signals collected by each listening monitoring device to obtain the preprocessed noise signals and time-frequency images; S40, based on the correlation principle, calculating the mutual correlation coefficients between the pre-processed noise signals collected by different listening monitoring devices to obtain a mutual correlation matrix; S50, matching the preset leakage noise template with the cross-correlation matrix to determine whether there is leakage and estimate the leakage position.

2. The method according to claim 1, characterized in that In step S10, each of the listening monitoring devices further includes a data compression module and an acceleration sensor array; the acceleration sensor array includes a plurality of acceleration sensors distributed in a circular array for collecting vibration signals in different directions; In step S20, the noise signal of the water supply pipeline is collected in real time by the data acquisition module, the vibration signal of the water supply pipeline is collected in real time by the acceleration sensor array, and the collected noise signal and vibration signal are adaptively compressed by the data compression module, the compression ratio is automatically adjusted according to the signal complexity and the wireless communication bandwidth, and then the compressed noise signal and vibration signal are transmitted to the data processing center through the wireless communication module; In step S30, in the data processing center, the noise signals and vibration signals collected and compressed by the various listening monitoring devices are adaptively decompressed and preprocessed, and the preprocessing includes adaptive filtering, wavelet noise reduction, normalization and time-frequency transformation operations to obtain preprocessed noise signals, vibration signals and time-frequency images; In step S40, based on the correlation principle, the mutual correlation coefficients and mutual correlation functions between the pre-processed noise signals collected by different listening monitoring devices are calculated to obtain a mutual correlation matrix and a mutual correlation function matrix; at the same time, the autocorrelation coefficients and autocorrelation functions of the multiple vibration signals collected by the acceleration sensor array inside each listening monitoring device are calculated to obtain an autocorrelation matrix and an autocorrelation function matrix; In step S50, according to the preset leakage noise vibration template, multi-feature fusion matching is performed with the mutual correlation matrix, mutual correlation function matrix, autocorrelation matrix and autocorrelation function matrix to determine whether there is leakage, and the leakage location, leakage degree and leakage type are estimated by combining the mutual correlation coefficient, mutual correlation function, autocorrelation coefficient and autocorrelation function.

3. The method according to claim 2, characterized in that In step S10, for a three-way pipe or a complex node with multiple branch pipes, at least one listening monitoring device is arranged on each branch to form a triangular array; In step S40, in addition to calculating the cross-correlation coefficients and cross-correlation functions between different listening monitoring devices, the cross-power spectrum density and phase difference between two of the listening monitoring devices on the multiple branches are also calculated; In step S50, the main frequency component of the leakage noise is determined by the peak frequency of the cross-power spectrum density, and the propagation direction and flow characteristics of the leakage noise in the three-way pipe are determined by the sign and size of the phase difference. Combined with the topological structure of the water supply network and the hydraulic model, the positioning information of the leakage position is quantified.

4. The method according to claim 3, characterized in that In step S10, at least two listening monitoring devices are arranged on each branch of the three-way pipe, and the number of listening monitoring devices arranged on each branch increases in proportion to the length and diameter of the branch, and the distance between adjacent listening monitoring devices does not exceed 10 meters; another listening monitoring device is arranged at the connection of the three-way pipe to form a triangular monitoring array; wherein the configuration formula for the number of listening monitoring devices arranged on each branch is: N=max(2,[L / (10×D)]) Where N is the number of devices, L is the length of the branch pipeline (m), and D is the pipe diameter (m).

5. The method according to claim 4, characterized in that Combine the listening monitoring devices on the same branch with strong correlation, calculate the mutual correlation coefficient and mutual correlation function between them, and judge the propagation direction and speed of the leakage noise on the branch through the extreme value of the mutual correlation coefficient and the time delay corresponding to the extreme value; The listening monitoring devices on different branches are strongly correlated and combined, and the cross-power spectrum density and phase difference between them are calculated. The main propagation path of the leakage noise in the three-way pipe is determined by the peak frequency of the cross-power spectrum density, and the transmission direction and attenuation law of the leakage noise between adjacent branches are determined by the sign and size of the phase difference. Combined with the pipe material and wall thickness parameters, the distance and azimuth between the leakage point and the center of the tee are estimated. Combined with the water supply network topology and hydraulic model, the location information of the leakage position is quantified.

6. The method according to claim 5, characterized in that The leakage judgment process of step S50 also includes: Carry out regional modeling of the water supply network, divide the water supply network into several monitoring zones, each zone contains multiple listening monitoring devices, and form a local monitoring network; In each monitoring zone, a spatial interpolation model and propagation inversion model of leakage noise are constructed, and the Gaussian process regression method is used to estimate the leakage noise intensity distribution at any location in the zone by combining the mutual correlation coefficient and the autocorrelation coefficient. The sound line method is used to simulate the propagation path and attenuation law of leakage noise in the pipeline network, and the spatial coordinates and source intensity of the leakage point are inverted through the least squares method; The positioning results of each partition are summarized and globally optimized to obtain the leakage distribution map and risk level map of the entire water supply network.

7. The method according to any one of claims 1 to 6, characterized in that: Also includes: Identify the shortcut area in the water supply network, and arrange a linear monitoring array consisting of multiple vibration sensors in the shortcut area, wherein the number of the vibration sensors is not less than 3, the arrangement interval is 5-10 meters, and the multiple vibration sensors adopt wireless signal self-organizing network; Noise collection and feature extraction are performed on typical vehicle models in the short-distance area to build a vehicle noise template library. The template parameters include vehicle model, speed, and road conditions. Cluster analysis and statistical modeling are used to characterize the spectral distribution of vehicle noise under different working conditions. Based on the real-time collected vehicle vibration signal, the vehicle position, speed and acceleration parameters are estimated. Combined with the vehicle noise template, a reference noise signal matching the current working condition is adaptively generated as the input of the adaptive filter. The filter weight is updated in real time, and the notch frequency and order are dynamically adjusted to eliminate the vehicle interference noise in the noise signal and time-frequency image.

8. The method according to any one of claims 2 to 6, characterized in that The preprocessing process also includes: The background noise template library constructed includes pipeline vibration noise, valve switch noise, water pump noise and other pipe network equipment noise, as well as bird calls, dog barking, and human voice noise. Large sample collection and data enhancement technology are used to generate noise samples. Noise classification adopts a hybrid architecture based on deep convolutional neural network and long short-term memory network (CNN-LSTM). CNN is used to extract local features and spectral features of noise signals, and LSTM is used to model the temporal association and contextual information of noise signals. Through end-to-end joint training, the weight parameters of CNN and LSTM are optimized to distinguish the noise category. Using sliding window and probability voting mechanism, a continuous noise signal is segmented into multiple scales and overlaps, and then sent to CNN-LSTM model for classification and probability estimation. The classification results of each segment are voted and fused to obtain the confidence distribution of the noise type. Based on this, the filtering and noise reduction strategy is adaptively selected to complete the identification and removal of background noise.

9. A system applied to the water supply pipeline leakage noise monitoring and early warning method based on the correlation principle as claimed in any one of claims 1 to 8, characterized in that: include: A plurality of listening monitoring devices, each of which is configured to be arranged near a preset node of a water supply network, and each of which includes a data acquisition module, a wireless communication module and a power supply module. The data acquisition module is used to collect noise signals of the water supply pipeline in real time, the wireless communication module is used to transmit the collected noise signals to a data processing center, and the power supply module is used to supply power to the data acquisition module and the wireless communication module. The data processing center is used to pre-process the noise signals collected by each listening monitoring device to obtain the pre-processed noise signals and time-frequency images; The correlation principle analysis module is used to calculate the mutual correlation coefficients between the pre-processed noise signals collected by different listening monitoring devices based on the correlation principle, obtain the mutual correlation matrix, match it with the mutual correlation matrix according to the preset leakage noise template, determine whether there is leakage, and estimate the leakage position.

10. The system according to claim 9, characterized in that The listening monitoring device also includes a data compression module and an acceleration sensor array. The acceleration sensor array includes a plurality of acceleration sensors distributed in a circular array and is used to collect vibration signals in different directions. The data compression module is used to adaptively compress the collected noise signals.

Images