Water supply pipeline leakage identification method and device based on active low-frequency sound wave excitation

Abstract

The invention discloses a water supply pipeline leakage identification method based on active low-frequency sound wave excitation, and the method comprises the steps: obtaining a sound signal in a pipeline through a low-frequency sound emission transducer and a hydrophone which are disposed at two ends of the pipeline, and generating a corresponding leakage point position based on a pre-constructed sound wave equation; processing the sound signal to obtain a corresponding time-frequency image, labeling the time-frequency image by judging whether the sound signal is a leakage signal or not, and forming a data set by a signal frequency spectrum, a leakage point position and a label; constructing an identification network; training an identification network by using the data set to obtain a water supply pipeline leakage identification model for identifying whether the pipeline is leaked or not; and inputting the sound signal in the to-be-identified pipeline into the water supply pipeline leakage identification model so as to output an identification result about whether the pipeline is leaked or not and an identification result about the position of a leakage point. The invention further provides a water supply pipeline leakage identification device. The method provided by the invention can effectively improve the accuracy of leakage identification.

Description

Technical Field

[0001] The present invention belongs to the technical field of urban water supply pipe networks, and in particular relates to a method and a device for identifying water supply pipe leakage based on active low-frequency acoustic wave excitation. Background Art

[0002] The problem of urban water supply pipeline leakage is becoming increasingly serious, bringing challenges to water supply management and residents' drinking water safety. According to data from the Ministry of Housing and Urban-Rural Development, the amount of water leaked nationwide is as high as 8 billion tons per year, and the urban leakage rate is generally between 10% and 15%.

[0003] Common methods for identifying water supply pipeline leaks include acoustic detection, hydraulic detection, radar detection, and robot endoscopy detection. Acoustic leak detection technology has become a mainstream technology due to its non-invasiveness, high sensitivity, and low cost. At present, all acoustic leak detection technologies on the market are passive acoustic methods, that is, passively sensing the sound source signal of the leak point. On the one hand, since the signal-to-noise ratio of the leak point sound signal is usually low, the signal is difficult to distinguish from the background noise, and the signal decays quickly due to the damping of the pipeline. In order to achieve effective leak diagnosis, the spacing between sensors in the existing technology generally cannot exceed 200 meters, and the large-scale installation of sensors makes the investment cost high. On the other hand, the sensor layout distance is subject to the distribution of the inspection well, and it is often difficult to achieve the requirement that the spacing between sensors is less than 200 meters. Since the passive acoustic method relying on sensors is far less than expected in actual applications, manual leak detection using listening rods is still the main detection method for water companies. However, this traditional passive acoustic leak detection technology relies on the experience of leak detection workers and requires a lot of manpower and time. Therefore, it is urgent to develop new leakage diagnosis methods to solve the limitations of current technologies.

[0004] Patent document CN103521424A discloses a low-frequency acoustic wave excitation device, including a liquid-filled pipe, a wall opening of the liquid-filled pipe is connected to an outer closed cavity through a connecting cavity, the shells of the cavity and the connecting cavity are fastened and sealed, the connected cavity and the connecting cavity form a measuring cavity, the excitation piston is sealed to the inner wall of the measuring cavity, the excitation piston is fixed to the piston rod, the piston rod passes through the cavity and is slidably connected to the cavity, the shell of the cavity and the piston rod are sealed, the exciter drives the excitation piston to slide in the measuring cavity through the piston rod, an air inlet is provided on the cavity, the excitation piston divides the measuring cavity into an upper cavity and a lower cavity, the air inlet is connected to the upper cavity, and an accelerometer is further arranged on the connecting surface between the hollow piston rod and the excitation piston.

[0005] Patent document CN118309943A discloses a monitoring device for a heating pipe network, the device comprising: an insulated collection component, a sensor, a data collection component and a server; the insulated collection component is filled with air and is used to collect the sound wave signal of the heating pipe network and transmit the sound wave signal of the heating pipe network to the sensor; the sensor is used to receive the sound wave signal of the heating pipe network, convert the sound wave signal into an electrical signal, and then send the electrical signal to the data collection component; the data collection component is used to convert the electrical signal into a digital signal and send the data signal to the server; the server is used to determine whether the heating pipe network has leakage based on the digital signal. Summary of the invention

[0006] The purpose of the present invention is to provide a method and device for identifying water supply pipeline leakage based on active low-frequency acoustic wave excitation, which can effectively improve the accuracy of leakage identification.

[0007] In order to achieve the first object of the present invention, the following technical solution is provided: a method for identifying leakage in a water supply pipeline based on active low-frequency acoustic wave excitation, comprising: obtaining an acoustic signal in the pipeline by means of a low-frequency acoustic emission transducer and a hydrophone arranged at both ends of the pipeline, and generating a corresponding leakage point position based on a pre-constructed acoustic wave equation;

[0008] Based on the time axis direction, the acquired acoustic signal is processed by the STFT analysis method to obtain the corresponding time-frequency image, and the time-frequency image is labeled according to whether it is a leakage signal, and the signal spectrum, leakage point position and label form a data set;

[0009] Constructing a recognition network, wherein the recognition network includes a feature extraction module, a fusion module and an analysis module;

[0010] The feature extraction module is used to extract time domain features and frequency domain features from the input signal spectrum;

[0011] The fusion module is used to fuse the time domain features and the frequency domain features to output fusion features;

[0012] The analysis module performs analysis based on the fusion features to obtain a recognition result;

[0013] Using the data set to train a recognition network to obtain a water supply pipeline leakage recognition model for identifying whether a pipeline is leaking;

[0014] The acoustic signal in the pipeline to be identified is input into the water supply pipeline leakage identification model to output the identification results of whether the pipeline is leaking and the location of the leakage point.

[0015] The present invention utilizes the advantages of slow attenuation of low-frequency excitation sound wave signals and the presence of plane waves in the pipe to overcome the shortcoming of a small monitoring range of a passive acoustic leakage identification method.

[0016] Specifically, the low-frequency acoustic emission transducer generates sound waves with optimal excitation frequency and intensity based on the pipe material, pipe diameter, wall thickness, water pressure in the pipe, flow rate in the pipe and soil cover depth of the target pipe, and the sound waves propagate forward and backward along the pipe.

[0017] Specifically, the acoustic wave equation includes an equation of motion, an equation of continuity and an equation of state.

[0018] Specifically, the expression of the acoustic wave equation is as follows:

[0019]

[0020] dP=c 2 dρ

[0021]

[0022] Where, ρ is liquid density; v is particle velocity; p is sound pressure; t is time; c is sound wave propagation velocity; P is pressure; ω is angular velocity; k is i With k r is the wave number; x, y and z are the spatial directions; i is the incident wave; r is the reflected wave.

[0023] Specifically, the recognition network is constructed based on the pre-trained neural network model SqueezeNet framework and the convolutional neural network model framework.

[0024] Specifically, the STFT analysis method uses short-time Fourier transform to analyze the input sound signal segment by segment to output the corresponding frequency spectrum.

[0025] Specifically, the expression of the short-time Fourier transform is as follows:

[0026] STFT(t,f)=∫[x(τ)ω(τ-t)]e -jfτ dτ

[0027] Where t is time, ω(t) is the window function, f is the frequency, and τ is the time variable used to window and Fourier transform the signal x(t).

[0028] Specifically, the low-frequency acoustic emission transducer is arranged in a fire hydrant at one end of the pipeline.

[0029] In order to achieve the second objective of the present invention, the following technical solution is provided: steps for executing the above-mentioned method for identifying water supply pipe leakage based on active low-frequency acoustic wave excitation.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention obtains the acoustic signal in the pipeline by setting a low-frequency acoustic emission transducer, and determines whether the pipeline is leaking and the location based on the preset model constructed. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 A flow chart of a water supply pipeline leakage identification method based on active low-frequency acoustic wave excitation provided in this embodiment;

[0033] Figure 2 A schematic diagram of a water supply pipeline leakage identification device provided in this embodiment;

[0034] Figure 3 Installation layout diagram of the low-frequency acoustic emission transducer provided in this embodiment;

[0035] In the figure, 1. low-frequency acoustic emission transducer; 2. near-end fire hydrant; 3. far-end fire hydrant; 4. hydrophone; 5. GPS equipment; 6. near-end fire hydrant control valve; 7. far-end fire hydrant control valve; 8. computer; 9. water supply pipeline; 10. leakage point; 11. fire hydrant; 12. flange; 13. valve; 14. water outlet pipeline; 15. skid-mounted shell DETAILED DESCRIPTION

[0036] In order to make the purpose, technical scheme and advantages of the embodiments of the present invention clearer, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0037] like Figure 1 As shown, the water supply pipeline leakage identification method provided by this embodiment has the following specific steps:

[0038] Obtain a training data set for the machine learning model: the active low-frequency sound waves and the sound wave reflection coefficient, transmission coefficient and leakage coefficient under different leakage system characteristics are affected by pipeline characteristics, hydraulic characteristics of water flow in the pipe, and characteristics of soil covering outside the pipe. Thousands of experiments were conducted, and the experimental variables involved sound source frequency, sound source sound pressure level, erosion cavity size, pipe diameter, pipe material, soil covering height, sand-soil ratio, water pressure in the pipe, leakage port shape, leakage port area and leakage port direction. The experiment also includes non-leakage conditions and leakage conditions, with a ratio of about 3:7, to jointly construct a data set of water-pipe-soil three-phase leakage system and low-frequency excitation sound waves in the pipe echo.

[0039] The signal data collected by the hydrophone is the signal amplitude information of a specific duration, which may contain various interference noises, leakage status information, and the signal itself may vary depending on the leakage situation, pipeline and environmental conditions. The method of the present invention combines the time domain characteristics and frequency domain characteristics of the acoustic signal when performing feature extraction, and extracts the characteristics of the nonlinear characteristics that describe the non-stationary characteristics of the leakage signal. At the same time, the signal time-frequency analysis method is also used to describe the change of the signal frequency distribution over time, providing effective data support for the automatic identification of leakage signals in the water supply network.

[0040] STFT is a simple time-frequency analysis tool. Its principle is to use a window function of appropriate width to divide the signal into several small segments in time, perform Fourier transform on each segment of data, and translate the window function on the time axis to analyze the signal segment by segment, and combine them to obtain the spectrum of the original signal.

[0041] For a continuous signal x(t) in the time domain, its short-time Fourier transform is defined as:

[0042] The expression of short-time Fourier transform is as follows:

[0043] STFT(t,f)=∫[x(τ)ω(τ-t)]e -jfτ dτ

[0044] Where t is time, ω(t) is the window function, f is the frequency, and τ is the time variable used to window and Fourier transform the signal x(t).

[0045] Using STFT to obtain the time-frequency image of the signal requires determining several important parameters, including the window function, fast Fourier transform length, frame length, and frame overlap length.

[0046] Training a water supply pipeline leakage identification model based on active low-frequency acoustic wave excitation: Various leakage signals and non-leakage signals collected based on experiments together form a data set, which is used to train the machine learning model and verify its identification effect.

[0047] The pre-trained neural network model SqueezeNet is used to classify the time-frequency images of different signals. SqueezeNet is a lightweight convolutional neural network with up to 18 deep learnable layers, which can classify images into 1,000 different categories. The training goal is to determine whether there is leakage. The data set is defined as only two types of data, one is leakage signal, labeled "1", and the other is non-leakage signal, labeled "0". Leakage signal and non-leakage signal together constitute the data set, which is used to train the machine learning model and verify its recognition effect. The training process is supervised learning, that is, the signal label is also used as model input information.

[0048] The specific training process is as follows:

[0049] (1) Use signal spectrum images to train the SqueezeNet model;

[0050] (2) Use the trained models to perform recognition applications on the validation set, and modify the hyperparameters of the corresponding models based on the application results to obtain better performance on the validation set;

[0051] (3) The test set is input into the trained model to determine whether the signal is leaked and obtain the recognition result.

[0052] In order to verify the accuracy of the leakage identification model in the method of the present invention, 20 groups of experiments were conducted to test its performance. These tests included 15 groups of leakage conditions and 5 groups of non-leakage conditions. In addition, different weather background conditions were considered in the experiments: 11 groups of sunny days, 5 groups of light rain days, and 4 groups of heavy rain days. The experimental results show that under the background condition of heavy rain days, the leakage identification accuracy of the model is 75%; while under the background of sunny days and light rain days, the leakage identification accuracy reaches 100%.

[0053] like Figure 2 As shown, this embodiment also provides a water supply pipe leakage identification device, which is used to execute the steps of the water supply pipe leakage identification method provided in the above embodiment.

[0054] The original pipeline includes a water supply pipeline 9 and a leakage point 7, a proximal fire hydrant 2 and a control valve 6, and a distal fire hydrant 3 and a control valve 7.

[0055] The method provided in this embodiment adds equipment to the original pipeline, including a skid-mounted low-frequency acoustic emission transducer 1 and a hydrophone 4 for acoustic signal monitoring, and the hydrophone 4 transmits the collected acoustic signal data to a remote computer 8 via a GPS device 5. The presence of the leakage point 10 will change the characteristics of the acoustic signal collected by the hydrophone.

[0056] like Figure 3As shown, the docking method of the low-frequency acoustic emission transducer and the fire hydrant provided in this embodiment, the low-frequency acoustic emission transducer 1 is built into the skid-mounted housing 15, and the power supply cable passes through the skid-mounted housing 15 to ensure the sealing performance. The low-frequency acoustic emission transducer 16 and the skid-mounted housing 15 are arranged on the water supply pipeline as a whole to replace the position of the fire hydrant. The installation process includes: first closing the water outlet control valve 13 of the fire hydrant 11, then disassembling the fire hydrant flange 12, then docking with the skid-mounted housing through the flange, and finally opening the fire hydrant water outlet control valve 13 to ensure smooth flow.

[0057] Since low-frequency sound waves present a relatively pure plane wave in the water supply pipe, they have less attenuation when propagating in water and can maintain a longer propagation distance in the pipe. In addition, the plane wave formed by low-frequency sound waves can maintain good directional propagation, and has higher transmission efficiency and stability when used to transmit information.

[0058] The device provided in this embodiment is based on the special scattering phenomenon that occurs when low-frequency sound waves encounter a leaky structure. The principle is to establish a connection between the changes in sound pressure over space and time. This connection is established based on the equation of motion, the equation of continuity and the equation of state.

[0059] The specific formula is as follows:

[0060]

[0061] dP=c 2 dρ

[0062] Where, ρ is liquid density; v is particle velocity; p is sound pressure; t is time; c is sound wave propagation velocity; P is pressure; ω is angular velocity; k is i With k r is the wave number; x, y and z are the spatial directions; i is the incident wave; r is the reflected wave.

[0063] Since the sound wave parameters p, v, ρ′ and their changes in space and time are tiny in the water supply pipe, the linear sound wave equation can be obtained by ignoring the trace quantities above the second order and solving the equation.

[0064] Low-frequency sound waves propagate in a bounded pipe. When they encounter a leak, a reflected wave p is generated in the pipe. r , and a transmitted wave p t , there will be leakage waves b At this time, the sound pressure p at any point in the pipe is the superposition of the incident wave and the reflected wave:

[0065]

[0066] The water-pipe-soil three-phase leakage system is the main reason for the scattering of low-frequency sound waves. Therefore, the size of the reflected wave, transmitted wave and leakage wave is not only related to the leakage degree and the characteristics of the low-frequency sound wave, but also closely related to the acoustic impedance of the erosion cavity outside the leak hole.

[0067] In addition, the terms "upper", "lower", "inner", "outer", "front", "rear" are used for descriptive purposes only and should not be understood as indicating or implying relative importance. Unless otherwise specifically stated, the relative steps, numerical expressions and values ​​of the components and steps described in these embodiments do not limit the scope of the present invention.

[0068] Of course, the above description is only a specific embodiment of the present invention and is not intended to limit the scope of implementation of the present invention. All equivalent changes or modifications made according to the structure, characteristics and principles described in the patent application scope of the present invention should be included in the patent application scope of the present invention.

[0069] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A water supply pipeline leakage identification method based on active low-frequency acoustic wave excitation, characterized in that: include: The acoustic signal in the pipeline is obtained by arranging low-frequency acoustic emission transducers and hydrophones at both ends of the pipeline, and the corresponding leakage point position is generated based on the pre-built acoustic wave equation; Based on the time axis direction, the acquired acoustic signal is processed by the STFT analysis method to obtain the corresponding time-frequency image, and the time-frequency image is labeled according to whether it is a leakage signal, and the signal spectrum, leakage point position and label form a data set; Constructing a recognition network, wherein the recognition network includes a feature extraction module, a fusion module and an analysis module; The feature extraction module is used to extract time domain features and frequency domain features from the input signal spectrum; The fusion module is used to fuse the time domain features and the frequency domain features to output fusion features; The analysis module performs analysis based on the fusion features to obtain a recognition result; Using the data set to train a recognition network to obtain a water supply pipeline leakage recognition model for identifying whether a pipeline is leaking; The acoustic signal in the pipeline to be identified is input into the water supply pipeline leakage identification model to output the identification results of whether the pipeline is leaking and the location of the leakage point.

2. The method for identifying water supply pipeline leakage based on active low-frequency acoustic wave excitation according to claim 1 is characterized in that: The low-frequency acoustic emission transducer generates sound waves with optimal excitation frequency and intensity based on the pipe material, pipe diameter, wall thickness, water pressure in the pipe, flow rate in the pipe and soil cover depth of the target pipe, and the sound waves propagate forward and backward along the pipe.

3. The method for identifying water supply pipeline leakage based on active low-frequency acoustic wave excitation according to claim 1 is characterized in that: The acoustic wave equation includes the equation of motion, the equation of continuity and the equation of state.

4. The method for identifying water supply pipeline leakage based on active low-frequency acoustic wave excitation according to claim 3 is characterized in that: The expression of the acoustic wave equation is as follows: dP=c 2 dρ Where, ρ is liquid density; v is particle velocity; p is sound pressure; t is time; c is sound wave propagation velocity; P is pressure; ω is angular velocity; k is i With k r is the wave number; x, y and z are the spatial directions; i is the incident wave; r is the reflected wave.

5. The method for identifying water supply pipeline leakage based on active low-frequency acoustic wave excitation according to claim 1 is characterized in that: The recognition network is constructed based on the pre-trained neural network model SqueezeNet framework and the convolutional neural network model framework.

6. The method for identifying water supply pipeline leakage based on active low-frequency acoustic wave excitation according to claim 1 is characterized in that: The STFT analysis method uses short-time Fourier transform to analyze the input sound signal segment by segment to output the corresponding frequency spectrum.

7. The method for identifying water supply pipeline leakage based on active low-frequency acoustic wave excitation according to claim 6 is characterized in that: The expression of the short-time Fourier transform is as follows: STFT(t,f)=∫[x(τ)ω(τ-t)]e -jfτ dt Where t is time, ω(t) is the window function, f is the frequency, and τ is the time variable used to window and Fourier transform the signal x(t).

8. The method for identifying water supply pipeline leakage based on active low-frequency acoustic wave excitation according to claim 1 is characterized in that: The low-frequency acoustic emission transducer is arranged in a fire hydrant at one end of the pipeline.

9. A water supply pipeline leakage identification device, characterized in that: Used to execute the steps of the water supply pipeline leakage identification method based on active low-frequency acoustic wave excitation as described in any one of claims 1 to 8.

Images