A method for locating tiny leaks in pipelines using infrasound waves

Abstract

This invention discloses an infrasound localization method for minor pipeline leaks, comprising: acquiring the original infrasound signal of the pipeline leak; performing variational mode decomposition on the original infrasound signal to obtain IMF components; based on the IMF components, using an improved generalized cross-correlation function to obtain the time difference between the infrasound reflection to the upstream and downstream ends of the pipeline; calculating the wave velocity of the infrasound according to the pipeline characteristics; and determining the location of the pipeline leak point based on the time difference and wave velocity. The improved generalized cross-correlation function method involves: constructing an analytical signal based on the IMF components; using an improved linear adaptive weighted local approximation algorithm to obtain the local optimal solution of the IMF components; and performing inverse attenuation dispersion compensation analysis based on the generalized cross-correlation function to obtain the time difference between the infrasound reflection to the upstream and downstream ends of the pipeline. This invention can accurately locate the pipeline leak point.

Description

Technical Field

[0001] This invention relates to an infrasound method for locating minor leaks in pipelines, belonging to the field of pipeline leak location technology. Background Technology

[0002] With the continuous advancement of policies aimed at building a modern industrial system and achieving carbon peaking and carbon neutrality, the safe and reliable operation of urban pipeline networks has received increasing attention. Due to the continuous expansion of urban pipeline networks, and the impacts of natural aging of equipment, extreme weather, geological disasters, and human-caused damage, incidents involving urban pipeline networks are becoming increasingly frequent. Urban pipeline networks are mostly distributed in a network structure, covering a wide area with numerous nodes. For minute leak signals from buried urban pipelines, the collected signals are mixed with a large amount of environmental noise, media noise, and other irrelevant signals. Therefore, effectively detecting and locating initial minute leaks in pipelines and accurately pinpointing the leak location has significant economic and social value.

[0003] In recent years, with the development of computer technology, the world is gradually entering a digital society, and pipeline leak detection and location technology is also developing towards a combination of hardware and software. Variational Mode Decomposition (VMD), proposed by Dragomiretskiy et al. in 2014, is a novel adaptive fault diagnosis method. As a decomposition algorithm, VMD, similar to EMD and LMD methods, can decompose fault signals into several intrinsic mode functions (IMFs) based on high and low frequencies. Generalized Cross Correlation (GCC) is one of the most commonly used time delay estimation algorithms. It calculates the cross-correlation function of two signals, and its peak value is the estimated TDOA. GCC is widely used in partial discharge source location due to its strong noise resistance, simple calculation, and low susceptibility to human factors. However, signal attenuation and dispersion can cause distortion of the PD signal waveform, reduce the correlation between signals, and increase the positioning error of the generalized cross-correlation method. In addition, GCC is based on an ideal signal transmission model, which is not only affected by noise and reverberation in practice, but also limited by the sampling frequency, resulting in low TDOA estimation accuracy.

[0004] When VMD and generalized cross-correlation algorithms are applied to the detection of minor leaks in pipelines, the infrasound leak signal, after being decomposed by VMD, exhibits a multi-peak oscillation signal characteristic when it propagates to the end sensor. Directly performing correlation calculations on the monitoring signals from the sensors at both ends may result in correlation peaks appearing at any peak point of the oscillation signal, making it difficult to accurately locate the leak point. Summary of the Invention

[0005] To address the problems existing in the prior art, this invention proposes an infrasound localization method for minor pipeline leaks. The method involves performing a Hilbert transform on the IMF components to construct an analytical signal, and then using an improved linear adaptive weighted local approximation algorithm to find the local optimum solution for each IMF component band. An improved inverse attenuation dispersion compensation is applied to the generalized cross-correlation function to reduce interference from attenuation and dispersion. Finally, through a Hilbert difference transform, the time delay value is changed from peak position search to zero-point detection, reducing the impact of distorted signal waveforms on localization and improving accuracy.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical means:

[0007] This invention proposes an infrasound method for locating minute leaks in pipelines, comprising the following steps:

[0008] Obtain the raw infrasound signal of the pipeline leak;

[0009] Variational mode decomposition was performed on the original infrasound signal to obtain the IMF components of each level;

[0010] Based on the IMF components, the time difference between the infrasound reflection to the upstream and downstream ends of the pipeline is obtained using an improved generalized cross-correlation function.

[0011] Calculate the wave velocity of infrasound based on the characteristics of the pipeline;

[0012] The location of the pipeline leak point is obtained using the time difference and wave velocity method.

[0013] Furthermore, based on the IMF components, the method for obtaining the time difference between infrasound reflections to the upstream and downstream ends of the pipeline using an improved generalized cross-correlation function is as follows:

[0014] Construct an analytical signal based on IMF components;

[0015] Based on the analytical signal, the local optimal solution of the IMF component is obtained using an improved linear adaptive weighted local approximation algorithm;

[0016] Based on the local optimal solution of the IMF component, an inverse attenuation dispersion compensation analysis based on generalized cross-correlation is performed to obtain the time difference between the infrasound reflection to the upstream and downstream ends of the pipeline.

[0017] Furthermore, the method for constructing an analytical signal based on IMF components is as follows:

[0018] Perform Hilbert transform on the denoised IMF components of the signal:

[0019]

[0020] Where p(t) represents the signal after the Hilbert transform, uk The signal is denoised as an IMF component, k is the number of IMF components in each layer, and t is time.

[0021] Construct an analytic signal based on signal p(t):

[0022] z(t) = u k +jp(t) (2)

[0023] Where z(t) is an analytic signal.

[0024] Furthermore, based on the analytical signal, the method for obtaining the local optimal solution of the IMF component using the improved linear adaptive weighted local approximation algorithm is as follows:

[0025] The analytic signal z(t) is divided into n segments, and the IMF components are weighted using a weighting function and a dynamic weighting factor:

[0026] z(t)=[z(1),z(2),...,z(x),...,z(n),x=1,...,n] (3)

[0027] g(x)=K(d)w(t)z(x) (4)

[0028] Where g(x) represents the weighted IMF component of the x-th segment, K(d) is the weighting function, w(t) is the dynamic weighting factor, and z(x) is the analytic signal of the x-th segment;

[0029] The expression for K(d) is as follows:

[0030]

[0031] Where d is the distance from the measurement domain to the signal center;

[0032] The expression for w(t) is as follows:

[0033] w(t+1)=4w(t)·(1-w(t)) (6)

[0034] Local approximation of the weighted IMF components:

[0035] g(x) p+1 =g(x) p +βe r z(x) (7)

[0036] Where, g(x) p+1 Let β represent the IMF component of the local approximation of segment x after the (p+1)th iteration, where β is the approximation rate, |β|≤1, and e r Indicates the signal approximation value.

[0037] When g(x) p+1 >g(x) p When the iteration stops, the local optimum of the IMF component is obtained:

[0038] h(n)=[g(1) * ,g(2) * ,g(3) * ,...,g(n) * (8)

[0039] Where h(n) represents the local optimal solution of the IMF component, and g(n) * This represents the optimal IMF component for the local approximation of the nth segment.

[0040] Furthermore, based on the local optimal solution of the IMF components, an inverse attenuation dispersion compensation analysis based on generalized cross-correlation is performed to obtain the time difference of infrasound reflection to the upstream and downstream ends of the pipeline.

[0041] The generalized cross-correlation function is obtained from the local optimal solution of the IMF components:

[0042]

[0043] Where GET represents the generalized cross-correlation function, Ψ 1，2 (ω) denotes the generalized cross-correlation weighting function, where ω is the signal frequency. h1(n) represents the cross-power spectrum of the signals received by the upstream and downstream sensors in the pipeline after processing, h2(n) represents the local optimal solution of the signal received by the upstream sensor after processing, and h2(n) represents the local optimal solution of the signal received by the downstream sensor after processing.

[0044] Calculate the inverse attenuation dispersion compensation coefficient of the signal characteristics based on the generalized cross-correlation function:

[0045] G -1 =[G -1 (1,t),...,G -1 (x,t),...,G -1 (n,t)] (10)

[0046]

[0047]

[0048]

[0049] Among them, G -1 G represents the anti-attenuation dispersion compensation coefficient. -1(x,t) represents the inverse attenuation dispersion compensation coefficient of the x-th segment of the IMF component at time t, A(x,t) represents the amplitude compensation operator of the x-th segment of the IMF component at time t, Q is a linear function of the attenuation factor with frequency, U(x,t) represents the phase compensation operator of the x-th segment of the IMF component at time t, Q>1, x=1,2,...,n, n is the number of segments of the IMF component, and t is time;

[0050] The generalized cross-correlation function after anti-attenuation dispersion compensation is obtained based on the anti-attenuation dispersion compensation coefficient:

[0051] GET' = GET·G -1 (14)

[0052] Where GET' represents the generalized cross-correlation function after inverse attenuation dispersion compensation;

[0053] Applying a Hilbert difference transform to the generalized cross-correlation function after inverse attenuation dispersion compensation changes the time delay value from peak position search to zero point detection:

[0054]

[0055] in, The improved generalized cross-correlation function is given, where τ is the time difference between the infrasound wave reflecting to the upstream and downstream ends of the pipeline.

[0056] Furthermore, the expression for calculating the infrasound wave velocity based on the pipe characteristics is as follows:

[0057]

[0058] Where v is the wave velocity of the infrasound, β is the compressibility coefficient of the transport medium, ρ is the density of the transport medium, D is the pipe diameter, E is the elastic modulus of the pipe, e is the thickness of the pipe, and C is the correction factor.

[0059] The formula for calculating β is as follows:

[0060]

[0061] Where Z is the compressibility factor of the conveying medium, and P is the conveying pressure of the conveying medium.

[0062] Furthermore, the conveying medium is dry air, and the compressibility factor of the conveying medium is... Where p is the absolute pressure of the gas.

[0063] Furthermore, the formula for calculating the location of the pipeline leak point using the time-difference positioning method is as follows:

[0064]

[0065] Where M is the location of the pipeline leak, l is the distance between the upstream and downstream sensors in the pipeline, v is the wave velocity of the infrasound, and τ is the time difference between the infrasound reflecting to the upstream and downstream ends of the pipeline.

[0066] The following advantages can be obtained by adopting the above technical means:

[0067] This invention proposes an infrasound-based method for locating minor pipeline leaks. First, variational mode decomposition (VMD) is performed on the acquired raw infrasound signal to determine the IMF components at each level. Second, improved generalized cross-correlation function analysis is applied to the IMF components to determine the time difference between the leak signal transmission to upstream and downstream sensors. Finally, the infrasound velocity is estimated based on pipeline characteristics, and the time difference positioning method is used to locate the pipeline leak. This method can accurately pinpoint the location of the pipeline leak.

[0068] In improving the generalized cross-correlation function, this invention sharpens the IMF classification to highlight the effective signal in the correlation peak, thus avoiding the problem of the correlation peak appearing at arbitrary peak points of the oscillating signal. This invention performs inverse attenuation dispersion compensation on the generalized cross-correlation function to reduce the interference of attenuation and dispersion, and through Hilbert difference transform, changes the time delay value from peak position search to zero point detection, reducing the impact of distorted signal waveforms on positioning, making the positioning results more accurate. Attached Figure Description

[0069] Figure 1 This is a flowchart illustrating the steps of an infrasound method for locating minor leaks in a pipeline according to the present invention.

[0070] Figure 2 This is a schematic diagram of the test pipeline system in an embodiment of the present invention;

[0071] Figure 3 This is a schematic diagram of the original infrasound signal in an embodiment of the present invention;

[0072] Figure 4 This is a VMD decomposition diagram in an embodiment of the present invention;

[0073] Figure 5 This is a schematic diagram of the analytical function of the upstream sensor in an embodiment of the present invention;

[0074] Figure 6 This is a schematic diagram of the analytical function of the downstream sensor in an embodiment of the present invention;

[0075] Figure 7 This is a schematic diagram illustrating the correlation between upstream and downstream sensors in an embodiment of the present invention;

[0076] Figure 8 This is a schematic diagram of the generalized cross-correlation function before improvement in this embodiment of the invention;

[0077] Figure 9 This is a schematic diagram of the improved generalized cross-correlation function in an embodiment of the present invention. Detailed Implementation

[0078] The technical solution of the present invention will be further described below with reference to the accompanying drawings:

[0079] This invention proposes an infrasound localization method for minor leaks in pipelines, such as... Figure 1 As shown, the specific steps include the following:

[0080] Step A: Obtain the raw infrasound signal of the pipeline leak. The method of this invention obtains the infrasound signal of the pipeline through an infrasound acquisition device (sensor) installed in the pipeline. When a leak occurs in the pipeline, the infrasound signal at the leak point propagates along the transport medium in the pipeline, and the raw infrasound signal X(t) of the pipeline leak is collected by the upstream sensor, where t is time.

[0081] Step B: Perform variational mode decomposition on the original infrasound signal to obtain the IMF components of each level.

[0082] Step B01: Decompose the original infrasound signal X(t) into k intrinsic mode function components, as follows:

[0083]

[0084] Among them, u k (t) represents the k-th intrinsic mode function component, k = 1, 2, ..., s, where s is the number of intrinsic mode function components.

[0085] Step B02: Calculate each mode function u using the Hilbert transform. k The analytic signal of (t) is obtained by its one-sided spectrum:

[0086]

[0087] in, The spectrum is a single-sided spectrum, δ(t) is the unit impulse function, and j is the imaginary sign.

[0088] Step B03: Mix the analytical signal with an estimated center frequency. Modulate the spectrum of each mode to the corresponding baseband:

[0089]

[0090] in, The spectrum of each mode is modulated to the corresponding fundamental frequency band, w k denoted as the center frequency of the modal function.

[0091] Step B04: Calculate the squared L2 norm of the gradient of the demodulated signal to estimate the bandwidth of each modal component. The constrained variational model is as follows:

[0092]

[0093] Where * denotes convolution, f(t) represents the leaked signal after preliminary screening, and u k This represents the IMF component after signal denoising.

[0094] Step B05: Introduce the quadratic penalty factor α and the Lagrange multiplication operator λ(t) to transform the constrained variational problem into an unconstrained variational problem. The extended Lagrange expression is as follows:

[0095]

[0096] Where λ is the Lagrange factor, used to ensure the strictness of the constraints, {u k} represents the modal components, {w k} represents the center frequency.

[0097] Step B06: Iterate repeatedly using the alternating direction multiplier algorithm to find the minimum saddle point of the extended Lagrange expression. The optimal solution is the modal components and center frequency of the IMF.

[0098]

[0099]

[0100]

[0101] in, This represents the modal components at the (y+1)th iteration. This represents the center frequency at the (y+1)th iteration. Let f(w) represent the saddle point of the augmented Lagrange function at the (y+1)th iteration, f(w) represent the input signal, w represent the frequency value, and ξ represent the noise tolerance, satisfying the fidelity requirement of signal decomposition.

[0102] Step B07: Use the alternating direction method of multipliers to solve for the extrema of the augmented Lagrange expression, and then decompose the original signal into k modal components. The solution process is as follows:

[0103] ① Initialization and y;

[0104] ②Let y = y + 1, and perform an iterative loop;

[0105] ③For uk w k Perform alternating updates, repeating this process k times.

[0106] ④ Update

[0107] ⑤ Repeat the above steps iteratively until the termination condition is met.

[0108] Step C: Based on the IMF components, the time difference between the infrasound reflection to the upstream and downstream ends of the pipeline is obtained using the improved generalized cross-correlation function.

[0109] To address the issue that during generalized cross-correlation analysis, leaked signals propagate to the end sensor exhibiting multi-peak oscillating characteristics, directly performing correlation calculations on the signals monitored by the sensors at both ends can lead to correlation peaks appearing at arbitrary peak points of the oscillating signal, affecting positioning accuracy. This invention sharpens the processed IMF component signal to highlight the effective signal within the correlation peaks. Furthermore, to address the problem that attenuation and dispersion phenomena cause signal waveform distortion, reducing the correlation between signals and affecting positioning accuracy, this invention first improves the generalized cross-correlation function with inverse attenuation and dispersion compensation to reduce interference from attenuation and dispersion. Then, through Hilbert difference transform, the time delay value is changed from peak position lookup to zero-point detection, reducing the impact of distorted signal waveforms on positioning.

[0110] Step C01: Construct an analytical signal based on the IMF components.

[0111] 1) Perform Hilbert transform on the denoised IMF components of the signal:

[0112]

[0113] Where p(t) represents the signal after the Hilbert transform, u k The denoising represents the IMF components after signal denoising, k represents the number of IMF components in each layer, and t represents time.

[0114] 2) Construct an analytic signal based on the signal p(t):

[0115] z(t) = u k +jp(t) (28)

[0116] Where z(t) is an analytic signal.

[0117] Step C02: Based on the analytical signal, obtain the local optimal solution of the IMF component using the improved linear adaptive weighted local approximation algorithm.

[0118] 1) Divide the analytic signal z(t) into n segments, and use a weighting function and dynamic weighting factor to weight the IMF components:

[0119] z(t)=[z(1),z(2),...,z(x),...,z(n),x=1,...,n] (29)

[0120] g(x)=K(d)w(t)z(x) (30)

[0121] Where g(x) represents the weighted IMF component of the x-th segment, K(d) is the weighting function, w(t) is the dynamic weighting factor, and z(x) is the analytic signal of the x-th segment.

[0122] The expression for K(d) is as follows:

[0123]

[0124] Where d is the distance from the measurement field to the signal center, and ranging refers to the detection length of each signal segment.

[0125] The expression for w(t) is as follows:

[0126] w(t+1)=4w(t)·(1-w(t)) (32)

[0127] 2) Perform a local approximation on the weighted IMF components:

[0128] g(x) p+1 =g(x) p +βe r z(x) (33)

[0129] Where, g(x) p+1 Let β represent the IMF component of the local approximation of segment x after the (p+1)th iteration, where β is the approximation rate, |β|≤1, and e r Indicates the signal approximation value.

[0130] 3) When g(x) p+1 >g(x) p When the iteration stops, the local optimum of the IMF component is obtained:

[0131] h(n)=[g(1) * ,g(2) * ,g(3) * ,...,g(n) * (34)

[0132] Where h(n) represents the local optimal solution of the IMF component, and g(n) * This represents the optimal IMF component for the local approximation of the nth segment.

[0133] Step C03: Based on the local optimal solution of the IMF component, perform inverse attenuation dispersion compensation analysis based on generalized cross-correlation to obtain the time difference of infrasound reflection to the upstream and downstream ends of the pipeline.

[0134] 1) By analyzing the relationship between the delay τ and frequency ω of the correlation function using the generalized cross-correlation transform, we obtain:

[0135]

[0136] Among them, x1' and x2' are the two optimal observation signals, specifically referring to the signals that are most correlated after processing of the upstream and downstream sensor signals. Let GET represent the time-frequency distribution of the cross-correlation function of the two best observed signals, where GET is the generalized cross-correlation function.

[0137] 2) Based on step C02, the local optimal solutions for the signals received by the upstream and downstream sensors in the pipeline are obtained, and then the generalized cross-correlation function is obtained:

[0138]

[0139] Among them, Ψ 1，2 (ω) denotes the generalized cross-correlation weighting function, where ω is the frequency of the cross-correlation function between the two best observed signals. Let h1(n) represent the cross-power spectrum of the signals received by the upstream and downstream sensors in the pipeline after processing, and h2(n) represent the local optimal solution of the signal received by the upstream sensor after processing.

[0140] 3) Calculate the inverse attenuation dispersion compensation coefficient of the signal characteristics based on the signal characteristics of the generalized cross-correlation function:

[0141] G -1 =[G -1 (1,t),...,G -1 (x,t),...,G -1 (n,t)] (37)

[0142]

[0143]

[0144]

[0145] Among them, G -1 G represents the anti-attenuation dispersion compensation coefficient. -1(x,t) represents the inverse attenuation dispersion compensation coefficient of the x-th segment of the IMF component at time t, A(x,t) represents the amplitude compensation operator of the x-th segment of the IMF component at time t, Q is a linear function of the attenuation factor with frequency, U(x,t) represents the phase compensation operator of the x-th segment of the IMF component at time t, Q>1, x=1,2,...,n, n is the number of segments of the IMF component, and t is time;

[0146] 4) Obtain the generalized cross-correlation function after inverse attenuation dispersion compensation based on the inverse attenuation dispersion compensation coefficient:

[0147] GET' = GET·G -1 (41)

[0148] Here, GET' represents the generalized cross-correlation function after inverse attenuation dispersion compensation.

[0149] 5) Perform Hilbert difference transform on the generalized cross-correlation function after inverse attenuation dispersion compensation, changing the time delay value from peak position search to zero point detection:

[0150]

[0151] in, The improved generalized cross-correlation function is given, where τ is the time difference between the infrasound wave reflecting to the upstream and downstream ends of the pipeline.

[0152] As shown in the above formula, the generalized cross-correlation function after inverse attenuation dispersion compensation transforms from an even function to an odd function, and the delay estimation changes from peak position finding to zero point detection, avoiding the problem of insufficient peak prominence. Since the peak of the improved generalized cross-correlation function corresponds to the zero point of its waveform, the peak position is preserved while the improved generalized cross-correlation function becomes sharper.

[0153] The time difference τ can be obtained by searching for the zero value using formula (40).

[0154] Step D: Calculate the wave velocity of the infrasound based on the characteristics of the pipeline.

[0155] The actual propagation speed of infrasound is affected by a variety of factors, such as the volume elasticity of the fluid medium, ambient temperature, and operating pressure. These factors work together to cause the propagation speed of infrasound to vary under different operating conditions and transportation conditions. This invention introduces an infrasound wave speed correction formula to correct the wave speed.

[0156] Step D01: Calculate the compressibility coefficient β of the transported medium. β is related to the compressibility factor.

[0157]

[0158] Where Z is the compressibility factor of the conveying medium, and P is the conveying pressure of the conveying medium, in Pa.

[0159] Step D02: The conveying medium is dry air. The formula for calculating the compressibility factor of the conveying medium is as follows:

[0160]

[0161] Where p is the absolute pressure of the gas, in MPa.

[0162] Step D03: The formula for calculating the wave velocity of infrasound is as follows:

[0163]

[0164] Where v is the wave velocity of the infrasound, in m / s, ρ is the density of the transported medium, in kg / m3, D is the pipe diameter, in m, E is the elastic modulus of the pipe, e is the thickness of the pipe, and C is the correction factor.

[0165] Step E: Based on the time difference and wave velocity, the location of the pipeline leak is determined using the time difference positioning method. The calculation formula is as follows:

[0166]

[0167] Where M is the location of the pipeline leak point, and l is the distance between the upstream and downstream sensors inside the pipeline.

[0168] To verify the effectiveness of the method of the present invention, the following experiments are provided in the embodiments of the present invention:

[0169] The experimental pipeline system involved in this invention consists of three parts: a non-metallic buried pipeline network, an infrasound acquisition device, and a pressure monitoring system. A simplified diagram of the site is shown below. Figure 2 As shown in Table 1, this invention uses an air compressor to simulate the operation of an urban gas pipeline. Specific pipeline parameters and transport medium parameters are shown in Table 1.

[0170] Table 1

[0171]

[0172] Step S1: Adjust the working pressure of the air compressor to 0.3 MPa, and use an infrasound acquisition device to obtain the infrasound signal of the pipeline. Simulate pipeline leakage by opening the leakage valve, and collect data every 1 minute for 1mm, 2mm, 3mm, and 4mm leakage conditions, continuously collecting data for 300 seconds to obtain a total of 3 sets of data; thus, a total of 12 sets of pipeline infrasound signals corresponding to the operating conditions are obtained, as shown in Table 2:

[0173] Table 2

[0174]

[0175]

[0176] Step S2: Randomly select the 8th group of acquired signals, perform variational mode decomposition to obtain several IMF components, and then obtain the VMD decomposition diagram. The signals acquired in the 8th group are as follows: Figure 3 As shown, the VMD decomposition diagram is as follows: Figure 4 As shown.

[0177] Step S3: Based on the IMF components, the time difference between the infrasound reflection and the upstream and downstream ends of the pipeline is obtained using the improved generalized cross-correlation function. Specifically, for the IMF components u... k A Hilbert transform is performed to construct an analytic signal z(x). Then, an improved linear adaptive weighted local approximation algorithm is used to sharpen the analytic signal, finding the local optimum h(n) for each IMF component in each band. The analytic function image from the upstream sensor is shown below. Figure 5 As shown, the downstream sensor analytical function graph is as follows: Figure 6 As shown. An improved generalized cross-correlation function analysis is performed on the local optimal solution h(n) of each IMF component to determine the time difference τ. The correlation between upstream and downstream sensors is as follows. Figure 7 As shown.

[0178] The generalized cross-correlation function before improvement is as follows: Figure 8 As shown, the improved generalized cross-correlation function for delay estimation, based on inverse attenuation dispersion compensation and zero-point detection, is as follows: Figure 9 As shown. By Figure 8 It can be seen that amplitude spikes appear in the time-domain waveforms of the generalized cross-correlation of the upstream and downstream leakage signals before the improvement, indicating that the time for the infrasound to travel to the downstream end sensor is 0.066405s. Therefore, the infrasound signal time difference τ = 0.066405s. Figure 9 It can be seen that the improved positioning delay value τ = 0.053086ms, and the positioning is performed based on the infrasound positioning principle.

[0179] Step S4: Calculate the infrasound velocity under this working condition based on the fluid sound velocity equation: Given a wall thickness e of 1.5 mm and an air density ρ of 1.29 kg / m³. 3 The elastic modulus E is 2.9 GPa, the leakage hole diameter D is 1 mm, the correction factor C is 0.91, the compressibility coefficient β of the conveyed medium is related to the pipeline pressure, P is the average value of the resting pressure before and after the leakage, which is 0.15 MPa, and the absolute pressure p is 0.3 MPa.

[0180] According to formula (44), the compressibility factor of dry air is:

[0181] According to formula (43), the compressibility coefficient β of the conveying medium can be obtained. According to formula (45), the infrasound wave velocity can be obtained.

[0182]

[0183] Step S5: Substitute the wave velocity v and time difference τ into formula (46) to obtain the location of the leakage point.

[0184] If the improved generalized cross-correlation function of this invention is not used, with wave velocity v = 342.72 m / s and time difference τ = 0.066405 ms, the original generalized cross-correlation positioning result is M = (19 + 0.066405 × 342.72) / 2 = 20.8791 m.

[0185] If the improved generalized cross-correlation function of this invention is used, with wave velocity v = 342.72 m / s and time difference τ = 0.053086 ms, the improved generalized cross-correlation positioning result is M = (19 + 0.053086 × 342.72) / 2 = 18.5968 m.

[0186] The original generalized cross-correlation infrasound leak detection location error was:

[0187] ε1=|(19-20.8791) / 38|×100%=4.945%

[0188] The improved generalized cross-correlation infrasound leak detection location error is:

[0189] ε2=|(19-18.5968) / 38|×100%=1.061%

[0190] The average positioning error of the coarse positioning based on the infrasound detection system is 7.06%.

[0191] As can be seen from the above errors, the improved generalized cross-correlation infrasound leak location accuracy > the unimproved generalized cross-correlation infrasound leak detection location accuracy > the coarse location accuracy based on the infrasound detection system. Therefore, the method of the present invention can effectively determine the leak and accurately locate it, thus improving the accuracy of pipeline leak location.

[0192] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for locating minute leaks in pipelines using infrasound, characterized in that, Includes the following steps: Obtain the raw infrasound signal of the pipeline leak; Variational mode decomposition was performed on the original infrasound signal to obtain the IMF components of each level; Based on the IMF components, the time difference between the infrasound reflection to the upstream and downstream ends of the pipeline is obtained using an improved generalized cross-correlation function. Calculate the wave velocity of infrasound based on the characteristics of the pipeline; The location of the pipeline leak point is obtained using the time difference and wave velocity method. The method for obtaining the time difference between infrasound reflection to the upstream and downstream ends of the pipeline based on IMF components and using an improved generalized cross-correlation function is as follows: Construct an analytical signal based on IMF components; Based on the analytical signal, the local optimal solution of the IMF component is obtained using an improved linear adaptive weighted local approximation algorithm; Based on the local optimal solution of the IMF component, an inverse attenuation dispersion compensation analysis based on generalized cross-correlation is performed to obtain the time difference between the infrasound reflection to the upstream and downstream ends of the pipeline. Based on the analytical signal, the method for obtaining the local optimal solution of the IMF component using the improved linear adaptive weighted local approximation algorithm is as follows: Analyze the signal Divided into n segments, the IMF components are weighted using a weighting function and a dynamic weighting factor: ； ； in, This represents the weighted IMF component of segment x. For weighted functions, As a dynamic weighting factor, Let x be the segment of the analytic signal; The expression is as follows: ； Where d is the distance from the measurement domain to the signal center; The expression is as follows: ； Local approximation of the weighted IMF components: ； in, This represents the IMF component of the local approximation in segment x after the (p+1)th iteration. To approximate the speed, , Indicates the signal approximation value. ; when When the iteration stops, the local optimum of the IMF component is obtained: ； in, This represents a local optimum for an IMF component. This represents the optimal IMF component for the local approximation of the nth segment.

2. The infrasound localization method for minor pipeline leaks according to claim 1, characterized in that, The method for constructing an analytical signal based on IMF components is as follows: Perform Hilbert transform on the denoised IMF components of the signal: ； in, This represents the signal after the Hilbert transform. The signal is denoised to represent the IMF components after noise reduction, k is the number of IMF components in each layer, and t is time. According to the signal Constructing an analytic signal: ； in, For the analytic signal, j is the imaginary sign.

3. The infrasound localization method for minor pipeline leaks according to claim 1, characterized in that, The method for obtaining the time difference of infrasound reflection to the upstream and downstream ends of the pipeline by performing anti-attenuation dispersion compensation analysis based on generalized cross-correlation of the local optimal solution of the IMF components is as follows: The generalized cross-correlation function is obtained from the local optimal solution of the IMF components: ； in, Represents the generalized cross-correlation function. This represents the weighted function of generalized cross-correlation. For signal frequency, This represents the cross-power spectrum of the signals received by upstream and downstream sensors within the pipeline after processing. This represents the local optimal solution after processing the signal received by the upstream sensor. This represents the local optimal solution after the signal received by the downstream sensor has been processed. Calculate the inverse attenuation dispersion compensation coefficient of the signal characteristics based on the generalized cross-correlation function: ； ； ； ； in, This represents the anti-attenuation dispersion compensation coefficient. This represents the inverse attenuation dispersion compensation coefficient of the x-th segment of the IMF at time t. Let represent the amplitude compensation operator for the x-th segment of the IMF component at time t, and Q be a linear function of the attenuation factor as a function of frequency. This represents the phase compensation operator for the x-th segment of the IMF component at time t. , n is the number of segments in the IMF component, and t is time; The generalized cross-correlation function after anti-attenuation dispersion compensation is obtained based on the anti-attenuation dispersion compensation coefficient: ； in, This represents the generalized cross-correlation function after inverse attenuation dispersion compensation; Applying a Hilbert difference transform to the generalized cross-correlation function after inverse attenuation dispersion compensation changes the time delay value from peak position search to zero point detection: ； in, For the improved generalized cross-correlation function, This is the time difference between the reflection of the infrasound wave to the upstream and downstream ends of the pipeline.

4. The infrasound localization method for minor pipeline leaks according to claim 1, characterized in that, The expression for calculating the infrasound wave velocity based on the pipe characteristics is as follows: ； Where v is the wave speed of the infrasound. The compressibility coefficient of the transported medium. Where D is the density of the transported medium, E is the pipe diameter, E is the elastic modulus of the pipe, e is the pipe thickness, and C is the correction factor. The calculation formula is as follows: ； Where Z is the compressibility factor of the conveying medium, and P is the conveying pressure of the conveying medium.

5. The infrasound localization method for minor pipeline leaks according to claim 4, characterized in that, The conveying medium is dry air, and the compressibility factor of the conveying medium is... , where p is the absolute pressure of the gas.

6. The infrasound localization method for minor pipeline leaks according to claim 1, characterized in that, The formula for calculating the location of a pipeline leak using the time-difference positioning method is as follows: ； Where M is the location of the pipeline leak, l is the distance between the upstream and downstream sensors in the pipeline, and v is the wave velocity of the infrasound. This is the time difference between the reflection of the infrasound wave to the upstream and downstream ends of the pipeline.

Images