Valve inner leakage monitoring and positioning method

Abstract

The invention discloses a valve inner leakage monitoring and positioning method which comprises the following steps: S1, arranging sensors on a valve, and eliminating burst noise; s2, obtaining vibration signal data of each frame; s3, inner leakage noise of the valve is determined; s4, determining a fractal dimension; s5, constructing a valve inner leakage detection model; s6, judging whether internal leakage occurs or not by using the valve internal leakage detection model; s7, establishing a vibration propagation theoretical model of the pipe network; s8, arranging a sensor on the pipeline; s9, collecting a standard vibration signal; s10, establishing a vibration propagation energy dissipation model; and S11, the distance between the broadband vibration sensor and the inner leakage is obtained through a vibration propagation energy dissipation model, and positioning is achieved. According to the technology, the broadband vibration signal is used as a processing object, the inner leakage of the valve can be timely and accurately found, the position of the inner leakage point can be rapidly positioned, and the positioning technology is not limited to be applied to the valve and can also be applied to positive pressure equipment including a tank body, an infusion tube network and the like.

Description

Technical Field

[0001] The invention relates to the technical field of valve internal leakage monitoring, and in particular to a valve internal leakage monitoring and positioning method. Background Art

[0002] Timely and accurate detection of valve internal leakage and online monitoring of valve internal leakage status are of great significance to improving the safety factor and economy of pipeline system operation and ensuring normal production and economic operation.

[0003] The detection methods of valve internal leakage include vibration monitoring method, temperature monitoring method, flow monitoring method, air tightness test method, sealing test method and acoustic wave detection method. The vibration monitoring method determines internal leakage by detecting the vibration of the valve body, but it may be interfered by background noise. The temperature monitoring method relies on temperature changes to identify internal leakage, but it is easily affected by the environment. The flow monitoring method determines internal leakage by abnormal flow, but it is not sensitive to small leakage. The air tightness test method requires shutdown and has a long cycle. The sealing test method tests the sealing of the valve and cannot achieve real-time monitoring. The literature [Vacuum, 2004, 41 (2): 55-56] adopts an ultrasonic leak detection method, using three piezoelectric sensors connected in series to form a phased array. During the detection, the sensor is scanned toward the device to be tested. If there is an internal leakage in a certain part, the sensor can receive the ultrasonic wave of the internal leakage and convert it into an electrical signal. The direction where the sensor receives the strongest ultrasonic signal of the internal leakage is the internal leakage point. In summary, the above-mentioned technologies are difficult to achieve online monitoring of the internal leakage status of the valve. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a valve internal leakage monitoring and positioning method, which uses broadband vibration signals as processing objects and can not only detect valve internal leakage in a timely and accurate manner, but also quickly locate the position of the internal leakage point.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0006] A valve internal leakage monitoring and positioning method is provided, comprising the following steps:

[0007] S1. Arrange a plurality of broadband vibration sensors and a vibration exciter for generating vibration signals on the valve housing, collect vibration signals on the pipeline housing simultaneously through the plurality of broadband vibration sensors, and remove burst noise from the collected vibration signals;

[0008] S2, performing frame processing on the vibration signal after removing the burst noise, determining the data amount of each frame, and obtaining vibration signal data of each frame;

[0009] S3. Determine the valve vibration signal and its corresponding frequency range according to each frame of vibration signal data, and screen out suspected internal leakage vibration signals;

[0010] S4, using a radius dimension algorithm to quantitatively determine the fractal dimension of the suspected endoleak vibration signal, and determine whether the suspected endoleak vibration signal is a true endoleak vibration signal;

[0011] S5. According to the fractal dimension and the frequency distribution range of the suspected internal leakage vibration signal, a valve internal leakage detection model is constructed, and the valve internal leakage detection model is trained using the real internal leakage vibration signal, and the trained valve internal leakage detection model is output;

[0012] S6, inputting the collected pipeline shell vibration signal into the trained valve internal leakage detection model to make a decision and determine whether the valve has internal leakage;

[0013] S7. Establish a vibration propagation theoretical model of the pipe network according to the pipe network parameters;

[0014] S8, arranging a plurality of broadband vibration sensors and a vibration exciter outside the pipeline, wherein the plurality of broadband vibration sensors are arranged in a star network, and the vibration exciter is arranged at the center of the star network;

[0015] S9, collecting a standard vibration signal using a broadband vibration sensor, and generating a standard vibration signal through a vibration exciter;

[0016] S10, extracting characteristic parameters from the standard vibration signal and combining them with the vibration propagation theoretical model to establish a vibration propagation energy dissipation model for vibration propagation in the pipeline;

[0017] S11. The standard vibration signal of the valve is collected by several wide-band vibration sensors, and the propagation distance of the suspected internal leakage vibration signal is calculated in combination with the vibration propagation energy dissipation model to locate the internal leakage point.

[0018] Furthermore, the specific steps of removing the burst noise in step S1 are as follows:

[0019] S11. Define the acquired vibration signal as x i (n), calculate the vibration signal x i The standard deviation of the amplitude of (n):

[0020]

[0021] Among them, σ A is the amplitude standard deviation, N′ is the number of vibration signal sampling points, A i is the amplitude value of the i-th sampling point, μ′ is the average amplitude value;

[0022] S12, take 4 times the amplitude standard deviation as the rejection threshold 4σ A ;

[0023] S13, extract the vibration signal in segments, calculate the amplitude standard deviation of each vibration signal, and compare the amplitude standard deviation of each vibration signal with the rejection threshold in segments; if σ A >4σ A , then the vibration signal is a burst noise, and the vibration signal is removed; if σ A ≤4σ A , then the vibration signal is not a burst noise, and the vibration signal is retained.

[0024] Furthermore, step S3 specifically includes the following steps:

[0025] S31, divide each frame of vibration signal data into N frequency segments according to the frequency, obtain N frequency distribution ranges, and obtain signal components x in N different frequency ranges. ij (n):

[0026] x ij (n)[j=1:N];

[0027] Wherein, i represents the number of the i-th broadband vibration sensor, and j represents the frequency band;

[0028] S32, using the autocorrelation signal processing method to process the signal components x within the N frequency range ij (n) is processed to calculate the autocorrelation coefficients corresponding to the signal components of N frequency bands. The autocorrelation coefficient calculation formula is as follows:

[0029] Rx ij (τ)=E{x ij (n)x ij (n-τ)};

[0030] Where R represents the autocorrelation function, x ij (τ) represents the signal component x ij (n) is the signal sequence relationship between them, τ represents the delay, E is the expected calculation, x ij (n-τ) is the delayed signal sequence, and n represents the discrete sampling interval of the signal;

[0031] S33, judging whether there is a suspected internal leakage vibration signal in the vibration signals corresponding to the N frequency distribution ranges according to the autocorrelation coefficients corresponding to the signal components of the N frequency segments;

[0032] If the autocorrelation coefficient is greater than or equal to the rejection threshold, there is no suspected internal leakage vibration signal in the frequency range;

[0033] If the autocorrelation coefficient is less than the rejection threshold, there is a suspected internal leakage vibration signal in the frequency range, and step S4 is executed.

[0034] Furthermore, step S4 specifically includes the following steps:

[0035] The suspected endoleak vibration signal is covered by gradually decreasing the box size ε, and the initial value of the box is set to the maximum value. The maximum value covers the distribution area of ​​the suspected endoleak vibration signal, and then the box size ε is gradually reduced to the minimum value. The signal range is divided into grids at each box size, and the minimum number of boxes N(ε) containing the signal points is calculated, and then the fractal dimension D is calculated:

[0036]

[0037] If the measured radius dimension is smaller, it indicates that the degree of centripetal aggregation of the signal is stronger and the signal change is more regular, then it is judged as a true endoleak vibration signal; if the measured radius dimension is larger and the amplitude layout is more dispersed, then it is judged as a non-endoleak vibration signal.

[0038] Furthermore, step S5 specifically includes the following steps:

[0039] S51, taking the fractal dimension obtained in step S4 and the frequency distribution range of the suspected endoleak vibration signal extracted in step S3 as two input feature parameters;

[0040] S52. Select a genetic algorithm as a decision algorithm, combine the confirmed real valve internal leakage signal as a training sample, and build a valve internal leakage detection model.

[0041] Furthermore, in step S6, the method for the valve internal leakage detection model to determine whether the suspected internal leakage vibration signal is an internal leakage signal is as follows:

[0042] If the valve internal leakage detection model determines that the signal is an internal leakage vibration signal, the next step is to locate the leakage point. If the valve internal leakage detection model determines that the signal is a non-internal leakage vibration signal, there is no leakage in the valve; thereby realizing the internal leakage identification and monitoring of the collection pipeline.

[0043] Furthermore, the vibration propagation theoretical model of step S7 is as follows:

[0044]

[0045] Among them, u(x, t) is the vibration displacement, η is the damping coefficient, c is the propagation velocity, F (x, t) is the excitation force.

[0046] Furthermore, step S10 specifically includes the following steps:

[0047] A1001. Extract characteristic parameters of the collected standard vibration signal;

[0048]

[0049] Wherein, μ is the time domain mean of the standard vibration signal, T is the total sampling time of the standard vibration signal, x(t) is the time domain signal amplitude of the standard vibration signal, X(f) is the frequency domain amplitude of the standard vibration signal at frequency f, N is the number of sampling points, x(n) is the discrete time domain signal amplitude, and n is the total number of sampling points;

[0050] A1002. Combined with the vibration propagation theoretical model of the pipeline network, the dissipation characteristics of energy in the propagation process are analyzed, and the pipeline vibration propagation energy dissipation model is established:

[0051] E dissipated (d) = E total ·e -αd ;

[0052] Among them, E dissipated (d) is the energy of the standard vibration signal after it propagates to a distance d, d is the propagation distance of the suspected internal leakage vibration signal, E total is the total energy of the suspected internal leakage vibration signal, α is the attenuation coefficient, and generally the high frequency part corresponds to a larger α value;

[0053] The vibration propagation energy dissipation model can be expressed as the law that the amplitude of the vibration signal decays exponentially with the propagation distance. It is necessary to analyze the propagation characteristics and dissipation behavior of the high-frequency component in particular. For the high-frequency component, by analyzing its attenuation rate separately, the relationship between the frequency domain amplitude X(f) and the attenuation coefficient α is obtained:

[0054]

[0055] Among them, γ is a constant related to the pipe material and fluid characteristics, c is the propagation velocity, and α(f) is the attenuation coefficient at different frequencies.

[0056] The beneficial effects of the present invention are:

[0057] The valve internal leakage monitoring and positioning method of the present invention uses broadband vibration signals as processing objects, which can not only detect valve internal leakage in a timely and accurate manner, but also quickly locate the position of the internal leakage point. Moreover, this positioning technology is not limited to valves, but can also be used for positive pressure equipment, including tanks, infusion pipelines, etc.

[0058] The monitoring method proposed by the present invention requires little quantity, has a short monitoring time and is highly efficient. The technology of using blind system identification to determine the location of the leak is suitable for pipelines with simple structures and has high requirements for the signal-to-noise ratio. The method of this patent is suitable for complex pipe network structures and complex interference environments, has low requirements for the signal-to-noise ratio and has higher environmental adaptability. The methods of Elman neural network and wavelet packet identification of leak occurrence require a large amount of data, are relatively time-consuming, and have a long monitoring time. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1A flow chart of valve internal leakage detection of the valve internal leakage monitoring and positioning method of the present invention;

[0060] Figure 2 A flow chart of valve internal leakage positioning of the valve internal leakage monitoring and positioning method of the present invention;

[0061] Figure 3 This is a comparison chart of the pump noise and valve internal leakage signal identification results. DETAILED DESCRIPTION

[0062] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.

[0063] A broadband vibration sensor is attached to the outer wall of the pipe where the valve exists. Through the self-adjusting circuit and A / D conversion, a digital signal reflecting the current state of the valve in the pipe network is obtained. There may be several types of components in the digital signal: 1. If there is an internal leakage in the valve, the internal leakage vibration sound signal caused by it; 2. The vibration sound noise in the environment where the valve is located, such as various pumps; 3. The circuit noise of the detection equipment, etc. Therefore, the purpose of valve internal leakage detection is to clearly indicate whether the valve has an internal leakage or not by analyzing the obtained digital signal. However, the differences in the structures of different valves make the internal leakage sound vibration signal itself have differences in amplitude, frequency and other characteristics, and due to the existence of various types of noise, it is difficult to accurately detect valve internal leakage using conventional time-frequency feature extraction methods.

[0064] Attach the broadband vibration sensor to the outer wall of the pipeline, and assume that the sensor obtains the signal model as follows:

[0065] x(n)=s(n)+n(n)

[0066] Among them, x(n) represents the independent observation of the unknown internal leakage source signal, s(n), and n(n) represents the additive noise in the observation, including the valve network environment noise and circuit noise, which are unrelated to each other. According to the theory of gas dynamic acoustics, when the valve is in an internal leakage state, the high-speed airflow flowing into the front end of the internal leakage valve forms a turbulent jet and flows out from the rear end of the valve. The internal leakage causes irregular movement of the gas and excites a large number of acoustic wave signals. This signal belongs to jet noise. The acoustic wave signal is coupled to the valve housing and propagates around. According to the above analysis, since the source of the signal is caused by the high-speed turbulent jet and the turbulent process is random, the signal s(n) is also random, which is the most important feature.

[0067] like Figure 1and 2 According to the above characteristics, a valve internal leakage monitoring and positioning method is proposed, comprising the following steps:

[0068] S1. Arrange a plurality of broadband vibration sensors and a vibration exciter for generating vibration signals on the valve housing, collect vibration signals on the pipeline housing simultaneously through the broadband vibration sensors, and remove burst noise from the collected vibration signals; the specific steps of removing burst noise are as follows:

[0069] S11. Define the acquired vibration signal as x i (n), calculate the vibration signal x i The standard deviation of the amplitude of (n):

[0070]

[0071] Among them, σ A is the amplitude standard deviation, N′ is the number of vibration signal sampling points, A i is the amplitude value of the i-th sampling point, μ′ is the average amplitude value;

[0072] S12, take 4 times the amplitude standard deviation as the rejection threshold 4σ A ;

[0073] S13, extract the vibration signal in segments, calculate the amplitude standard deviation of each vibration signal, and compare the amplitude standard deviation of each vibration signal with the rejection threshold in segments; if σ A >4σ A , then the vibration signal is a burst noise, and the vibration signal is removed; if σ A ≤4σ A , then the vibration signal is not a burst noise, and the vibration signal is retained; the burst noise elimination method has the advantages of being fast, small in computational complexity, and being able to accurately determine the burst noise;

[0074] S2, performing frame processing on the vibration signal after removing the burst noise, determining the data amount of each frame, and obtaining the vibration signal data of each frame; specifically:

[0075] Assume that the sampling frequency of the vibration signal is fs. The internal leakage sound signal is a continuous process caused by turbulence. The sampling data of 5fs is sufficient to include the process of sound generated by a single turbulence. Therefore, 5fs is used as the data sampling period, that is, 5fs is used as the data size of each frame, so as to perform frame processing on the collected vibration signal.

[0076] S3. Determine the vibration signal and its corresponding frequency range according to each frame of vibration signal data, and screen out suspected internal leakage vibration signals; specifically:

[0077] S31, divide each frame of vibration signal data into N frequency segments according to the frequency, obtain N frequency distribution ranges, and obtain signal components x in N different frequency ranges. ij (n):

[0078] x ij (n)[j=1:N];

[0079] Wherein, i represents the number of the i-th broadband vibration sensor, and j represents the frequency band;

[0080] S32, using the autocorrelation signal processing method to process the signal components x within the N frequency range ij (n) is processed to calculate the autocorrelation coefficients corresponding to the signal components of N frequency bands. The autocorrelation coefficient calculation formula is as follows:

[0081] Rx ij (τ)=E{x ij (n)x ij (n-τ)};

[0082] Where R represents the autocorrelation function, x ij (τ) represents the signal component x ij (n) is the signal sequence relationship between them, τ represents the delay, E is the expected calculation, x ij (n-τ) is the delayed signal sequence, and n represents the discrete sampling interval of the signal;

[0083] S33, judging whether there is a suspected internal leakage vibration signal in the vibration signals corresponding to the N frequency distribution ranges according to the autocorrelation coefficients corresponding to the signal components of the N frequency segments;

[0084] If the autocorrelation coefficient is greater than or equal to the rejection threshold, there is no suspected internal leakage vibration signal in the frequency range;

[0085] If the autocorrelation coefficient is less than the rejection threshold, there is a suspected internal leakage vibration signal in the frequency range, and step S4 is executed; wherein the rejection threshold can be determined by multiple statistical methods, thereby determining the suspected internal leakage signal in each frame and its frequency range;

[0086] S4. Using a radius dimension algorithm to quantitatively determine the fractal dimension of the suspected endoleak vibration signal, and determine whether the suspected endoleak vibration signal is a true endoleak vibration signal; specifically:

[0087] The suspected endoleak vibration signal is covered by gradually decreasing the box size ε, and the initial value of the box is set to the maximum value. The maximum value covers the distribution area of ​​the suspected endoleak vibration signal, and then the box size ε is gradually reduced to the minimum value. The signal range is divided into grids at each box size, and the minimum number of boxes N(ε) containing the signal points is calculated, and then the fractal dimension D is calculated:

[0088]

[0089] If the measured radius dimension is smaller, it indicates that the centripetal aggregation of the signal is stronger and the signal changes more regularly, then it is determined to be a true endoleak vibration signal; if the measured radius dimension is larger, the amplitude layout is more dispersed, then it is determined to be a non-endoleak vibration signal;

[0090] S5. According to the fractal dimension and the frequency distribution range of the suspected internal leakage vibration signal, a valve internal leakage detection model is constructed, and the valve internal leakage detection model is trained using the real internal leakage vibration signal, and the trained valve internal leakage detection model is output; specifically:

[0091] S51, taking the fractal dimension obtained in step S4 and the frequency distribution range of the suspected endoleak vibration signal extracted in step S3 as two input feature parameters;

[0092] S52, selecting a genetic algorithm as a decision algorithm, combining the confirmed real valve internal leakage signal as a training sample, and constructing a valve internal leakage detection model, the valve internal leakage detection model is used to make a decision on the input features of the fractal dimension and the frequency distribution range;

[0093] S6. Input the collected pipeline shell vibration signal into the trained valve internal leakage detection model to make a decision and determine whether the valve has internal leakage; specifically, if the valve internal leakage detection model determines that it is an internal leakage vibration signal, the next step of leak point location is performed; if the valve internal leakage detection model determines that it is a non-internal leakage vibration signal, the valve does not have leakage; thereby realizing the internal leakage identification and monitoring of the collection pipeline;

[0094] S7. Based on the pipe network parameters, such as size, material, etc., combined with the classical structural vibration propagation theory, and considering the effects of damping and fluid coupling, a vibration propagation theoretical model of the pipe network is established; specifically, the vibration propagation theoretical model is as follows:

[0095]

[0096] Among them, u(x, t) is the vibration displacement, η is the damping coefficient, c is the propagation speed, F (x, t) is the excitation force, and this vibration propagation theoretical model is used to reflect the energy characteristics and frequency characteristics;

[0097] S8, arranging a plurality of broadband vibration sensors and a vibration exciter outside the pipeline, wherein the plurality of broadband vibration sensors are arranged in a star network, and the vibration exciter is arranged at the center of the star network;

[0098] S9, collecting a standard vibration signal using a broadband vibration sensor, and generating a standard vibration signal through a vibration exciter;

[0099] S10. Extract characteristic parameters from the standard vibration signal and combine them with the vibration propagation theoretical model to establish a vibration propagation energy dissipation model for vibration propagation in the pipeline; specifically:

[0100] A1001. Extract characteristic parameters of the collected standard vibration signal:

[0101]

[0102] Wherein, μ is the time domain mean of the standard vibration signal, T is the total sampling time of the standard vibration signal, x(t) is the time domain signal amplitude of the standard vibration signal, X(f) is the frequency domain amplitude of the standard vibration signal at frequency f, N is the number of sampling points, x(n) is the discrete time domain signal amplitude, and n is the total number of sampling points;

[0103] A1002. Combined with the vibration propagation theoretical model of the pipeline network, the dissipation characteristics of energy in the propagation process are analyzed, and the pipeline vibration propagation energy dissipation model is established:

[0104] E dissipated (d) = E total ·e -αd ;

[0105] Among them, E dissipated (d) is the energy of the standard vibration signal after it propagates to a distance d, d is the propagation distance of the suspected internal leakage vibration signal, E total is the total energy of the suspected internal leakage vibration signal, α is the attenuation coefficient, and generally the high frequency part corresponds to a larger α value;

[0106] The vibration propagation energy dissipation model can be expressed as the law that the amplitude of the vibration signal decays exponentially with the propagation distance. It is necessary to analyze the propagation characteristics and dissipation behavior of the high-frequency component in particular. For the high-frequency component, by analyzing its attenuation rate separately, the relationship between the frequency domain amplitude X(f) and the attenuation coefficient α is obtained:

[0107]

[0108] Among them, γ is a constant related to the pipe material and fluid characteristics, c is the propagation velocity, and α(f) is the attenuation coefficient at different frequencies;

[0109] S11. The standard vibration signal of the valve is acquired by a number of wide-band vibration sensors, and the frequency domain amplitude X(f) of the standard vibration signal is calculated. The attenuation coefficient α is calculated and substituted into the vibration propagation energy dissipation model to obtain the propagation distance of each of the wide-band vibration sensors from the internal leakage point, thereby obtaining the three-dimensional propagation characteristics of the vibration on the valve and locating the internal leakage point. Figure 2 The distributed sensors in the are broadband vibration sensors.

[0110] The energy dissipation characteristics of the characteristic vibration signal generated by the internal leakage point at the valve are compared with the energy dissipation characteristics in the vibration propagation energy dissipation model. By analyzing the attenuation characteristics, frequency components and amplitude changes of the vibration signal, combined with the corresponding parameters of the signal obtained in the actual sensor, and combining the influence of the internal leakage point position on signal propagation, the internal leakage point can be accurately located.

[0111] Combining the dissipation characteristics with the vibration propagation model on the valve, the 3D propagation characteristics of the vibration on the valve are obtained, thereby locating the internal leakage point.

[0112] The method of using Elman neural network and wavelet packet to identify leakage requires a large amount of data, so it is time-consuming and takes a long time to monitor; while the monitoring method proposed by the present invention does not require a large amount of data, has a short monitoring time and is highly efficient. The technology of using blind system identification to determine the location of the leak is suitable for pipelines with simple structures and has high requirements for signal-to-noise ratio. The method of this patent is suitable for complex pipe network structures and complex interference environments, has low requirements for signal-to-noise ratio, and has higher environmental adaptability.

[0113] Figure 3 The pump noise and internal leakage signal identification results in the valve environment were statistically analyzed. Figure 3 It can be seen that the valve internal leakage missed detection rates of LPCC+HMM and AR+SVM are 25.0% and 22.5% respectively, which are higher than the missed detection rate of 0% of the present invention; the internal leakage false detection rates of LPCC+HMM and AR+SVM are 17.5% and 20.0% respectively, which are higher than the false detection rate of 5.0% of this patent.

Claims

1. A valve internal leakage monitoring and positioning method, characterized in that: The steps include: S1. Arrange a plurality of broadband vibration sensors and a vibration exciter for generating vibration signals on the valve housing, collect vibration signals on the pipeline housing simultaneously through the plurality of broadband vibration sensors, and remove burst noise from the collected vibration signals; S2, performing frame processing on the vibration signal after removing the burst noise, determining the data amount of each frame, and obtaining vibration signal data of each frame; S3. Determine the valve vibration signal and its corresponding frequency range according to each frame of vibration signal data, and screen out suspected internal leakage vibration signals; S4, using a radius dimension algorithm to quantitatively determine the fractal dimension of the suspected endoleak vibration signal, and determine whether the suspected endoleak vibration signal is a true endoleak vibration signal; S5. According to the fractal dimension and the frequency distribution range of the suspected internal leakage vibration signal, a valve internal leakage detection model is constructed, the valve internal leakage detection model is trained using the real internal leakage vibration signal, and the trained valve internal leakage detection model is output; S6, inputting the collected pipeline shell vibration signal into the trained valve internal leakage detection model to make a decision and determine whether the valve has internal leakage; S7. Establish a vibration propagation theoretical model of the pipe network according to the pipe network parameters; S8, arranging a plurality of broadband vibration sensors and a vibration exciter outside the pipeline, wherein the plurality of broadband vibration sensors are arranged in a star network, and the vibration exciter is arranged at the center of the star network; S9, collecting a standard vibration signal using a broadband vibration sensor, and generating a standard vibration signal through a vibration exciter; S10, extracting characteristic parameters from the standard vibration signal and combining them with the vibration propagation theoretical model to establish a vibration propagation energy dissipation model for vibration propagation in the pipeline; S11. The standard vibration signal of the valve is collected by several wide-band vibration sensors, and the propagation distance of the suspected internal leakage vibration signal is calculated in combination with the vibration propagation energy dissipation model to locate the internal leakage point.

2. The valve internal leakage monitoring and positioning method according to claim 1 is characterized in that: The specific steps of burst noise removal in step S1 are as follows: S11. Define the acquired vibration signal as x i (n), calculate the vibration signal x i The standard deviation of the amplitude of (n): Among them, σ A is the amplitude standard deviation, N′ is the number of vibration signal sampling points, A i is the amplitude value of the i-th sampling point, μ′ is the average amplitude value; S12, take 4 times the amplitude standard deviation as the rejection threshold 4σ A ; S13, extract the vibration signal in segments, calculate the amplitude standard deviation of each vibration signal, and compare the amplitude standard deviation of each vibration signal with the rejection threshold in segments; if σ A >4σ A , then the vibration signal is a burst noise, and the vibration signal is removed; if σ A ≤4σ A , then the vibration signal is not a burst noise, and the vibration signal is retained.

3. The valve internal leakage monitoring and positioning method according to claim 1 is characterized in that: Step S3 specifically includes the following steps: S31, divide each frame of vibration signal data into N frequency segments according to the frequency, obtain N frequency distribution ranges, and obtain signal components x in N different frequency ranges. ij (n): x ij (n)[j=1:N]? Wherein, i represents the number of the i-th broadband vibration sensor, and j represents the frequency band; S32, using the autocorrelation signal processing method to process the signal components x within the N frequency range ij (n) is processed to calculate the autocorrelation coefficients corresponding to the signal components of N frequency bands. The autocorrelation coefficient calculation formula is as follows: Rx ij (τ)=E{x ij (n)x ij (n-τ)}; Where R represents the autocorrelation function, x ij (τ) represents the signal component x ij (n) is the signal sequence relationship between them, τ represents the delay, E is the expected calculation, x ij (n-τ) is the delayed signal sequence, and n represents the discrete sampling interval of the signal; S33, judging whether there is a suspected internal leakage vibration signal in the vibration signals corresponding to the N frequency distribution ranges according to the autocorrelation coefficients corresponding to the signal components of the N frequency segments; If the autocorrelation coefficient is greater than or equal to the rejection threshold, there is no suspected internal leakage vibration signal in the frequency range; If the autocorrelation coefficient is less than the rejection threshold, there is a suspected internal leakage vibration signal in the frequency range, and step S4 is executed.

4. The valve internal leakage monitoring and positioning method according to claim 1 is characterized in that: Step S4 specifically includes the following steps: The suspected endoleak vibration signal is covered by gradually decreasing the box size ε, and the initial value of the box is set to the maximum value. The maximum value covers the distribution area of ​​the suspected endoleak vibration signal, and then the box size ε is gradually reduced to the minimum value. The signal range is divided into grids at each box size, and the minimum number of boxes N(ε) containing the signal points is calculated, and then the fractal dimension D is calculated: If the measured radius dimension is smaller, it indicates that the degree of centripetal aggregation of the signal is stronger and the signal change is more regular, then it is judged as a true endoleak vibration signal; if the measured radius dimension is larger and the amplitude layout is more dispersed, then it is judged as a non-endoleak vibration signal.

5. The valve internal leakage monitoring and positioning method according to claim 1 is characterized in that: Step S5 specifically includes the following steps: S51, taking the fractal dimension obtained in step S4 and the frequency distribution range of the suspected endoleak vibration signal extracted in step S3 as two input feature parameters; S52. Select a genetic algorithm as a decision algorithm, combine the confirmed real valve internal leakage signal as a training sample, and build a valve internal leakage detection model.

6. The valve internal leakage monitoring and positioning method according to claim 1 is characterized in that: In step S6, the method for the valve internal leakage detection model to determine whether the suspected internal leakage vibration signal is an internal leakage signal is as follows: If the valve internal leakage detection model determines that the signal is an internal leakage vibration signal, the next step is to locate the leakage point. If the valve internal leakage detection model determines that the signal is a non-internal leakage vibration signal, there is no leakage in the valve; thereby realizing the internal leakage identification and monitoring of the collection pipeline.

7. The valve internal leakage monitoring and positioning method according to claim 1 is characterized in that: The vibration propagation theoretical model of step S7 is as follows: Among them, u(x, t) is the vibration displacement, η is the damping coefficient, c is the propagation velocity, F (x, t) is the excitation force.

8. The valve internal leakage monitoring and positioning method according to claim 1 is characterized in that: Step S10 specifically includes the following steps: A1001. Extract characteristic parameters of the collected standard vibration signal; Wherein, μ is the time domain mean of the standard vibration signal, T is the total sampling time of the standard vibration signal, x(t) is the time domain signal amplitude of the standard vibration signal, X(f) is the frequency domain amplitude of the standard vibration signal at frequency f, N is the number of sampling points, x(n) is the discrete time domain signal amplitude, and n is the total number of sampling points; A1002. Combined with the vibration propagation theoretical model of the pipeline network, the dissipation characteristics of energy in the propagation process are analyzed, and the pipeline vibration propagation energy dissipation model is established: Edissipated(d)=Etotal·e-αd; Among them, E dissipated (d) is the energy of the standard vibration signal after it propagates to a distance d, d is the propagation distance of the suspected internal leakage vibration signal, E total is the total energy of the suspected internal leakage vibration signal, α is the attenuation coefficient, and generally the high frequency part corresponds to a larger α value; The vibration propagation energy dissipation model can be expressed as the law that the amplitude of the vibration signal decays exponentially with the propagation distance. It is necessary to analyze the propagation characteristics and dissipation behavior of the high-frequency component in particular. For the high-frequency component, by analyzing its attenuation rate separately, the relationship between the frequency domain amplitude X(f) and the attenuation coefficient α is obtained: Among them, γ is a constant related to the pipe material and fluid characteristics, c is the propagation velocity, and α(f) is the attenuation coefficient at different frequencies.

Images