A method and system for leak detection in valves

Abstract

This application relates to the field of valve leakage detection technology, specifically to a valve leakage detection method and system. The method includes: acquiring acoustic emission data and valve flow velocity from the valve; reconstructing the acoustic emission data to obtain reconstructed emission data for each monitoring period; acquiring the acoustic emission signal fluctuation irregularity coefficient and the acoustic emission signal frequency distribution disorder, thereby obtaining the valve leakage determination coefficient for each monitoring period; obtaining the leakage status assessment value for each period based on the correlation between the valve flow velocity and the valve leakage determination coefficient; and using a neural network to detect valve leakage based on the distribution characteristics of the reconstructed acoustic emission data for each period. This application can improve the accuracy of valve leakage detection.

Description

Technical Field

[0001] This application relates to the field of valve leakage detection technology, specifically to a leakage detection method and system for valves. Background Technology

[0002] Valves are devices used to control fluid pipelines. Their main function is to manage the medium flowing in the pipeline. They often operate in environments with high temperature, high pressure, and strong corrosion, making them highly susceptible to leakage. Valve leakage is generally divided into two categories: external leakage and internal leakage. External leakage is visible on the outside of the valve, making it relatively easy to detect and identify. In contrast, internal leakage occurs inside the valve and is difficult to detect directly.

[0003] Acoustic emission detection technology offers high accuracy and low cost, making it suitable for leak detection. However, in actual testing, the acquisition of acoustic emission signals is easily affected by external environmental vibrations, which can influence the accurate identification of valve leak characteristics. Conventional detection methods, failing to adequately consider these interference factors, result in inaccurate valve leak detection.

[0004] Publication No. CN102323013B discloses a valve leakage detection device that uses an acoustic sensor to collect acoustic leakage signals from a valve. After signal processing, fast Fourier transform, and spectral analysis, the signal is compared with a stored non-leakage spectral image to determine the leakage status. However, this method fails to adequately consider interference from external environmental vibrations and turbulent noise during signal acquisition, resulting in inaccurate valve leakage detection. Summary of the Invention

[0005] To address the aforementioned technical problems, the purpose of this application is to provide a leakage detection method and system for valves, and the specific technical solution adopted is as follows:

[0006] This application provides a leakage detection method for valves, including the following steps:

[0007] Acquire acoustic emission data and valve flow rate from the valve;

[0008] Modal decomposition was performed on the acoustic emission data within each monitoring period. Effective modal components were extracted based on the correlation between the modal components and the acoustic emission data. The reconstructed emission data within each monitoring period was then obtained.

[0009] The distribution fluctuation degree and the instability characteristics of the fluctuation time of the reconstructed acoustic emission data are analyzed to obtain the irregularity coefficient of acoustic emission signal fluctuation within each monitoring period. Based on the degree of disorder of the amplitude distribution of the reconstructed acoustic emission data in the frequency domain, and combined with the amplitude difference characteristics of the reconstructed acoustic emission data and the acoustic emission data of the valve in the non-leakage state in the frequency domain, the frequency distribution disorder of the acoustic emission signal within each monitoring period is obtained, and then the valve leakage judgment coefficient within each monitoring period is obtained.

[0010] Based on the correlation between the valve flow rate and the valve leakage judgment coefficient in each time period, the leakage status assessment value of each time period is obtained. Combined with the distribution characteristics of the reconstructed acoustic emission data in each time period, a neural network is used to detect valve leakage.

[0011] Preferably, the Pearson correlation coefficient between each modal component and the acoustic emission data within each monitoring period is calculated, and all Pearson correlation coefficients are clustered. The modal component corresponding to the cluster with the largest central value is taken as the effective modal component within each monitoring period.

[0012] Preferably, curve fitting is performed on the reconstructed acoustic emission data within each monitoring period, and the dispersion of the time difference corresponding to all two adjacent maxima and the dispersion of the time difference corresponding to all two adjacent minima on the fitted curve are statistically analyzed. The sum of the two dispersions is used as the instability coefficient of the fluctuation period of the reconstructed acoustic emission data within each monitoring period.

[0013] Preferably, the acquisition of the irregularity coefficient of acoustic emission signal fluctuation within each monitoring period is further described as follows: ,in, Let be the standard deviation of the amplitude difference between all adjacent maxima and minima on the fitted curve corresponding to the reconstructed acoustic emission data within the i-th monitoring duration. It is the instability coefficient of the reconstructed acoustic emission data fluctuation period within the i-th monitoring duration.

[0014] Preferably, the acquisition of the disorder of the frequency distribution of acoustic emission signals within each monitoring period is further as follows:

[0015] ,in, Let i be the disorder of the frequency distribution of the acoustic emission signal within the i-th monitoring period. Let fractal dimension be the amplitude corresponding to all frequencies in the reconstructed acoustic emission data spectrum for the i-th monitoring duration. The sum of all frequency amplitudes in the reconstructed acoustic emission data spectrum for the i-th monitoring duration is the ratio of the sum of all frequency amplitudes in the spectrum corresponding to the non-leakage state. The corresponding spectrum can be obtained by performing a frequency domain transformation on the reconstructed acoustic emission data.

[0016] Preferably, the valve leakage determination coefficient within each monitoring period is the product of the irregularity coefficient of acoustic emission signal fluctuation within each monitoring period and the disorder of acoustic emission signal frequency distribution.

[0017] Preferably, a time period includes multiple monitoring durations, and the average flow rate of the valve at each monitoring duration within any time period is arranged in ascending order to form a flow rate sequence for that time period.

[0018] Preferably, the leakage status assessment values ​​for each time period are further obtained as follows:

[0019] Calculate the correlation coefficient between the flow velocity sequence for each time period and all leakage judgment coefficients within each time period. Then, multiply the correlation coefficient by the average value of all valve leakage judgment coefficients within each time period to obtain the leakage status assessment value for each time period.

[0020] Preferably, the distribution characteristics of the reconstructed acoustic emission data in each time period include the mean, standard deviation, kurtosis, peak factor, margin factor, and leakage status assessment value of the reconstructed acoustic emission data. The reconstructed acoustic emission data and its distribution characteristics in each time period are used as input to the neural network to obtain the valve leakage detection results.

[0021] This application also provides a valve-oriented leakage detection system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any of the above-described valve-oriented leakage detection methods.

[0022] As can be seen from the above, the valve leakage detection method and system provided in this application have at least the following beneficial effects:

[0023] This application utilizes the correlation characteristics between noise signals and effective acoustic emission signals and the acquired valve acoustic emission data for denoising analysis. Its advantage lies in its ability to effectively remove noise from the valve acoustic emission data and reduce its interference. Simultaneously, this application deeply analyzes the irregularity of the fluctuations and cycles of the acoustic emission signals in the time domain, as well as the chaotic distribution and richness characteristics in the frequency domain. It also considers the degree to which leakage anomalies are affected by fluid flow velocity to determine the likelihood of valve leakage. The application uses an LSTM network model for valve leakage detection, which helps improve the accuracy of valve leakage detection. Attached Figure Description

[0024] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This application provides a flowchart of the steps for a valve leakage detection method. Detailed Implementation

[0026] To further illustrate the technical means and effects adopted by this application to achieve the intended inventive purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a valve leakage detection method and system proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0027] Unless otherwise specified and limited, terms such as “comprising,” “including,” or any other variations thereof are intended to cover a non-exclusive inclusion, such that a circuit structure, article, or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the article or device that includes said element. Furthermore, the term “and / or” as used herein includes any and all combinations of one or more of the associated listed items. All technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0028] The following description, in conjunction with the accompanying drawings, details a specific scheme for a valve leakage detection method and system provided in this application.

[0029] Please see Figure 1 The diagram illustrates a flowchart of a valve leakage detection method according to an embodiment of this application, including the following steps:

[0030] Step 1: Obtain the acoustic emission data of the valve and the valve flow rate.

[0031] Acoustic emission testing (AE) is a non-destructive testing method used to monitor and diagnose minute cracks, defects, stress changes, and other anomalies in materials or structures. When a valve leaks, the stress generated by fluid flow causes changes in the valve's internal structure, such as crack propagation or minute displacements. These stress-induced internal changes trigger the release of energy, which propagates as tiny sound waves. Therefore, this embodiment utilizes acoustic emission technology to detect valve leaks.

[0032] Typically, when a valve malfunctions, the resulting sound waves propagate along the flow direction of the pipe or equipment. Therefore, in this embodiment, an acoustic emission sensor is installed near the downstream position of the valve to collect the acoustic emission signal. This allows for better reception of the sound signal propagating from the fault point. The data acquisition frequency is set to 1kHz to obtain real-time monitored acoustic emission data.

[0033] Step 2: Perform modal decomposition on the acoustic emission data within each monitoring period, extract the effective modal components based on the correlation between the modal components and the acoustic emission data, and reconstruct the emission data.

[0034] When a valve leaks, it generates abnormal acoustic emission signals. These signals are primarily caused by several factors, such as cavitation sound generated by the collapse of air bubbles back into liquid under high pressure, interference and impact from the internal medium on certain parts of the valve causing vibration, and broadband noise generated by turbulent flow. In actual testing, interference from external environmental vibrations or noise can occur. Conventional methods typically compare the spectrum of the measured acoustic emission signal with that of a leak-free signal for leak detection, which can lead to misjudgments under interference. Therefore, a deeper analysis of the acoustic emission signal characteristics of a leaking valve is necessary for leak detection.

[0035] The acoustic emission signal generated after valve leakage changes extremely rapidly. To accurately monitor the acoustic emission signal status, this application sets each monitoring duration to 1 second. Firstly, the collected acoustic emission data may be affected by external environmental noise, impacting the accuracy of subsequent feature extraction. Therefore, this embodiment first performs denoising processing on the collected acoustic emission data. However, due to varying noise levels in different environments, conventional denoising techniques struggle to retain effective information while effectively suppressing noise. The correlation between noise signals and actual acoustic emission signals in the valve's collected acoustic emission signals is weak; therefore, noise removal from the valve's acoustic emission data can be achieved.

[0036] Therefore, this embodiment uses the i-th monitoring duration as an example. Preferably, the CEEMDAN (Fully Adaptive Ensemble Empirical Mode Decomposition) algorithm is used to perform mode decomposition on the acoustic emission data, and the maximum number of mode decompositions is set to 10. The output of this algorithm is multiple mode components. The obtained mode components reflect the signal characteristics of the acoustic emission signal in different frequency domains. Some intrinsic mode components contain more noise signals, and the correlation between these mode components and the original signal is low, while the mode components containing more effective information from the real acoustic emission data have a higher correlation with the original signal.

[0037] Therefore, in this embodiment, the Pearson correlation coefficient between each modal component and the acoustic emission data within the monitoring period is calculated. The larger the Pearson correlation coefficient, the stronger the correlation between the modal component and the real acoustic emission data. Since modal components with more noise information have a high difference from modal components containing effective acoustic emission information, in order to extract the modal components that are more correlated with the real acoustic emission data, preferably, this embodiment uses the K-means clustering algorithm to cluster all Pearson correlation coefficients, setting the number of clusters to 2. Then, the numerical values ​​of the cluster centers of the two clusters are compared, and the modal component corresponding to the cluster with the largest cluster center value is taken as the effective modal component. Further, all the effective modal components are reconstructed to obtain reconstructed acoustic emission data. Its advantage is that it can better remove noise signals from the valve acoustic emission data.

[0038] Step 3: Analyze the distribution fluctuation degree and the instability characteristics of the fluctuation time of the reconstructed acoustic emission data to obtain the irregularity coefficient of acoustic emission signal fluctuation within each monitoring period. Based on the degree of disorder in the amplitude distribution of the reconstructed acoustic emission data in the frequency domain, and combined with the amplitude difference characteristics between the reconstructed acoustic emission data and the acoustic emission data of the valve in the non-leakage state in the frequency domain, obtain the degree of disorder in the frequency distribution of acoustic emission signal within each monitoring period, and then obtain the valve leakage judgment coefficient within each monitoring period.

[0039] Furthermore, based on the reconstructed acoustic emission data, potential leakage characteristics in the valve are extracted. Specifically, when the valve is operating normally, the amplitude of the acoustic emission signal is relatively small and stable. When a leak occurs in the valve, the frequent impacts of the fluid inside the valve cause the amplitude of the acoustic emission to become significantly larger, and the fluctuations in the acoustic emission data become extremely irregular.

[0040] Therefore, in this embodiment, the reconstructed acoustic emission data will be fitted. Preferably, a quadratic polynomial fitting technique is used to fit the reconstructed acoustic emission data to a curve, and then all extreme points of the obtained fitted curve are obtained. Since the obtained fitted curve exhibits periodic fluctuation characteristics, the maxima and minima are interspersed. Then, the amplitude difference between all adjacent maxima and minima is calculated, and the standard deviation of the amplitude difference between all adjacent maxima and minima is denoted as [the standard deviation of the amplitude difference between all adjacent maxima and minima on the fitted curve corresponding to the reconstructed acoustic emission data within the i-th monitoring period]. The result The larger the value, the more irregular the fluctuations in the acoustic emission signal.

[0041] Furthermore, the acoustic emission signal fluctuations caused by leakage are relatively unstable. Therefore, in this embodiment, the dispersion of the time difference between all adjacent two maxima on the fitted curve is obtained. Variance is used to measure the dispersion in this embodiment; however, existing dispersion measurement methods such as standard deviation can be used in practical applications. Correspondingly, the dispersion of the time difference between all adjacent two minima is statistically analyzed, and the sum of the two dispersions is used as the instability coefficient of the reconstructed acoustic emission data fluctuation period within the corresponding monitoring duration. The instability coefficient of the reconstructed acoustic emission data fluctuation period within the i-th monitoring duration is denoted as... Understandably, the result... This reflects the unstable characteristics of the acoustic emission signal fluctuation period.

[0042] Furthermore, the fluctuation irregularity coefficient of the acoustic emission signal within each monitoring period is calculated. Preferably, in this embodiment, the calculation formula for the fluctuation irregularity coefficient is: ,in, Let be the standard deviation of the amplitude difference between all adjacent maxima and minima on the fitted curve corresponding to the reconstructed acoustic emission data within the i-th monitoring duration. Let be the instability coefficient of the reconstructed acoustic emission data fluctuation period within the i-th monitoring duration, where is obtained The larger the value, the greater the fluctuation of the acoustic emission signal within the monitoring period, and the more irregular the fluctuation period.

[0043] Furthermore, when the valve is leak-free, the amplitude distribution of the acoustic emission data at various frequencies is relatively even. When the valve leaks, the frequency range of the acoustic emission data becomes wider, and as the leakage becomes more severe, the amplitude distribution across different frequency ranges becomes more chaotic, and the overall amplitude of the frequencies increases significantly. Therefore, in this embodiment, taking the i-th monitoring duration as an example, the Fast Discrete Fourier Transform (FFT) technique is used to obtain the spectrum of the reconstructed acoustic emission data within that monitoring duration. Then, the Higuchi algorithm is used to calculate the fractal dimension of the amplitude corresponding to all frequencies in the spectrum. The fractal dimension of the amplitude corresponding to all frequencies in the spectrum of the reconstructed acoustic emission data within the i-th monitoring duration is denoted as […]. The result This reflects the chaotic distribution of amplitudes across all frequencies.

[0044] Simultaneously, according to the above process in this embodiment, the spectrum of reconstructed acoustic emission data under the non-leakage state of the valve is obtained. It should be noted that the acoustic emission data under the non-leakage state of the valve is selected and set manually, and the reconstructed acoustic emission data and corresponding spectrum under the non-leakage state are obtained using the above process in this embodiment. Further, the ratio of the sum of all frequency amplitudes in the spectrum to the sum of all frequency amplitudes in the spectrum corresponding to the non-leakage state is calculated and denoted as... The result This reflects the richness of the acoustic emission signal at different frequencies. Therefore, the frequency distribution disorder of the acoustic emission signal within the i-th monitoring period is calculated using the following formula: ,in, Let i be the disorder of the frequency distribution of the acoustic emission signal within the i-th monitoring period. Let fractal dimension be the amplitude corresponding to all frequencies in the reconstructed acoustic emission data spectrum for the i-th monitoring duration. The ratio of the sum of all frequency amplitudes in the reconstructed acoustic emission data spectrum for the i-th monitoring duration to the sum of all frequency amplitudes in the spectrum corresponding to the non-leakage state is obtained. The larger the value, the more obvious the characteristics of the frequency distribution and richness of the acoustic emission signal within the monitoring period.

[0045] Thus, the leakage anomaly characteristics of the denoised acoustic emission data in both the time and frequency domains are obtained. Furthermore, the product of the irregularity coefficient of the acoustic emission signal fluctuation and the disorder of the acoustic emission signal frequency distribution obtained within the i-th monitoring period is used as the valve leakage determination coefficient for that monitoring period. The valve leakage determination coefficient within the i-th monitoring period is denoted as... The result This reflects the fluctuations in the acoustic emission signal of the valve and the degree of abnormality in its frequency domain distribution.

[0046] Step 4: Based on the correlation between the valve flow rate and the valve leakage judgment coefficient in each time period, obtain the leakage status assessment value for each time period. Combined with the distribution characteristics of the reconstructed acoustic emission data in each time period, use a neural network to detect valve leakage.

[0047] Furthermore, the degree of valve leakage is significantly affected by the internal fluid flow rate. For example, the higher the fluid flow rate, the greater the stress generated by the fluid causes changes in the internal structure of the valve, further leading to abnormal changes in the acoustic emission signal. Therefore, in this embodiment, the length of each detection period is set to 5 minutes. For any given period, the average valve flow rate for each monitoring duration is obtained and arranged in ascending order to obtain a flow rate sequence. Since each average flow rate has a corresponding valve leakage determination coefficient, the correlation coefficient between the flow rate sequence for each period and all valve leakage determination coefficients within each period is calculated. In this embodiment, the Spearman correlation coefficient is used. The larger the obtained correlation coefficient, the stronger the positive correlation between the valve leakage determination coefficient and the flow rate sequence, and the more significant the influence of flow rate on valve leakage anomalies. The product of the obtained correlation coefficient and the average value of all valve leakage determination coefficients within that period is then used as the leakage status assessment value for that period. This value reflects the degree of probability that the valve is leaking.

[0048] Furthermore, this embodiment employs an LSTM neural network model for leak detection. First, the distribution characteristics of all reconstructed acoustic emission data within each time period are obtained. Preferably, the distribution characteristics in this embodiment include the mean, standard deviation, kurtosis, peak factor, margin factor, and leak status assessment value of the reconstructed acoustic emission data. The reconstructed acoustic emission data for each time period and all corresponding feature values ​​are used as input to the LSTM neural network model to obtain the valve leak detection results for each time period. The optimizer for the LSTM neural network model is set to Adam, and the loss function is set to the cross-entropy loss function. Performing leak detection according to the above process in this embodiment helps improve the accuracy of valve leak detection. It should be noted that the specific usage process of the LSTM neural network model is prior art known to those skilled in the art, and will not be elaborated upon in this embodiment.

[0049] Based on the same inventive concept as the above method, this application also provides a valve-oriented leakage detection system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any of the above-described valve-oriented leakage detection methods.

[0050] It is understood that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0051] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0052] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the protection scope of this application.

Claims

1. A leakage detection method for valves, characterized in that, Includes the following steps: Acquire acoustic emission data and valve flow rate from the valve; Modal decomposition was performed on the acoustic emission data within each monitoring period. Effective modal components were extracted based on the correlation between the modal components and the acoustic emission data. The reconstructed emission data within each monitoring period was then obtained. The distribution fluctuation degree and the instability characteristics of the fluctuation time of the reconstructed acoustic emission data are analyzed to obtain the irregularity coefficient of acoustic emission signal fluctuation within each monitoring period. Based on the degree of disorder of the amplitude distribution of the reconstructed acoustic emission data in the frequency domain, and combined with the amplitude difference characteristics of the reconstructed acoustic emission data and the acoustic emission data of the valve in the non-leakage state in the frequency domain, the frequency distribution disorder of the acoustic emission signal within each monitoring period is obtained, and then the valve leakage judgment coefficient within each monitoring period is obtained. Based on the correlation between the valve flow rate and the valve leakage judgment coefficient in each time period, the leakage status assessment value of each time period is obtained. Combined with the distribution characteristics of the reconstructed acoustic emission data in each time period, a neural network is used to detect valve leakage.

2. The leakage detection method for valves as described in claim 1, characterized in that, The Pearson correlation coefficient between each modal component and the acoustic emission data within each monitoring period is calculated. All Pearson correlation coefficients are then clustered, and the modal component corresponding to the cluster with the largest central value is taken as the effective modal component within each monitoring period.

3. The leakage detection method for valves as described in claim 1, characterized in that, Curve fitting was performed on the reconstructed acoustic emission data within each monitoring period. The dispersion of the time difference corresponding to all two adjacent maxima and the dispersion of the time difference corresponding to all two adjacent minima on the fitted curve were statistically analyzed. The sum of the two dispersions was used as the instability coefficient of the fluctuation period of the reconstructed acoustic emission data within each monitoring period.

4. The leakage detection method for valves as described in claim 3, characterized in that, The acquisition of the irregularity coefficient of acoustic emission signal fluctuation within each monitoring period is further as follows: ,in, Let be the standard deviation of the amplitude difference between all adjacent maxima and minima on the fitted curve corresponding to the reconstructed acoustic emission data within the i-th monitoring duration. It is the instability coefficient of the reconstructed acoustic emission data fluctuation period within the i-th monitoring duration.

5. A leakage detection method for valves as described in claim 1, characterized in that, The degree of disorder in the frequency distribution of acoustic emission signals within each monitoring period was further obtained as follows: ,in, Let i be the disorder of the frequency distribution of the acoustic emission signal within the i-th monitoring period. Let fractal dimension be the amplitude corresponding to all frequencies in the reconstructed acoustic emission data spectrum for the i-th monitoring duration. The sum of all frequency amplitudes in the reconstructed acoustic emission data spectrum for the i-th monitoring duration is the ratio of the sum of all frequency amplitudes in the spectrum corresponding to the non-leakage state. The corresponding spectrum can be obtained by performing a frequency domain transformation on the reconstructed acoustic emission data.

6. The leakage detection method for valves as described in claim 1, characterized in that, The valve leakage determination coefficient for each monitoring period is the product of the irregularity coefficient of acoustic emission signal fluctuation and the disorder of acoustic emission signal frequency distribution within each monitoring period.

7. A leakage detection method for valves as described in claim 1, characterized in that, A time period includes multiple monitoring durations. The average flow rate of the valve at each monitoring duration within any time period is arranged in ascending order to form a flow rate sequence for that time period.

8. A leakage detection method for valves as described in claim 7, characterized in that, The leakage status assessment values ​​for each time period are further obtained as follows: Calculate the correlation coefficient between the flow velocity sequence for each time period and all leakage judgment coefficients within each time period. Then, multiply the correlation coefficient by the average value of all valve leakage judgment coefficients within each time period to obtain the leakage status assessment value for each time period.

9. A leakage detection method for valves as described in claim 1, characterized in that, The distribution characteristics of the reconstructed acoustic emission data in each time period include the mean, standard deviation, kurtosis, peak factor, margin factor, and leakage status assessment value of the reconstructed acoustic emission data. The reconstructed acoustic emission data and its distribution characteristics in each time period are used as input to the neural network to obtain the valve leakage detection results.

10. A valve-oriented leakage detection system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the valve-oriented leakage detection method as described in any one of claims 1-9.

Images