Leak hole jet noise identification method, device, electronic equipment and storage medium

Abstract

The present invention discloses a method and device for identifying leakage hole jet noise, the method comprising the following steps: A. Collecting acoustic emission signals at leakage holes of different shapes at different flow rates in a simulated pipeline; B. Extracting the Mel-frequency cepstral coefficients and time-frequency entropy of each jet noise signal from the preprocessed acoustic emission signals; C. Constructing the Mel-frequency cepstral coefficients and time-frequency entropy into a characteristic numerical matrix, using the characteristic numerical matrix as an input variable of a leakage hole jet noise identification model, and training the model; D. Using the trained model to identify leakage and the shape of leakage holes at different flow rates. The present invention extracts the frequency domain voiceprint feature Mel-frequency cepstral coefficients and the time domain feature time-frequency entropy from the collected abnormal sound signals of the leakage hole, uses the combined feature parameters as the input of the model, and identifies leakage holes of different shapes under different flow rate states, which can effectively improve the accuracy of identification.

Description

Technical Field

[0001] The invention relates to the technical field of industrial nondestructive testing, and in particular to a method, a device, an electronic device and a storage medium for identifying jet noise from a leak hole. Background Art

[0002] During the deep processing of crude oil, leakage holes are prone to occur in the equipment, which not only affects the output and quality of petroleum products, but may also cause serious safety accidents. Therefore, simulating the test of the jet noise of leakage holes of different shapes in a high flow rate environment can provide specific indicators for the safe range of the perforation size of the leakage holes during operation, which plays a vital role in ensuring the normal operation of actual working conditions.

[0003] Acoustic Emission (AE) is a phenomenon in which a material emits transient elastic stress waves due to the rapid release of energy locally. As a non-destructive monitoring method, it is widely used in the industrial field. The friction between the leakage of the medium and the leakage hole forms an acoustic emission signal, but there are few technical solutions in the prior art that use acoustic emission technology to monitor the leakage holes formed during the deep processing of crude oil, and there are even fewer technical solutions that use the combination of characteristic parameters as the input of the algorithm model to identify leakage holes at different flow rates and different shapes.

[0004] There are schemes in the prior art for detecting leakage by collecting audio data. For example, Chinese patent application CN111022941A discloses a method for detecting natural gas pipeline leakage based on LSTM recurrent neural network. The method comprises the following steps: collecting pipeline leakage audio data; extracting features from audio data using an improved Mel frequency cepstral coefficient MFCC method; constructing a natural gas pipeline leakage detection model based on long short-term memory LSTM recurrent neural network; training and optimizing parameters of the pipeline micro-leakage detection model, and verifying the robustness of the model. The scheme uses a single Mel frequency cepstral coefficient as the input parameter of the model, and the efficiency and accuracy of identifying leakage holes need to be further improved.

[0005] Therefore, there is an urgent need for a method that applies the principle of acoustic emission technology to effectively identify the jet noise of leakage holes through an algorithm model, and can greatly improve the identification accuracy of leakage holes of different shapes at different flow rates.

[0006] The information disclosed in this background technology section is only intended to enhance the understanding of the overall background of the invention and should not be regarded as an acknowledgment or any form of suggestion that the information constitutes the prior art already known to a person skilled in the art. Summary of the invention

[0007] The purpose of the present invention is to provide a method and device for identifying leakage hole jet noise. By collecting abnormal sound signals from leakage holes, the frequency domain soundprint feature Mel-frequency cepstral coefficients and the time domain feature time-frequency entropy are extracted, and the combined feature parameters are used as the input of the model to identify leakage holes of different shapes under different flow rate states, which can effectively improve the recognition accuracy.

[0008] To achieve the above-mentioned object, according to a first aspect of the present invention, a method for identifying jet noise of a leakage hole is provided, comprising the following steps: A, collecting acoustic emission signals at leakage holes of different shapes at different flow rates in a simulated pipeline; B, extracting the Mel-frequency cepstral coefficients and time-frequency entropy of each jet noise signal from the preprocessed acoustic emission signals; C, constructing the Mel-frequency cepstral coefficients and time-frequency entropy into a characteristic numerical matrix, using the characteristic numerical matrix as an input variable of a leakage hole jet noise identification model, and training the model; D, identifying leakages and the shapes of leakage holes at different flow rates through the trained model.

[0009] Furthermore, in the above technical solution, the preprocessing in step B may specifically include: performing VMD decomposition on the acoustic emission signal, and then combining and reconstructing the decomposed similar IMF components. The preprocessing may also include noise reduction processing and pre-emphasis processing.

[0010] Furthermore, in the above technical solution, the extraction process of the Mel-frequency cepstral coefficients in step B is as follows: converting the original acoustic emission signal into a time domain signal, and converting the time domain signal into a linear spectrum through a discrete Fourier transform or a fast Fourier transform; converting the linear spectrum into a Mel spectrum through a Mel-frequency filter group, taking the logarithmic energy of the Mel spectrum and obtaining the logarithmic spectrum; performing a discrete cosine transform on the logarithmic spectrum to obtain a matrix of Mel-frequency cepstral coefficients.

[0011] Further, in the above technical solution, the extraction process of the time-frequency entropy in step B is as follows: perform Hilbert transform on each frame of the acoustic emission signal to obtain the amplitude function and phase function corresponding to each frame; calculate the instantaneous frequency and instantaneous amplitude of each frame through the amplitude function and phase function; superimpose the amplitudes of the points with the same instantaneous frequency in each frame to construct a time-frequency matrix; divide the time-frequency matrix into H blocks of equal area; let the energy of each block be E j , the energy of the entire time-frequency plane is TE, and the energy of the H block area is normalized, that is:

[0012]

[0013] When the normalization condition is met, the calculation formula of the time-frequency entropy is as follows:

[0014]

[0015] The time-frequency entropy value is calculated and extracted using formula (1).

[0016] Furthermore, in the above technical solution, the step C of constructing the Mel-frequency cepstral coefficients and time-frequency entropy into a characteristic numerical matrix can be specifically as follows: suppose the matrix of Mel-frequency cepstral coefficients is A=M×N, where M is the number of frames of a signal and N is the filter group coefficient; calculate the transposed matrix B=N×M of the matrix A; calculate the matrix C=M×M and obtain the characteristic column vector D=M×1 corresponding to the maximum eigenvalue of the matrix C; obtain all frame samples of a signal, and construct a characteristic column vector E=M×1 by calculating the time-frequency entropy value of each frame; combine the characteristic column vector D of the Mel-frequency cepstral coefficients and the characteristic column vector E of the time-frequency entropy by direct superposition to obtain the characteristic numerical matrix in step C.

[0017] Furthermore, in the above technical solution, the leakage hole jet noise identification model can adopt a Gaussian mixture model.

[0018] Furthermore, in the above technical solution, the noise reduction process can be specifically as follows: based on the db4 wavelet function in the MATLAB wavelet toolbox, the original acoustic emission signal is decomposed into four layers into acoustic signals of different frequency bands; the acoustic signals of different frequency bands are fitted and high-frequency noise interference is filtered out; and the decomposition functions with high correlation are reconstructed.

[0019] Furthermore, in the above technical solution, the pre-emphasis processing can be specifically: emphasizing the high-frequency part of the acoustic emission signal to improve the high-frequency resolution of the signal, so that the spectrum of the signal remains in the entire frequency band and the spectrum is analyzed using the same signal-to-noise ratio.

[0020] According to the second aspect of the present invention, the present invention provides a leakage hole jet noise identification device, comprising: a signal acquisition module, used to collect acoustic emission signals at leakage holes of different shapes at different flow rates in a simulated pipeline; a variable extraction module, used to extract the Mel-frequency cepstral coefficients and time-frequency entropy of each jet noise signal from the preprocessed acoustic emission signal; a matrix construction and training module, used to construct the Mel-frequency cepstral coefficients and time-frequency entropy into a characteristic numerical matrix, and use the characteristic numerical matrix as an input variable of a leakage hole jet noise identification model to train the model; a shape recognition module, used to identify leaks and the shapes of leakage holes at different flow rates through the trained model.

[0021] Furthermore, the above technical solution may also include a preprocessing module, which further includes: a VMD decomposition submodule, which is used to perform VMD decomposition on the acoustic emission signal, and then combine and reconstruct the decomposed similar IMF components; a noise reduction processing submodule, which is used to decompose the acoustic emission signal into acoustic signals of different frequency bands, and perform fitting processing on the acoustic signals of different frequency bands to filter out high-frequency noise interference; a pre-emphasis processing submodule, which is used to emphasize the high-frequency part of the acoustic emission signal and thereby improve the high-frequency resolution of the signal.

[0022] According to the third aspect of the present invention, the present invention provides an electronic device for identifying leakage hole jet noise, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor executes the leakage hole jet noise identification method as described above.

[0023] According to a fourth aspect of the present invention, the present invention provides a non-transitory computer-readable storage medium, which stores computer-executable instructions, and the computer-executable instructions are used to enable the computer to execute the leakage hole jet noise identification method as described above.

[0024] Compared with the prior art, the present invention has one or more of the following beneficial effects:

[0025] 1) The present invention can collect acoustic emission signals at leakage holes of different shapes at different flow rates through simulated pipelines, extract the Mel-frequency cepstrum coefficients and time-frequency entropy of the jet noise in the acoustic emission signals, and construct the two into a new characteristic numerical matrix as the input variables of the leakage hole jet noise identification model, so as to more effectively identify the shape of the leakage hole. Tests have shown that it is more accurate than identifying the leakage situation through the Mel-frequency cepstrum coefficients of the acoustic emission signals alone;

[0026] 2) Mel-frequency cepstral coefficients (matrix form) and time-frequency entropy (numerical form) need to be converted and processed before being used as input parameters of the model. Based on their different expressions, the present invention constructs a characteristic numerical matrix, combines the two, and converts them into an input mode that the model can accept;

[0027] 3) The two characteristic variables, Mel-frequency cepstrum coefficient and time-frequency entropy, can provide a deeper understanding of the corresponding change relationship between the characteristic parameters of the faulty acoustic emission signal (i.e., the leakage hole jet noise) and the leakage holes in different states, providing a theoretical basis for real-time monitoring of acoustic signal analysis, and laying a foundation for proposing specific indicators for the real-time status of leakage hole faults during actual operating conditions, as well as for timely identification and early warning of leakage hole faults.

[0028] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, and to make the above and other purposes, technical features and advantages of the present invention easier to understand, one or more preferred embodiments are listed below and described in detail with the accompanying drawings as follows. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flow chart of the leakage hole jet noise identification method according to embodiment 1 of the present invention.

[0030] Figure 2 It is a flow chart of the leakage hole jet noise identification method according to embodiment 2 of the present invention.

[0031] Figure 3 It is a schematic diagram of the extraction process of Mel-frequency cepstral coefficients in Example 2 of the present invention.

[0032] Figure 4 It is a schematic diagram of the calculation and extraction process of time-frequency entropy in Example 2 of the present invention.

[0033] Figure 5 It is a data graph of the first column of the Mel-frequency cepstral coefficient matrix calculated by the present invention when the leakage hole is circular.

[0034] Figure 6 It is a data graph of the first column of the Mel-frequency cepstral coefficient matrix calculated by the present invention when the leakage hole is a square.

[0035] Figure 7 It is a data graph of the first column of the Mel-frequency cepstral coefficient matrix calculated by the present invention when the leakage hole is elliptical.

[0036] Figure 8 It is a schematic diagram of the structure of a leakage hole jet noise identification device according to embodiment 3 of the present invention.

[0037] Fig. 9 It is a schematic diagram of the structure of the leakage hole jet noise identification electronic device according to embodiment 4 of the present invention. DETAILED DESCRIPTION

[0038] The specific implementation modes of the present invention are described in detail below in conjunction with the accompanying drawings, but it should be understood that the protection scope of the present invention is not limited by the specific implementation modes.

[0039] Unless explicitly stated otherwise, throughout the specification and claims, the term “comprise” or variations such as “include” or “comprising”, etc., will be understood to include the stated elements or components but not to exclude other elements or components.

[0040] In this document, for the convenience of description, spatial relative terms such as "below", "below", "down", "above", "above", "upper", etc. may be used to describe the relationship between one element or feature and another element or feature in the accompanying drawings. It should be understood that the spatial relative terms are intended to include different orientations of the object in use or operation in addition to the orientation depicted in the figure. For example, if the object in the figure is turned over, the element described as being "below" or "below" other elements or features will be oriented "above" the element or feature. Therefore, the exemplary term "below" can include both the below and above directions. Objects may also have other orientations (rotated 90 degrees or other orientations) and the spatial relative terms used in this document should be interpreted accordingly.

[0041] In this document, the terms "first", "second", etc. are used to distinguish two different elements or parts, and are not used to limit a specific position or relative relationship. In other words, in some embodiments, the terms "first", "second", etc. can also be interchangeable.

[0042] The method, system, electronic device and storage medium of the present invention are described in more detail below by way of specific embodiments. It should be understood that the embodiments are merely exemplary and the present invention is not limited thereto.

[0043] First of all, it should be noted that in any leak hole identification system, it is important to extract features, and it is necessary to extract and characterize the identifiable components in the sound wave signal formed by the leak. The medium generates sound through the hole, and the shape of the hole determines what kind of sound is emitted. The shape of the hole is displayed in the envelope of the short-time power spectrum of the sound signal. The Mel-Frequency Cepstrum is a linear transformation of the logarithmic energy spectrum of the nonlinear Mel scale (mel scale) based on the sound frequency. It can describe a feature of this envelope as a characteristic parameter to characterize the sound wave signal of leak holes of different shapes. Although there is a method in the prior art to identify whether there is a leak in the pipeline through the Mel-Frequency Cepstrum coefficient, the recognition is not high, and the shape of the leak hole cannot be identified. In order to further improve the recognition of the leak hole sound wave signal, the present invention quantitatively characterizes the difference in energy distribution on the time-frequency plane, because the time-frequency distribution of leak holes in different states is different, and the test shows that the uniformity of the energy distribution in each time-frequency zone can reflect the difference in the state of the leak hole. Therefore, the present invention combines the voiceprint feature parameters (Mel-frequency cepstral coefficients) with information entropy (time-frequency entropy) to measure the uniformity of energy probability distribution in different time-frequency bands, and combines the two feature parameters to identify whether there is a leak and determine the shape of the leak hole.

[0044] Example 1

[0045] like Figure 1As shown, the leakage hole jet noise identification method of this embodiment includes the following steps:

[0046] Step S101, collect acoustic emission signals at leakage holes of different shapes at different flow rates in the simulated pipeline. Specifically, experimental equipment including fans, buffer boxes, flow meters, ball valves, nozzles, and pipe flanges can be prepared in advance to build a simulated pipeline. Simulated leakage holes of different shapes (such as round, square or elliptical, etc.) are set at the end of the simulated pipeline. The fan continuously supplies air to the entire set of equipment. The acoustic emission detection system can convert mechanical waves into electrical signals through sensors with piezoelectric effect, so that the acoustic emission signals have an intuitive expression in the form of waveforms, and the acoustic emission signal data is obtained through the signal acquisition of the sensor.

[0047] Step S102, extracting the Mel-frequency cepstrum coefficients and time-frequency entropy of each jet noise signal from the preprocessed acoustic emission signal. The preprocessing of the acoustic emission signal may include, but is not limited to: VMD decomposition of the acoustic emission signal, and then combining and reconstructing the decomposed similar IMF components; it may also include noise reduction processing and pre-emphasis processing. This embodiment identifies whether there is a leak and the shape of the leak hole by extracting the Mel-frequency cepstrum coefficients MFCC as the voiceprint feature and the time-frequency entropy s(q) as the time domain feature.

[0048] Step S103, constructing a characteristic numerical matrix from the Mel frequency cepstral coefficients extracted in step S102 and the calculated time-frequency entropy, and using the characteristic numerical matrix as an input variable of the leakage hole jet noise recognition model to train the model. The extracted Mel frequency cepstral coefficients are a matrix, and the calculated time-frequency entropy is a numerical value, so it is necessary to construct a new characteristic numerical matrix to combine the two. Preferably, but not restrictively, the leakage hole jet noise recognition model of this embodiment can adopt a Gaussian mixture model GMM.

[0049] Step S104: identifying leakage and the shape of leakage holes at different flow rates using the trained model.

[0050] The leakage hole jet noise identification method of the present embodiment 1 can collect acoustic emission signals at leakage holes of different shapes at different flow rates through a simulated pipeline, and extract the two characteristic variables of the Mel-frequency cepstrum coefficient and the time-frequency entropy of the jet noise in the acoustic emission signal and construct the two into a new characteristic numerical matrix as the input variable of the leakage hole jet noise identification model, so as to more effectively identify whether there is a leakage and the shape of the leakage hole. Compared with identifying the leakage situation only through the Mel-frequency cepstrum coefficient of the acoustic emission signal, it is more accurate, and can more deeply understand the corresponding change relationship between the characteristic parameters of the faulty acoustic emission signal and the leakage holes in different states, so as to provide a theoretical basis for real-time monitoring of acoustic signal analysis, and lay a foundation for proposing specific indicators for the real-time status of the leakage hole fault during actual working conditions and for timely identification and early warning of the leakage hole fault.

[0051] Example 2

[0052] like Figure 2 As shown, the leakage hole jet noise identification method of this embodiment is a more specific embodiment based on embodiment 1, and includes the following steps:

[0053] Step S201, collect acoustic emission signals at leakage holes of different shapes at different flow rates in the simulated pipeline. Specifically, experimental equipment including fans, buffer boxes, flow meters, ball valves, nozzles, and pipe flanges can be prepared in advance to build a simulated pipeline. Simulated leakage holes of different shapes (such as round, square or elliptical, etc.) are set at the end of the simulated pipeline. The fan continuously supplies air to the entire set of equipment. The acoustic emission detection system can convert mechanical waves into electrical signals through sensors with piezoelectric effect, so that the acoustic emission signals have an intuitive expression in the form of waveforms, and the acoustic emission signal data is obtained through the signal acquisition of the sensor.

[0054] Step S202: preprocessing the collected acoustic emission signal. The preprocessing may include VMD decomposition (variational mode decomposition), noise reduction processing and pre-emphasis processing.

[0055] VMD decomposition is a time-frequency decomposition algorithm. Compared with other traditional time-frequency decomposition methods, this method has great advantages in processing the center frequency and bandwidth between each IMF component. First, this method adaptively decomposes the original signal according to the frequency to ensure that the center frequencies between the decomposed IMFs do not overlap, thus avoiding the problem of modal overlap between IMF components. During the decomposition process, the center frequency and bandwidth of the IMF separation continue to change. The iterative optimization method can be used to ensure the bandwidth of each IMF decomposed by VMD, and then the combination reconstruction of the same type of IMF components can be realized by using Matlab software to obtain the acoustic signals of leakage holes of different shapes.

[0056] The collected acoustic emission signals may be mixed with a lot of other noises that are irrelevant to the formation of the leak hole jet noise. In order to remove these irrelevant noise signals to the maximum extent, after the collected acoustic emission signals are intercepted in stages, noise reduction and pre-emphasis can be performed. The noise reduction process can be based on the db4 wavelet function in the matlab wavelet toolbox to perform a four-layer decomposition of the original acoustic emission signal, decomposing it into acoustic signals of different frequency bands; then the acoustic signals of different frequency bands are fitted and the high-frequency noise interference is filtered out; finally, the decomposition function with high correlation is reconstructed to improve the signal-to-noise ratio. When analyzing the spectrum, the spectrum components of the high-frequency part are small and difficult to study and analyze. Therefore, the sampled and quantized acoustic signal is pre-emphasized, the high-frequency part of the signal is emphasized, the high-frequency resolution of the signal is improved, the spectrum of the signal becomes flatter, and the spectrum can be analyzed with the same signal-to-noise ratio in the entire frequency band, which is more conducive to the study of the sound channel parameters and spectrum. The pre-emphasis method can improve the spectrum of the high-frequency part and increase the high-frequency resolution of the signal.

[0057] Step S203, extracting the voiceprint feature parameters of the acoustic emission signal, Mel-Frequency Cepstral Coefficients (MFCC). The MFCC extraction process diagram is shown in FIG. Figure 3 As shown, the details are as follows:

[0058] (1) After preprocessing, the original acoustic emission signal s(n) becomes a time domain signal x(n); the time domain signal x(n) is padded with a number of zeros to form a sequence of length N (usually N = 512), and then the discrete Fourier transform (DFT) or fast Fourier transform (FFT) is used to transform these time domain signals into a linear spectrum X(k). The transformation formula is:

[0059]

[0060] (2) The obtained linear spectrum X(k) is passed through the Mel frequency filter bank, so that the linear spectrum is converted into the Mel spectrum, and then the logarithmic energy is taken to obtain the logarithmic spectrum S(m). The Mel frequency filter bank is a group of triangular bandpass filters, 0≤m≤M is the number of filters. These filters have triangular filtering characteristics, and the corresponding center frequency is f(m). When the m value increases, the distance between f(m) also increases, and when the m value decreases, the distance between f(m) also decreases. The transfer function of each bandpass filter is as follows:

[0061]

[0062] In order to improve the robustness of the voiceprint recognition system, the logarithmic energy of the Mel spectrum is usually taken, and the total transfer function from the linear spectrum X(k) to the logarithmic spectrum S(m) is:

[0063]

[0064] (3) S(m) is subjected to discrete cosine transform (DCT) to obtain the MFCC parameters C(n):

[0065]

[0066] In an actual voiceprint recognition system, not all MFCC dimensions are used. Experiments have shown that the MFCC coefficients of the first and last dimensions have the best signal discrimination performance. In this embodiment, the MFCC coefficients of the first 12 dimensions are used.

[0067] Step S204, extract the time-domain characteristic time-frequency entropy s(q) of the acoustic emission signal. The extraction process is actually a calculation process, and the result of the calculation is in the form of a numerical value. Entropy refers to the uncertainty of the data signal, which is a quantitative description of the uncertainty of a certain signal system. It is a measure of the overall information of the signal system from an average perspective. The greater the uncertainty of the information contained in the signal system, the more information sources it contains, and the greater the entropy value of the signal system.

[0068] The extraction process of the time-domain feature parameter time-frequency entropy s(q) is as follows Figure 4 As shown in Figure 2, the calculation of time-frequency entropy is as follows:

[0069] (1) For each frame of the decomposed acoustic emission signal Perform a Hilbert transform:

[0070]

[0071] (2) Construct an analytical function to obtain the amplitude function and phase function corresponding to each frame:

[0072]

[0073]

[0074]

[0075] In the above formula, z n (t) is the constructed analytical function; a n (t) is the amplitude function; φ n (t) Phase function. According to the above formula, the instantaneous frequency and instantaneous amplitude of each frame can be further obtained:

[0076]

[0077]

[0078] (3) The amplitudes of the points with the same instantaneous frequency in each frame obtained in (2) are superimposed to construct a large sparse matrix (i.e., time-frequency matrix). The time-frequency matrix is ​​divided into H blocks of equal area. In this embodiment, the number of blocks for calculating the time-frequency entropy H is 8, and the energy of each block is E. j , j = 1, 2, ... H, the energy of the entire time-frequency plane is TE, and the energy of the H block area is normalized as follows:

[0079]

[0080] From the above formula, we can get:

[0081]

[0082] According to entropy theory, if the normalization conditions for calculating entropy are met, the calculation formula for time-frequency entropy can be obtained:

[0083]

[0084] Step S205, construct a characteristic numerical matrix by using the voiceprint feature MFCC (expressed as a matrix) extracted in step S203 and the time-domain feature time-frequency entropy s(q) (expressed as a numerical value) calculated in step S204, and use the characteristic numerical matrix as the input variable of the Gaussian mixture model to train the model. Since the MFCC of a signal is a matrix, and the time-frequency entropy s(q) of a signal is a numerical value, the dimensions are different and cannot be directly used as the input of the Gaussian mixture model. This embodiment solves this problem by constructing a new characteristic numerical matrix. The specific steps are as follows: the voiceprint feature MFCC matrix obtained by step S203 is A=M×N, where M is the number of frames of a signal, and N is the filter group coefficient. The transposed matrix B=N×M of the matrix A is obtained, and the matrix C=M×M is calculated. The characteristic column vector D=M×1 corresponding to the maximum eigenvalue of the matrix C is further obtained. At the same time, the function of Matlab software is used to obtain all frame samples of a signal, and the time-frequency entropy of each frame is calculated to form a feature column vector E=M×1. The feature column vector D and the feature column vector E are combined by direct superposition to obtain the final feature numerical matrix F=M×2. The feature numerical matrix F is substituted into the Gaussian mixture model for training and recognition. Tests have shown that the recognition rate is better than that of using MFCC recognition alone.

[0085] The following is a specific example to illustrate the process of constructing the characteristic value matrix of step S205:

[0086] This example uses the acoustic emission data of three different shapes of leakage holes at a flow rate of 5.53 m / s for illustration. For each signal, 10 waveforms are taken for analysis. One waveform has 2048 voltage values, 256 voltage values ​​are used as one frame, and the frame shift is 80. There are 253 frames of data in total, that is, one signal has 253 samples, of which 12 filter groups are set, that is, the dimension of each frame is 12 dimensions, so the MFCC matrix A is in the form of 253×12. Figure 5 This is the data graph of the first column of the MFCC matrix calculated when the leakage hole is circular. Figure 6 This is the data graph of the first column of the MFCC matrix calculated when the leakage hole is a square. Figure 7 This is the data graph of the first column of the MFCC matrix calculated when the leakage hole is elliptical.

[0087] At the same time, according to the above step S205, the column vector corresponding to the maximum eigenvalue of the 253×12 matrix (i.e., matrix A) is obtained to obtain the MFCC column vector of the 253×1 form (i.e., feature column vector D), and the time-frequency entropy data column vector of each frame of the 253 frames (i.e., feature column vector E) is obtained, and a 253×2 feature vector group is formed by direct combination and superposition. In this embodiment, the first waveform is taken for data description. A waveform has 23 frames of data, i.e., 23 samples. The MFCC and time-frequency entropy of each frame are obtained. After direct combination and superposition, as shown in Table 1, each row value in the table represents a sample.

[0088] Table 1

[0089] Mel coefficient Time-Frequency Entropy Frame 1 -2.404 2.629 Frame 2 -2.485 2.965 Frame 3 -3.280 2.286 Frame 4 -2.198 2.564 Frame 5 -3.262 2.203 Frame 6 -2.545 2.228 Frame 7 -2.466 2.271 Frame 8 -3.210 3.896 Frame 9 -2.099 4.552 Frame 10 -2.681 2.945 Frame 11 -2.484 2.955 Frame 12 -2.873 1.767 Frame 13 -2.877 2.676 Frame 14 -4.348 2.353 Frame 15 -2.902 2.000 Frame 16 -2.859 2.232 Frame 17 -3.111 3.582 Frame 18 -2.821 2.612 Frame 19 -4.435 2.275 Frame 20 -2.902 2.487 Frame 21 -2.936 2.284 Frame 22 -3.070 2.627 Frame 23 -2.723 3.269

[0090] The test set and training set were processed at a ratio of 1:4, that is, 200 sample matrices were used as training sets and 53 samples were used as test sets. The three vector groups and the single MFCC group were randomly assigned by computer to the Gaussian mixture model (GMM) for training and recognition comparison. The recognition rate comparison of the MFCC and time-frequency entropy s(q) combined recognition method (i.e., MFCC_S) and the single MFCC recognition method for leakage holes of different shapes at the same flow rate of 5.53 m / s is shown in Table 2.

[0091] Table 2

[0092]

[0093] Step S206, using the trained Gaussian mixture model (GMM) to obtain the detection results, realize the recognition of the soundprint features and time domain features of the leakage holes with different flow rates and different shapes, thereby realizing the signal recognition and early warning of the leakage holes.

[0094] The leakage hole jet noise identification method of this embodiment 2 is a preferred solution based on embodiment 1. The acoustic emission signals at leakage holes of different shapes at different flow rates can be collected by simulating pipelines. By extracting the two characteristic variables of the Mel-frequency cepstrum coefficient and the time-frequency entropy of the jet noise in the acoustic emission signal and constructing the two into a new characteristic numerical matrix as the input variable of the leakage hole jet noise identification model, it is possible to more effectively identify whether there is a leak and the shape of the leakage hole. This embodiment further proves through test data that the method of the present invention is more accurate than identifying the leakage situation through the Mel-frequency cepstrum coefficient of the acoustic emission signal alone.

[0095] Example 3

[0096] like Figure 8 As shown, the leakage hole jet noise identification device of this embodiment includes a signal acquisition module 301, a variable extraction module 303, a matrix construction and training module 304 and a shape recognition module 305. Among them, the signal acquisition module 301 is used to collect acoustic emission signals at leakage holes of different shapes at different flow rates in the simulated pipeline; the variable extraction module 303 is used to extract the Mel frequency cepstral coefficients and time-frequency entropy of each jet noise signal from the pre-processed acoustic emission signal; the matrix construction and training module 304 is used to construct the Mel frequency cepstral coefficients and time-frequency entropy into a characteristic numerical matrix, and use the characteristic numerical matrix as an input variable of the leakage hole jet noise identification model to train the model; the shape recognition module 305 is used to identify the leakage and the shape of the leakage hole at different flow rates through the trained model.

[0097] Furthermore, the leakage hole jet noise identification device of this embodiment may also include a preprocessing module 302, which may further include: a VMD decomposition submodule, which is used to perform VMD decomposition on the acoustic emission signal, and then combine and reconstruct the decomposed similar IMF components; a noise reduction processing submodule, which is used to decompose the acoustic emission signal into acoustic signals of different frequency bands, and perform fitting processing on the acoustic signals of different frequency bands to filter out high-frequency noise interference; a pre-emphasis processing submodule, which is used to emphasize the high-frequency part of the acoustic emission signal and thereby improve the high-frequency resolution of the signal.

[0098] The device of this embodiment corresponds to the method of Embodiment 1 and Embodiment 2, and can achieve the same technical effect.

[0099] Example 4

[0100] Fig. 96 is a hardware structure diagram of the electronic device for identifying the jet noise of the leakage hole in this embodiment. The device (such as a terminal, a server, etc.) includes one or more processors 610 and a memory 620. Taking a processor 610 as an example, the device may also include: an input device 630 and an output device 640.

[0101] The processor 610 , the memory 620 , the input device 630 , and the output device 640 may be connected via a bus or in other ways.

[0102] The memory 620 is a non-transitory computer-readable storage medium that can be used to store non-transitory software programs, non-transitory computer executable programs, and modules. The processor 610 executes various functional applications and data processing of the electronic device by running the non-transitory software programs, instructions, and modules stored in the memory 620, that is, the processing method of the above method embodiment is implemented.

[0103] The memory 620 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data, etc. In addition, the memory 620 may include a high-speed random access memory, and may also include a non-transitory memory, such as at least one disk storage device, a flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory 620 may optionally include a memory remotely arranged relative to the processor 610, and these remote memories may be connected to the processing device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0104] The input device 630 can receive input digital or character information and generate signal input. The output device 640 can include a display device such as a display screen.

[0105] The one or more modules are stored in the memory 620, and when executed by the one or more processors 610, the following steps are performed: A. Acquiring acoustic emission signals at leakage holes of different shapes at different flow rates in the simulated pipeline; B. Extracting the Mel-frequency cepstral coefficients and time-frequency entropy of each jet noise signal from the preprocessed acoustic emission signals; C. Constructing the Mel-frequency cepstral coefficients and time-frequency entropy into a characteristic numerical matrix, using the characteristic numerical matrix as an input variable of a leakage hole jet noise identification model, and training the model; D. Identifying leaks and the shapes of leakage holes at different flow rates through the trained model.

[0106] The electronic device can execute the method provided by the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the methods provided by other embodiments of the present invention.

[0107] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0108] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0109] Example 5

[0110] In this embodiment, a non-transitory computer-readable storage medium is provided, and the non-transitory computer-readable storage medium stores computer-executable instructions, and the instructions are used to enable a computer to execute the following leakage hole jet noise identification method:

[0111] A. Collect acoustic emission signals at leakage holes of different shapes at different flow rates in the simulated pipeline; B. Extract the Mel-frequency cepstral coefficients and time-frequency entropy of each jet noise signal from the preprocessed acoustic emission signals; C. Construct the Mel-frequency cepstral coefficients and time-frequency entropy into a characteristic numerical matrix, use the characteristic numerical matrix as the input variable of the leakage hole jet noise identification model, and train the model; D. Use the trained model to identify leaks and the shapes of leakage holes at different flow rates.

[0112] The foregoing description of specific exemplary embodiments of the present invention is for the purpose of illustration and demonstration. These descriptions are not intended to limit the present invention to the precise form disclosed, and it is clear that many changes and variations can be made based on the above teachings. The purpose of selecting and describing the exemplary embodiments is to explain the specific principles of the present invention and its practical application, so that those skilled in the art can realize and utilize various different exemplary embodiments of the present invention and various different selections and changes. Any simple modifications, equivalent changes and modifications made to the above exemplary embodiments should fall within the scope of protection of the present invention.

Claims

1. A method for identifying leakage hole jet noise, characterized in that: The steps include: A. Collect acoustic emission signals at leak holes of different shapes at different flow rates in simulated pipelines; B. extracting the Mel-frequency cepstrum coefficients and time-frequency entropy of each jet noise signal from the pre-processed acoustic emission signal; The extraction process of the Mel-frequency cepstrum coefficients is as follows: converting the original acoustic emission signal into a time domain signal, converting the time domain signal into a linear spectrum through discrete Fourier transform or fast Fourier transform; converting the linear spectrum into a Mel spectrum through a Mel-frequency filter bank, taking the logarithmic energy of the Mel spectrum and obtaining the logarithmic spectrum; performing discrete cosine transform on the logarithmic spectrum to obtain a matrix of the Mel-frequency cepstrum coefficients; The extraction process of the time-frequency entropy is as follows: performing Hilbert transform on each frame of the acoustic emission signal to obtain the amplitude function and phase function corresponding to each frame; calculating the instantaneous frequency and instantaneous amplitude of each frame through the amplitude function and phase function; The amplitudes of the points with the same instantaneous frequency in each frame are superimposed to construct a time-frequency matrix; the time-frequency matrix is ​​divided into H blocks of equal area; the energy of each block is set to E j , the energy of the entire time-frequency plane is TE, and the energy of the H block area is normalized, that is: When the normalization condition is met, the calculation formula of the time-frequency entropy is as follows: Calculate and extract the time-frequency entropy value by formula (1); C. constructing the Mel-frequency cepstrum coefficients and time-frequency entropy into a characteristic numerical matrix, using the characteristic numerical matrix as an input variable of a leakage hole jet noise recognition model, and training the model; The Mel-frequency cepstral coefficients and time-frequency entropy are constructed into a characteristic numerical matrix as follows: assuming that the matrix of the Mel-frequency cepstral coefficients is A=M×N, where M is the number of frames of a signal and N is the filter bank coefficient; calculating the transposed matrix B=N×M of the matrix A; calculating the matrix C=M×M and obtaining the characteristic column vector D=M×1 corresponding to the maximum eigenvalue of the matrix C; Obtain all frame samples of the signal, and construct a feature column vector E=M×1 by calculating the time-frequency entropy value of each frame; combine the feature column vector D of the Mel-frequency cepstrum coefficient and the feature column vector E of the time-frequency entropy by direct superposition to obtain the feature value matrix in step C; D. The trained model is used to identify leaks and the shapes of leak holes at different flow rates.

2. The method for identifying leakage hole jet noise according to claim 1, characterized in that: The preprocessing in step B specifically includes: performing VMD decomposition on the acoustic emission signal, and then combining and reconstructing the decomposed IMF components of the same type.

3. The method for identifying leakage hole jet noise according to claim 2, characterized in that: The preprocessing also includes: noise reduction processing and pre-emphasis processing.

4. The method for identifying leakage hole jet noise according to claim 1, characterized in that: The leakage hole jet noise identification model adopts a Gaussian mixture model.

5. The method for identifying leakage hole jet noise according to claim 3, characterized in that: The noise reduction process is specifically as follows: Based on the db4 wavelet function in the MATLAB wavelet toolbox, the original acoustic emission signal is decomposed into four layers into acoustic signals of different frequency bands; Performing fitting processing on the acoustic signals of different frequency bands and filtering out high-frequency noise interference; Reconstruct the decomposition functions with high correlation.

6. The method for identifying leakage hole jet noise according to claim 3, characterized in that: The pre-emphasis processing is specifically: emphasizing the high-frequency part of the acoustic emission signal to improve the high-frequency resolution of the signal, so that the spectrum of the signal is maintained in the entire frequency band and the spectrum is analyzed using the same signal-to-noise ratio.

7. A leakage hole jet noise identification device, characterized in that: The method according to any one of claims 1 to 6 comprises: A signal acquisition module is used to collect acoustic emission signals at leak holes of different shapes at different flow rates in the simulated pipeline; A variable extraction module, used for extracting the Mel-frequency cepstrum coefficient and time-frequency entropy of each jet noise signal from the preprocessed acoustic emission signal; A matrix construction and training module, used for constructing the Mel-frequency cepstrum coefficients and time-frequency entropy into a characteristic numerical matrix, using the characteristic numerical matrix as an input variable of a leakage hole jet noise recognition model, and training the model; The shape recognition module is used to recognize the shapes of leaks and leak holes at different flow rates through the trained model.

8. The leakage hole jet noise identification device according to claim 7, characterized in that: Also includes a pre-processing module, the pre-processing module further includes: A VMD decomposition submodule, used for performing VMD decomposition on the acoustic emission signal, and then combining and reconstructing the decomposed similar IMF components; A noise reduction processing submodule, used for decomposing the acoustic emission signal into acoustic signals of different frequency bands, and performing fitting processing on the acoustic signals of different frequency bands to filter out high-frequency noise interference; The pre-emphasis processing submodule is used to emphasize the high-frequency part of the acoustic emission signal to improve the high-frequency resolution of the signal.

9. An electronic device for identifying jet noise from a leak hole, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the leakage hole jet noise identification method according to any one of claims 1 to 6.

10. A non-transitory computer-readable storage medium, characterized in that: The non-transitory computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable the computer to execute the leakage hole jet noise identification method according to any one of claims 1 to 6.

Images