A method, system and device for automatically identifying abnormal high-frequency vibration of a PCCP pipeline
By combining modal decomposition and Hilbert spectrum calculation with high-pass filtering and normalization, the problem of high-frequency identification of abnormal vibration in PCCP pipelines was solved, realizing automated identification of abnormal vibration in PCCP pipelines and accurate differentiation of high-frequency signals.
Patent Information
- Application Number
- CN202311671312.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-07
- Publication Date
- 2026-05-15
- Estimated Expiration
- 2043-12-07
AI Technical Summary
Existing technologies are insufficient to effectively identify the high-frequency components of abnormal vibrations in PCCP pipelines, which affects the safe operation of the pipeline.
By combining modal decomposition and Hilbert spectrum calculation with high-pass filtering and normalization, the target time period and frequency threshold are selected by forming the filtering threshold-time-filtered energy matrix and the energy-time curve before filtering, and the highest frequency of abnormal vibration is calculated.
It enables automated identification of high-frequency abnormal vibrations in PCCP pipelines, accurately distinguishes actual signals from random noise, and improves identification efficiency and accuracy.
Smart Images
Figure CN117685515B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of automatic vibration monitoring technology for concrete pipelines, and more specifically to a high-frequency automatic identification method, system and equipment for abnormal vibration of PCCP pipelines. Background Technology
[0002] Prestressed concrete cylinder pipes (PCCPs) have become highly competitive in pipeline water transportation due to their advantages such as high pressure resistance and good durability. However, some PCCP pipelines may gradually corrode and deteriorate after a period of service, eventually leading to failure. As the core component of the entire pipeline, the reliable and stable operation of the prestressed steel wire is crucial for the safe operation of the pipeline. Wire breakage monitoring is essential for assessing the condition of the PCCP and avoiding catastrophic consequences. The breakage of prestressed steel wires in PCCPs generates strong sound waves (vibrations), making the identification of the vibration characteristics generated by prestressed steel wire breakage critical.
[0003] The vibration modes that may occur after a steel wire breaks include: 1. Vibration caused by the combined action of the steel wire and concrete, where the contact surface between them remains intact and they move synchronously; 2. Vibration caused by the non-coordinated deformation of the steel wire and concrete, resulting in separation of the contact surface and high-frequency vibration due to dynamic friction during their movement; 3. Impact vibration caused by the violent rebound of the steel wire. Due to the variability of operating conditions, the vibration signal may contain any combination of the above three vibration modes. In addition, the vibration signal also includes the dynamic response of the pipeline and the dynamic response of the optical fiber. Various combinations of situations may occur during the rebound of a broken steel wire, and the bandwidth of the acoustic vibration generated by different situations varies. Bandwidth is an important parameter for identifying broken wires, and how to identify the bandwidth of vibration signals urgently needs to be solved, among which high-frequency vibration identification is key.
[0004] Therefore, how to provide a high-frequency automated identification method for abnormal vibrations in PCCP pipelines is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0005] In view of this, the present invention provides a method, system, and device for high-frequency automated identification of abnormal vibration in PCCP pipelines. To achieve the above objectives, the present invention provides the following technical solution:
[0006] A high-frequency automated identification method for abnormal vibration in PCCP pipelines includes:
[0007] S1: Monitor abnormal vibration data of PCCP pipeline, perform modal decomposition and Hilbert spectrum calculation on the abnormal vibration data of PCCP pipeline to obtain the matrix and time spectrum characterizing time-frequency-energy;
[0008] Among them, monitoring abnormal vibration data of PCCP pipelines includes data on broken wire vibration, etc.
[0009] S2: High-pass filtering is performed based on the frequency resolution of the time spectrum. The energy values corresponding to all frequency points at each time point after filtering are superimposed to obtain the energy value at each time point, forming the filtered vibration energy-time curve.
[0010] S3: Normalize the vibration energy-time curve after each filtering, and convert the normalized one-dimensional vibration energy data into multi-dimensional data to form a filtering threshold-time-filtered energy matrix;
[0011] S4: Based on the time spectrum of the unfiltered data, superimpose the energy values corresponding to all frequency points at each time point to obtain the energy value at each time point, forming the energy-time curve before filtering. Select the target time period in the energy-time curve before filtering and filter out the sub-matrix of the filtering threshold-time-filtered energy matrix corresponding to the target time period.
[0012] S5: Select the upper quartile of the row or column corresponding to each filtering frequency threshold of the submatrix, and average the normalized energy values that are greater than the upper quartile for each frequency threshold to form the filtering frequency threshold-normalized filtered energy mean curve.
[0013] S6: Select the maximum value on the filter frequency threshold-normalized filtered energy mean curve, and subtract the frequency resolution from the filter frequency threshold corresponding to the maximum value to obtain the highest frequency of abnormal vibration of the PCCP pipeline.
[0014] Optionally, the mode decomposition used in S1 includes adding white noise.
[0015] Optionally, the mode decomposition method used is a mode decomposition method based on Empirical Mode Decomposition (EMD). The mode decomposition method must include a white noise step, such as EEMD, CEEMD, CEEMDAN, and ICEEMDAN, but the mode decomposition method including the white noise step is not limited to the mentioned mode decomposition methods.
[0016] Optionally, in S2, the high-pass filtering is used to filter out components that are below the filtering frequency threshold corresponding to the time-frequency-energy matrix.
[0017] Optionally, in step S3, the vibration energy-time curve after each filtering is normalized, including the following steps:
[0018] Let [x] be the one-dimensional numerical sequence of the filtered vibration energy as a function of time. nThe sequence has a length of n, and the normalized sequence is [x']. n Calculate the sequence [x] n The maximum value of max(x);
[0019] For each sequence [x] n Normalize each data point in the dataset using the following formula:
[0020]
[0021] In the formula, the subscript i represents the i-th data point.
[0022] Optionally, selecting the target time period in the energy-time curve before filtering includes the following steps:
[0023] Normalize the energy before filtering;
[0024] The time period corresponding to the normalized, unfiltered energy being less than the threshold Tv is taken as the target time period.
[0025] Optionally, the energy before filtering can be normalized, including the following steps:
[0026] Let [y] be the one-dimensional numerical sequence of energy change over time before filtering. n The sequence length is n, and the normalized sequence is [y']. n Calculate [y] n The maximum value of y is max(y) and the minimum value of y is min(y).
[0027] [y] n Subtract the minimum value min(y) from each data point in the sequence to obtain a new sequence. The formula is:
[0028]
[0029] In the formula, the subscript i represents the i-th data point;
[0030] calculate Maximum value
[0031] right Normalization is performed using the following formula:
[0032]
[0033] In the formula, It is a one-dimensional numerical sequence after removing the minimum value; for The i-th element in the middle; y i ' represents the i-th element of the normalized numerical sequence.
[0034] A high-frequency automated identification system for abnormal vibration in PCCP pipelines includes:
[0035] The monitoring module monitors abnormal vibration data of the PCCP pipeline, performs modal decomposition and Hilbert spectrum calculation on the abnormal vibration data of the PCCP pipeline, and obtains a matrix and time spectrum characterizing time-frequency-energy.
[0036] The filtering module performs high-pass filtering based on the frequency resolution of the time spectrum, and superimposes the energy values corresponding to all frequency points at each time point after filtering to obtain the energy value at each time point, forming the filtered vibration energy-time curve.
[0037] The normalization module normalizes the vibration energy-time curve after each filtering and transforms the normalized one-dimensional vibration energy data into multi-dimensional data, forming a filtering threshold-time-filtered energy matrix.
[0038] The module selects the energy values corresponding to all frequency points at each time point based on the time spectrum of the unfiltered data, and obtains the energy value at each time point to form the energy-time curve before filtering. The target time period in the energy-time curve before filtering is selected, and the sub-matrix of the filtering threshold-time-filtered energy matrix corresponding to the target time period is filtered out.
[0039] The statistics module selects the upper quartile of the row or column corresponding to each filtering frequency threshold of the submatrix, and averages the normalized energy values that are greater than the upper quartile for each frequency threshold, forming a curve of filtering frequency threshold - normalized filtered energy mean.
[0040] The calculation module selects the maximum value on the filter frequency threshold-normalized filtered mean energy curve, and subtracts the frequency resolution from the filter frequency threshold corresponding to the maximum value to obtain the highest frequency of abnormal vibration in the PCCP pipeline.
[0041] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a high-frequency automated identification method for abnormal vibration of PCCP pipelines.
[0042] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a method that realizes the automatic identification of abnormal high-frequency vibrations. By obtaining the difference between the frequency band where the actual signal is located and the frequency band containing only random noise, the abnormal high-frequency vibrations can be accurately identified. At the same time, this method is also applicable to the analysis of other signals that have high-frequency impact vibration components. Attached Figure Description
[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0044] Figure 1 This is a schematic diagram of the method flow of the present invention;
[0045] Figure 2 This is a schematic diagram of the broken wire vibration data provided in this embodiment;
[0046] Figure 3 This is a schematic diagram of the modal decomposition results provided in this embodiment;
[0047] Figure 4(a) is a three-dimensional time-frequency spectrum diagram provided in this embodiment;
[0048] Figure 4(b) is a two-dimensional time-frequency spectrum diagram provided in this embodiment;
[0049] Figure 5(a) is a three-dimensional diagram of the energy matrix after filtering frequency threshold-time-normalized filtering provided in this embodiment;
[0050] Figure 5(b) is a two-dimensional diagram of the energy matrix after filtering frequency threshold-time-normalized filtering provided in this embodiment;
[0051] Figure 6 This is the energy-time curve before filtering provided in this embodiment;
[0052] Figure 7 This is the energy-time curve before normalized filtering provided in this embodiment;
[0053] Figure 8(a) is a three-dimensional submatrix diagram of the normalized filter frequency threshold-time-filtered energy matrix within the key time period provided in this embodiment;
[0054] Figure 8(b) is a two-dimensional submatrix diagram of the normalized filter frequency threshold-time-filtered energy matrix within the key time period provided in this embodiment;
[0055] Figure 9 This embodiment provides a curve of the filter frequency threshold and the mean energy after normalization filtering.
[0056] Figure 10 This is a schematic diagram showing the location of the maximum value point provided in this embodiment;
[0057] Figure 11 This is a schematic diagram of the system structure of the present invention. Detailed Implementation
[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0059] See Figure 1 This invention discloses a high-frequency automated identification method for abnormal vibration in PCCP pipelines, comprising:
[0060] S1: Monitor abnormal vibration data of PCCP pipeline, perform modal decomposition and Hilbert spectrum calculation on the abnormal vibration data of PCCP pipeline to obtain the matrix and time spectrum characterizing time-frequency-energy;
[0061] S2: Perform high-pass filtering based on the frequency resolution of the time spectrum, and calculate the vibration energy as a function of time after each filtering to obtain the vibration energy-time curve;
[0062] S3: Normalize the vibration energy-time curve after each filtering, and embed the normalized one-dimensional vibration energy data into the dimension to form a filtering threshold-time-filtered energy matrix;
[0063] Specifically, the vibration energy-time curve after each filtering is 1-dimensional data. Combining all the 1-dimensional data into a matrix transforms the 1-dimensional data into multi-dimensional data, forming a filtering threshold-time-filtered energy matrix.
[0064] S4: Select the target time period in the energy-time curve before filtering, and filter out the sub-matrix of the filtering threshold-time-filtered energy matrix corresponding to the target time period;
[0065] Specifically, the method for obtaining the energy-time curve before filtering is as follows: based on the time spectrum before filtering, the energy values corresponding to all frequency points at each time point are superimposed to obtain the energy value at each time point, thereby forming the energy-time curve before filtering.
[0066] S5: Select the upper quartile of the row or column corresponding to each filtering frequency threshold of the submatrix, and average the normalized energy values that are greater than the upper quartile for each frequency threshold to form the filtering frequency threshold-normalized filtered energy mean curve.
[0067] Specifically, quartiles are a concept in statistics used to describe the central tendency of a set of data.
[0068] Specifically, the upper quartile (Q3) is the third quartile, meaning that 25% of the data are smaller than it, while 75% of the data are larger than it.
[0069] S6: Calculate the maximum value on the filter frequency threshold-normalized filtered mean energy curve, and subtract the frequency resolution from the filter frequency threshold corresponding to the maximum value. The maximum value minus the frequency resolution is the highest frequency of abnormal vibration.
[0070] In one specific embodiment, the mode decomposition used in S1 includes adding white noise.
[0071] In one specific embodiment, the mode decomposition method used is a mode decomposition method based on Empirical Mode Decomposition (EMD). The mode decomposition method must be a mode decomposition method that includes a white noise step, such as EEMD, CEEMD, CEEMDAN, and ICEEMDAN, but the mode decomposition method that includes a white noise step is not limited to the mentioned mode decomposition methods.
[0072] In one specific embodiment, the high-pass filtering in S2 is used to filter out components that are below the filtering frequency threshold corresponding to the time-frequency-energy matrix.
[0073] In one specific embodiment, S3 normalizes the vibration energy-time curve after each filtering, including the following steps:
[0074] Let [x] be the one-dimensional numerical sequence of the filtered vibration energy as a function of time. n The sequence has a length of n, and the normalized sequence is [x']. n Calculate the sequence [x] n The maximum value of max(x);
[0075] For each sequence [x] n Normalize each data point in the dataset using the following formula:
[0076]
[0077] In the formula, the subscript i represents the i-th data point.
[0078] In one specific embodiment, selecting the target time period from the energy-time curve before filtering includes the following steps:
[0079] Normalize the energy before filtering;
[0080] The time period corresponding to the normalized, unfiltered energy being less than the threshold Tv is taken as the target time period.
[0081] In one specific embodiment, normalizing the energy before filtering includes the following steps:
[0082] Let [y] be the one-dimensional numerical sequence of energy change over time before filtering. n The sequence length is n, and the normalized sequence is [y']. n Calculate [y] n The maximum value of y is max(y) and the minimum value of y is min(y).
[0083] [y] n Subtract the minimum value min(y) from each data point in the sequence to obtain a new sequence. The formula is:
[0084]
[0085] In the formula, the subscript i represents the i-th data point;
[0086] calculate Maximum value
[0087] right Normalization is performed using the following formula:
[0088]
[0089] In the formula, It is a one-dimensional numerical sequence after removing the minimum value; for The i-th element in the middle; y i ' represents the i-th element of the normalized numerical sequence.
[0090] In one specific embodiment, a high-frequency automated identification method for abnormal vibration in a PCCP pipeline can be implemented as follows:
[0091] S1: Modal decomposition is performed on the monitored wire breakage vibration data, and Hilbert spectrum calculation is performed to obtain the time spectrum. In this embodiment, the sampling frequency of the wire breakage vibration is 10000Hz, and the modal decomposition method used in this embodiment is ICEEMDAN. The wire breakage vibration data is as follows: Figure 2 As shown, the mode decomposition results are as follows: Figure 3 As shown, the time spectrum is shown in Figures 4(a)-(b).
[0092] S2: Design a high-pass filter with a filtering threshold that is a multiple of the frequency resolution of the time spectrum, and calculate the vibration energy change curve over time after each filtering step. In this embodiment, the frequency resolution of the spectrum is 5 Hz.
[0093] S3: Normalize the vibration energy-time curve after each filtering, and embed the normalized one-dimensional vibration energy data into a dimension-time-normalized filtered energy matrix, as shown in Figures 5(a)-(b).
[0094] S4: Based on the energy-time curve before filtering, select key analysis time periods and filter out the sub-matrices of the normalized filter frequency threshold-time-normalized filtered energy matrix within the key time periods. First, normalize the energy before filtering; the energy-time curve before filtering is shown below. Figure 6 As shown; then, the time period corresponding to the normalized, unfiltered energy being less than the threshold Tv is taken as the key analysis period. Tv can be adjusted according to the actual measurement conditions. The energy-time curve before normalization filtering is shown in the figure. Figure 7 As shown in the figure. In this example, the value Tv = 0.1. The description of the submatrices of the normalized filter frequency threshold-time-filtered energy matrix corresponding to the key time period is shown in Figure 8(a)-(b).
[0095] S5: Select the upper quartile for each row (or column) corresponding to the filtering frequency threshold of the submatrix, and average the normalized energy values greater than the upper quartile for each frequency threshold to form a curve of filtering frequency threshold - normalized filtered energy mean, such as... Figure 9 As shown.
[0096] S6: Calculate the maximum value on the filter threshold-normalized energy mean curve. This maximum value corresponds to the filter frequency threshold minus the frequency resolution, which is the highest frequency of the abnormal vibration. The maximum value point is shown below. Figure 10 As shown, the filtering frequency threshold corresponding to this maximum value is 2373Hz, so the highest frequency of abnormal vibration is 2373-5=2368Hz.
[0097] Specifically, because the modal decomposition algorithm involved adds noise during signal decomposition, extremely low-energy noise data inevitably remains. Based on this characteristic, this invention normalizes the energy-time variation curve within the original signal's frequency band, which is equivalent to suppressing the energy of the original signal data; normalizing the energy-time variation curve of random noise in frequency bands outside the actual signal is equivalent to doubling the energy value of the random noise data. High-energy impact signals only exist for a very short period. Normalizing the energy-time variation curve within the actual signal's frequency band results in energy values far less than 1 during periods of intensified vibration, while the signal energy in the frequency band containing only random noise is multiplied to 1. At this point, there is a significant difference between the frequency band of the actual signal and the frequency band containing only random noise. This difference can characterize the highest frequency of the effective vibration signal, thus facilitating the accurate identification of abnormal high-frequency vibrations.
[0098] See Figure 11 This embodiment discloses a high-frequency automated identification system for abnormal vibration in PCCP pipelines, comprising:
[0099] The monitoring module monitors abnormal vibration data of the PCCP pipeline, performs modal decomposition and Hilbert spectrum calculation on the abnormal vibration data of the PCCP pipeline, and obtains a matrix and time spectrum characterizing time-frequency-energy.
[0100] The filtering module performs high-pass filtering based on the frequency resolution of the time spectrum, and superimposes the energy values corresponding to all frequency points at each time point after filtering to obtain the energy value at each time point, forming the filtered vibration energy-time curve.
[0101] The normalization module normalizes the vibration energy-time curve after each filtering and transforms the normalized one-dimensional vibration energy data into multi-dimensional data, forming a filtering threshold-time-filtered energy matrix.
[0102] The module selects the energy values corresponding to all frequency points at each time point based on the time spectrum of the unfiltered data, and obtains the energy value at each time point to form the energy-time curve before filtering. The target time period in the energy-time curve before filtering is selected, and the sub-matrix of the filtering threshold-time-filtered energy matrix corresponding to the target time period is filtered out.
[0103] The statistics module selects the upper quartile of the row or column corresponding to each filtering frequency threshold of the submatrix, and averages the normalized energy values that are greater than the upper quartile for each frequency threshold, forming a curve of filtering frequency threshold - normalized filtered energy mean.
[0104] The calculation module selects the maximum value on the filter frequency threshold-normalized filtered mean energy curve, and subtracts the frequency resolution from the filter frequency threshold corresponding to the maximum value to obtain the highest frequency of abnormal vibration in the PCCP pipeline.
[0105] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a high-frequency automated identification method for abnormal vibration of PCCP pipelines.
[0106] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.
[0107] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A high-frequency automated identification method for abnormal vibration in PCCP pipelines, characterized in that, include: S1: Monitor abnormal vibration data of PCCP pipeline, perform modal decomposition and Hilbert spectrum calculation on the abnormal vibration data of PCCP pipeline to obtain the matrix and time spectrum characterizing time-frequency-energy; S2: High-pass filtering is performed based on the frequency resolution of the time spectrum. The energy values corresponding to all frequency points at each time point after filtering are superimposed to obtain the energy value at each time point, forming the filtered vibration energy-time curve. S3: Normalize the vibration energy-time curve after each filtering, and convert the normalized one-dimensional vibration energy data into multi-dimensional data to form a filtering threshold-time-filtered energy matrix; S4: Based on the time spectrum of the unfiltered data, superimpose the energy values corresponding to all frequency points at each time point to obtain the energy value at each time point, forming the energy-time curve before filtering. Select the target time period in the energy-time curve before filtering and filter out the sub-matrix of the filtering threshold-time-filtered energy matrix corresponding to the target time period. S5: Select the upper quartile of the row or column corresponding to each filtering frequency threshold of the submatrix, and average the normalized energy values that are greater than the upper quartile for each frequency threshold to form the filtering frequency threshold-normalized filtered energy mean curve. S6: Select the maximum value on the filter frequency threshold-normalized filtered energy mean curve, and subtract the frequency resolution from the filter frequency threshold corresponding to the maximum value to obtain the highest frequency of abnormal vibration in the PCCP pipeline. Selecting the target time period from the energy-time curve before filtering includes the following steps: Normalize the energy before filtering; The energy before filtering after normalization is less than the threshold. The corresponding time period is used as the target time period.
2. The high-frequency automated identification method for abnormal vibration of PCCP pipelines according to claim 1, characterized in that, The mode decomposition used in S1 includes the addition of white noise.
3. The high-frequency automated identification method for abnormal vibration of PCCP pipelines according to claim 1, characterized in that, In S2, the high-pass filter removes components below the filtering frequency threshold corresponding to the time-frequency-energy matrix.
4. The high-frequency automated identification method for abnormal vibration of PCCP pipelines according to claim 1, characterized in that, The S3 step involves normalizing the vibration energy-time curve after each filtering process, including the following steps: Let the filtered one-dimensional numerical sequence of vibration energy changing with time be... The length of the sequence is The normalized sequence is Calculate the sequence maximum value ; For each one-dimensional numerical sequence Normalize each data point in the dataset using the following formula: In the formula, the subscript Indicates the first Data points.
5. The high-frequency automated identification method for abnormal vibration of PCCP pipelines according to claim 1, characterized in that, Normalizing the energy before filtering includes the following steps: Let the one-dimensional numerical sequence of energy change over time before filtering be... The sequence length is The normalized sequence is ,calculate maximum value and minimum value ; right Subtract the minimum value from each data point. A new sequence is obtained. The formula is: ； In the formula, the subscript Representing the One data point; calculate Maximum value ; right Normalization is performed using the following formula: ； In the formula, It is a one-dimensional numerical sequence after removing the minimum value; for The Middle i One element; For the normalized numerical sequence, the first... i Each element.
6. A high-frequency automated identification system for abnormal vibration of PCCP pipelines using the high-frequency automated identification method for abnormal vibration of PCCP pipelines according to any one of claims 1-5, characterized in that, include: The monitoring module monitors abnormal vibration data of the PCCP pipeline, performs modal decomposition and Hilbert spectrum calculation on the abnormal vibration data of the PCCP pipeline, and obtains a matrix and time spectrum characterizing time-frequency-energy. The filtering module performs high-pass filtering based on the frequency resolution of the time spectrum, and superimposes the energy values corresponding to all frequency points at each time point after filtering to obtain the energy value at each time point, forming the filtered vibration energy-time curve. The normalization module normalizes the vibration energy-time curve after each filtering and transforms the normalized one-dimensional vibration energy data into multi-dimensional data, forming a filtering threshold-time-filtered energy matrix. The module selects the energy values corresponding to all frequency points at each time point based on the time spectrum of the unfiltered data, and obtains the energy value at each time point to form the energy-time curve before filtering. The target time period in the energy-time curve before filtering is selected, and the sub-matrix of the filtering threshold-time-filtered energy matrix corresponding to the target time period is filtered out. The statistics module selects the upper quartile of the row or column corresponding to each filtering frequency threshold of the submatrix, and averages the normalized energy values that are greater than the upper quartile for each frequency threshold, forming a curve of filtering frequency threshold - normalized filtered energy mean. The calculation module selects the maximum value on the filter frequency threshold-normalized filtered energy mean curve, and subtracts the frequency resolution from the filter frequency threshold corresponding to the maximum value to obtain the highest frequency of abnormal vibration of the PCCP pipeline. Selecting the target time period from the energy-time curve before filtering includes the following steps: Normalize the energy before filtering; The energy before filtering after normalization is less than the threshold. The corresponding time period is used as the target time period.
7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the high-frequency automated identification method for abnormal vibration of PCCP pipelines as described in any one of claims 1 to 5.