A method, device and equipment for detecting a shock cycle of a vibration signal

Abstract

The application relates to the technical field of rotating machinery equipment fault diagnosis, and specifically discloses a vibration signal impact cycle detection method, device and equipment, which comprises the following steps: firstly, performing windowing processing on the vibration signal of the rotating machinery equipment, and extracting an upper envelope equal in length to the windowed signal; secondly, determining the estimated impact cycle of the windowed signal through an unbiased autocorrelation function and a biased autocorrelation function; then, obtaining the impact sequence from the upper envelope of the windowed signal through morphological top-hat transformation, and extracting a first principal feature component from the extended feature sequence by using a PCA algorithm to obtain an abnormal score sequence, so as to determine the impact index of the windowed signal by using a clustering algorithm; finally, determining whether the period of the impact in the windowed signal is found by judging whether the ratio of the maximum length of the arithmetic sequence in the impact index of the windowed signal to the estimated impact number exceeds a specified predetermined proportion threshold.

Description

Technical Field

[0001] The invention relates to the technical field of fault diagnosis of rotating mechanical equipment, and in particular discloses a vibration signal impact period detection method, device and equipment. Background Art

[0002] Periodic shocks often occur in the vibration signals of rotating machinery faults (such as metal collisions, crack propagation, and gear tooth breakage). Accurately detecting the period of these shocks is crucial for identifying the fault type (e.g., inner race, outer race, or rolling element faults). The periodicity of the shocks is primarily reflected in time domain intervals, rather than at a single frequency point. Commonly used techniques such as cepstrum analysis and spectrum analysis have significant limitations when processing such signals. Cepstrum analysis is highly sensitive to the periodicity of harmonic families, but has limited ability to detect the periodicity of transient shock sequences. The shock response is inherently a broadband signal, its energy dispersed across a wide frequency range. The shock energy generated by early faults is weak and easily overwhelmed by strong background noise (e.g., mechanical operation noise, electromagnetic interference) and other vibration sources (e.g., gear meshing, imbalance), rendering the shock signature invisible in the spectrum. Summary of the Invention

[0003] The present invention provides an impact period detection method with stronger robustness and higher noise resistance, which can effectively overcome the inherent defects of spectrum analysis and cepstrum analysis, and provide reliable technical support for early fault warning and accurate diagnosis of rotating machinery.

[0004] The present invention specifically provides a method for detecting the impact period of a vibration signal, the method comprising the following steps:

[0005] S101, obtaining a vibration signal of a rotating mechanical device, performing windowing processing on the vibration signal to obtain a windowed signal, and extracting an upper envelope having a sample point length consistent with that of the windowed signal; wherein the time length of the windowed signal is determined according to the impact cycle detection requirement;

[0006] S102: setting an allowable range of the estimated impact period according to the impact period detection requirement, calculating an unbiased autocorrelation coefficient sequence consistent with the sample point length of the windowed signal using an unbiased autocorrelation function on the upper envelope of the windowed signal, then determining the estimated impact period of the windowed signal using a biased autocorrelation function on the unbiased autocorrelation coefficient sequence, and obtaining an estimated number of impacts by dividing the sample point length of the windowed signal by the estimated impact period;

[0007] S103, extracting the impulse sequence of the windowed signal from the upper envelope of the windowed signal, generating an extended feature sequence using the maximum-minimum normalized impulse sequence, calculating the anomaly score of each point in the impulse sequence using the PCA algorithm and the extended feature sequence to generate an anomaly score sequence, clustering the indices corresponding to all elements in the anomaly score sequence that exceed the anomaly score threshold using a preset anomaly score threshold, and determining the impulse index of the windowed signal based on the index corresponding to the maximum anomaly score in each cluster;

[0008] S104, searching for an arithmetic sequence in the impact index of the windowed signal according to the estimated impact period, and determining whether the period of the impact in the windowed signal is found by judging whether the ratio of the maximum length of the arithmetic sequence in the impact index of the windowed signal to the estimated number of impacts exceeds a specified predetermined ratio threshold.

[0009] Preferably, the impulse sequence is a sequence of the windowed signals obtained by subtracting a base extraction of the upper envelope of the windowed signal from the upper envelope of the windowed signal.

[0010] Preferably, the maximum lag of the unbiased autocorrelation function is set to the number of sample points of the windowed signal minus 1; the maximum lag of the biased autocorrelation function is set to half the number of sample points of the windowed signal minus 1.

[0011] Preferably, the expression of the biased autocorrelation function is:

[0012] ,

[0013] in , represents the unbiased autocorrelation coefficient sequence corresponding to the upper envelope of the current windowed signal, is a positive integer representing the hysteresis, Indicates the total number of sample points of the envelope of the current windowed signal, Dummy variable representing the summation operation.

[0014] Preferably, the top-hat transform in morphology is used to extract the basis of the envelope of the windowed signal, and the window length of the top-hat transform is set according to the sampling rate of the windowed signal.

[0015] Preferably, the square, cubic, fourth and fifth powers of each element in the maximum-minimum normalized impact sequence are calculated, and then together with the maximum-minimum normalized impact sequence, an extended feature sequence with a dimension of 5 is formed in index order. The PCA algorithm and the extended feature sequence are used to calculate the anomaly score of each point in the impact sequence to generate an anomaly score sequence. A clustering algorithm is used to cluster the indices corresponding to all elements in the anomaly score sequence that exceed the preset anomaly score threshold. The impact index in the windowed signal is determined according to the index corresponding to the maximum anomaly score in each cluster.

[0016] Preferably, it also includes a preset proportional threshold, and the estimated number of impacts of the windowed signal is obtained by dividing the total number of sample points of the windowed signal by the estimated impact period, and it is determined whether the ratio of the length of the arithmetic sequence to the estimated number of impacts of the windowed signal exceeds the predetermined proportional threshold.

[0017] A device for detecting an impact period of a vibration signal, comprising:

[0018] an envelope acquisition unit, configured to acquire a vibration signal of the rotating mechanical equipment, perform windowing processing on the vibration signal, select a time length of the windowed signal according to the impact period detection requirement, and extract an upper envelope having the same sample point length as the windowed signal;

[0019] An estimated impact period unit is configured to set an allowable range of the estimated impact period according to the impact period detection requirement, calculate an unbiased autocorrelation coefficient sequence consistent with the sample point length of the windowed signal using an unbiased autocorrelation function on the upper envelope of the windowed signal, then determine the estimated impact period of the windowed signal using a biased autocorrelation function on the unbiased autocorrelation coefficient sequence, and obtain an estimated number of impacts by dividing the sample point length of the windowed signal by the estimated impact period;

[0020] An impact index extraction unit is configured to extract an impact sequence of the windowed signal from the upper envelope of the windowed signal, generate an extended feature sequence using the impact sequence after maximum and minimum normalization, calculate an anomaly score for each point in the impact sequence using a PCA algorithm and the extended feature sequence to generate an anomaly score sequence, cluster indices corresponding to all elements in the anomaly score sequence that exceed the anomaly score threshold by presetting the anomaly score threshold, and determine an impact index of the windowed signal based on the index corresponding to the maximum anomaly score in each cluster;

[0021] A judgment unit is used to search for an arithmetic sequence in the impact index of the windowed signal according to the estimated impact period, and determine whether the period of the impact in the windowed signal is found by judging whether the ratio of the maximum length of the arithmetic sequence in the impact index of the windowed signal to the estimated number of impacts exceeds a specified predetermined ratio threshold.

[0022] Preferably, it includes a memory and a processor, wherein the memory stores a computer program, and the computer program can be executed by the processor to implement the above-mentioned method for detecting the impact period of the vibration signal.

[0023] By adopting the above scheme, the present invention has the following advantages and beneficial effects: the present invention first performs window processing on the vibration signal of the rotating mechanical equipment and extracts an upper envelope of the same length as the window signal; secondly, the estimated impact period of the window signal is determined by an unbiased autocorrelation function and a biased autocorrelation function; then the impact sequence is obtained from the upper envelope of the window signal by morphological top hat transformation, and the first principal characteristic component is extracted from the expanded characteristic sequence using the PCA algorithm to obtain the abnormal score sequence, thereby determining the impact index of the window signal using a clustering algorithm; finally, by judging whether the ratio of the maximum length of the arithmetic sequence in the impact index of the window signal to the estimated number of impacts exceeds a specified predetermined ratio threshold, it is determined whether the impact period in the window signal is found. The present invention can effectively overcome the inherent defects of spectral analysis and cepstrum analysis in the prior art, and provide reliable technical support for early fault warning and accurate diagnosis of rotating machinery. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic flow chart of a vibration signal shock period detection method based on envelope autocorrelation provided in the first embodiment of the present invention;

[0025] Figure 2 This is a schematic structural diagram of a vibration signal impact period detection device based on envelope autocorrelation provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the invention for which protection is sought, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0027] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0028] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature identified as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0029] In the present invention, unless otherwise expressly specified or limited, the terms "mounted," "connected," "connect," "fixed," etc. should be understood broadly. For example, they may refer to fixed connection, detachable connection, or integral connection; mechanical connection or electrical connection; direct connection or indirect connection through an intermediate medium; or internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0030] The following describes in detail the preferred embodiments of the present invention in conjunction with the accompanying drawings, so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making a clearer and more precise definition of the protection scope of the present invention.

[0031] refer to Figure 1 As shown, the first embodiment of the present invention discloses a vibration signal impact period detection method, which can be performed by a vibration signal impact period detection device based on envelope autocorrelation (hereinafter referred to as a period detection device), and in particular, is executed by one or more processors in the period detection device to implement the following method:

[0032] S101, obtaining a vibration signal of a rotating mechanical device, performing window processing on the vibration signal, selecting a time length of the window signal according to an impact cycle detection requirement, and extracting an upper envelope having a sample point length consistent with the window signal;

[0033] S102: setting an allowable range of the estimated impact period according to the impact period detection requirement, calculating an unbiased autocorrelation coefficient sequence consistent with the sample point length of the windowed signal using an unbiased autocorrelation function on the upper envelope of the windowed signal, then determining the estimated impact period of the windowed signal using a biased autocorrelation function on the unbiased autocorrelation coefficient sequence, and obtaining an estimated number of impacts by dividing the sample point length of the windowed signal by the estimated impact period;

[0034] In this embodiment, the period detection device may be a computer, a server, a workstation, a gateway, or other device with data processing capabilities, which is not specifically limited in the present invention.

[0035] In this embodiment, a vibration sensor is first installed at a suitable location of the rotating machine equipment to be identified. Then, according to the current scenario requirements, the sensor sampling rate and acquisition time length of the vibration sensor are set to adapt to the scenario requirements. For example, the sampling rate of the vibration sensor is set to 25600 and the acquisition time length is set to 50 seconds.

[0036] Then, preprocessing the vibration time domain signal of the machine equipment using a Butterworth filter with a set filter order and frequency cutoff range, wherein, in particular, the filter order is 34 and the frequency cutoff range is [0.1, 10000], but not limited thereto;

[0037] Assuming that the impact period detection requirement is to determine whether the vibration signal experiences two impacts per second, i.e., the impact frequency is 2 Hz, then the time length of each windowed signal can be set to 2 seconds. In this way, a total of 25 windowed signals are included in the vibration signal. Then, the Hilbert transform is used to extract the upper envelope that is consistent with the sample point length of the windowed signal. The following processing flow is applied to each windowed signal.

[0038] The envelope of the windowed signal is normalized using the mean and standard deviation of the envelope of the windowed signal, and is recorded as , ,in Represents the total number of sample points of the envelope of the windowed signal, where ; Calculate the unbiased autocorrelation coefficient sequence corresponding to the envelope of the windowed signal using the following unbiased autocorrelation function:

[0039]

[0040] in, is a positive integer representing the hysteresis, the maximum hysteresis is 51199, and , the length of the unbiased autocorrelation coefficient sequence is 51200;

[0041] Since the impact frequency to be detected is 2 Hz and the sampling rate of the windowed signal is 25600, the biased autocorrelation coefficient sequence corresponding to the unbiased autocorrelation coefficient sequence is calculated using the following calculation formula for the biased autocorrelation function:

[0042]

[0043] in, Is a positive integer representing the hysteresis, where the maximum hysteresis is 25599;

[0044] Next, we will use the hysteresis The biased autocorrelation coefficient sequence between determines the estimated impact period of each of the windowed signals, specifically:

[0045] According to domain knowledge and business experience, a suitable impact period search range can be set. Taking the impact frequency equal to 2Hz as an example, in fact, the window signal sampling rate shows that the impact period is about 12800. Therefore, in the first embodiment of the present invention, the allowable range of the estimated impact period is set to , by collecting the biased autocorrelation coefficient series in The hysteresis corresponding to all local maximum points between the two, and finally selecting the hysteresis corresponding to the maximum biased autocorrelation coefficient from the local maximum point as the estimated impact period;

[0046] S103, extracting the impulse sequence of the windowed signal from the upper envelope of the windowed signal, generating an extended feature sequence using the maximum-minimum normalized impulse sequence, calculating the anomaly score of each point in the impulse sequence using the PCA algorithm and the extended feature sequence to generate an anomaly score sequence, clustering the indices corresponding to all elements in the anomaly score sequence that exceed the anomaly score threshold using a preset anomaly score threshold, and determining the impulse index of the windowed signal based on the index corresponding to the maximum anomaly score in each cluster;

[0047] In this embodiment, specifically:

[0048] First, the top hat transform in morphology is used to extract the basis of the envelope of the windowed signal, wherein the window length of the top hat transform can be set to the sampling rate. , i.e., 640 sample points;

[0049] Next, extracting the impulse sequence of the windowed signal by subtracting the basis of the upper envelope of the windowed signal from the upper envelope of the windowed signal;

[0050] Then, in order to suppress the high-frequency noise in the impact sequence, the square, cubic, fourth, and fifth powers of each element in the maximum and minimum normalized impact sequence are calculated, and then together with the maximum and minimum normalized impact sequence, an extended feature sequence with a dimension of 5 is formed in index order. The index of the extended feature sequence starts from 0.

[0051] Secondly, the PCA algorithm is used to reduce the dimension of the extended feature sequence, extract the first principal component vector, and then the Euclidean distance between each point in the extended feature and the first principal component vector is calculated as the anomaly score to generate an anomaly score sequence of the same length as the extended feature sequence;

[0052] Next, a triangular threshold method is applied to the anomaly score sequence to obtain an anomaly score threshold, and a weight coefficient is set to adjust the size of the anomaly score threshold;

[0053] Finally, on the one hand, the DBSCAN algorithm is used to cluster the indexes corresponding to all elements in the anomaly score sequence that exceed the anomaly score threshold. After removing noise points, the index corresponding to the maximum anomaly score in each cluster is collected and sorted according to the size of the anomaly score to generate candidate impact indexes. On the other hand, the estimated number of impacts of the windowed signal is obtained by dividing the total number of sample points of the windowed signal by the estimated impact period in step S103. Then, the index with the largest anomaly score that is the same as the estimated number of impacts is screened out from the sorted candidate impact indexes as the impact index of the windowed signal.

[0054] S104, searching for an arithmetic sequence in the impact index of the windowed signal according to the estimated impact period, and determining whether the impact period in the windowed signal is found by judging whether the ratio of the maximum length of the arithmetic sequence in the impact index of the windowed signal to the estimated number of impacts exceeds a specified predetermined ratio threshold.

[0055] In this embodiment, specifically:

[0056] First, the impact index of the windowed signal is sorted in ascending order. Assume that the impact index of the sorted windowed signal is ;

[0057] Secondly, the estimated impact period is used as the tolerance , determine whether there is a sufficiently long impulse index in the sorted window signal is an arithmetic sequence with a common difference, and the upper bound of the index of the arithmetic sequence is , the index error ratio value ;

[0058] Then, any two different impact indices in the impact indices of the sorted windowed signals are combined in ascending order to obtain For example, , verify whether the following inequality holds:

[0059]

[0060] in Indicates rounding operation, the above inequality judgment Whether they belong to the same arithmetic sequence. If the inequality does not hold, then determine the next combination;

[0061] If the inequality holds, then perform the following steps, specifically:

[0062] First of all, for ,use right Expand as follows, , making

[0063]

[0064]

[0065] Established, this will result in Candidate arithmetic sequence with the first term , where the candidate arithmetic sequence The value of each element in is less than the upper bound of the index of the arithmetic sequence , where the total number of sample points of the windowed signal is set as the upper bound of the index of the arithmetic sequence, that is, ;

[0066] Secondly, traverse the Candidate arithmetic sequence with the first term Each index in is recorded as , determine the impact index of the windowed signal after the sorting Is there a The difference ratio is less than Index , that is, determine whether the following inequality holds:

[0067]

[0068] If it exists, use the Replace the If it does not exist, then Candidate arithmetic sequence with the first term Delete the , repeat the above comparison process, and finally obtain a Repeat the above process to find all the impulse indices of the sorted windowed signal with is an arithmetic sequence of common differences;

[0069] Then, a ratio threshold is preset, and the estimated number of impacts of the window signal is obtained by dividing the total number of sample points of the window signal by the estimated impact period in step 103, and the estimated number of impacts of the window signal is determined. whether the ratio of the maximum length of the arithmetic sequence with a tolerance to the estimated number of shocks of the windowed signal exceeds the predetermined ratio threshold; if so, determining that the periodic shock with a frequency of 2 Hz can be detected in the vibration signal; if not, determining the next windowed signal until all windowed signals are traversed; if all windowed signals are determined to be negative, determining that the periodic shock with a frequency of 2 Hz does not exist in the vibration signal;

[0070] Preferably, based on business knowledge and domain knowledge, the index error ratio Can be set between 0.02 and 0.1;

[0071] Preferably, the preset ratio threshold can be set between 0.3 and 1;

[0072] To summarize, this embodiment first performs windowing processing on the vibration signal of the rotating mechanical equipment and extracts an upper envelope of the same length as the windowed signal; secondly, the estimated impact period of the windowed signal is determined by an unbiased autocorrelation function and a biased autocorrelation function; then, the impact sequence is obtained from the upper envelope of the windowed signal by a morphological top-hat transform, and the first principal feature component is extracted from the extended feature sequence using the PCA algorithm to obtain the abnormal score sequence, thereby determining the impact index of the windowed signal using the DBSCAN algorithm; finally, by judging whether the ratio of the maximum length of the arithmetic sequence in the impact index of the windowed signal to the estimated number of impacts exceeds a specified predetermined ratio threshold, it is determined whether the impact period in the windowed signal is found.

[0073] See also Figure 2 The second embodiment of the present invention further provides a vibration signal impact period detection device based on envelope autocorrelation, which is characterized by comprising:

[0074] An envelope acquisition unit 210 is configured to acquire a vibration signal of the rotating mechanical device, perform windowing processing on the vibration signal, select a time length of the windowed signal based on the impact period detection requirement, and extract an upper envelope having a sample point length consistent with the windowed signal;

[0075] The estimated impact period unit 220 is used to set an allowable range of the estimated impact period according to the impact period detection requirement, use an unbiased autocorrelation function on the upper envelope of the windowed signal to calculate an unbiased autocorrelation coefficient sequence that is consistent with the sample point length of the windowed signal, then use a biased autocorrelation function on the unbiased autocorrelation coefficient sequence to determine the estimated impact period of the windowed signal, and use the sample point length of the windowed signal divided by the estimated impact period to obtain the estimated number of impacts;

[0076] An impact index extraction unit 230 is configured to extract an impact sequence of the windowed signal from the upper envelope of the windowed signal, generate an extended feature sequence using the impact sequence after maximum and minimum normalization, calculate an anomaly score for each point in the impact sequence using a PCA algorithm and the extended feature sequence to generate an anomaly score sequence, cluster the indices corresponding to all elements in the anomaly score sequence that exceed the anomaly score threshold using a preset anomaly score threshold, and determine the impact index of the windowed signal based on the index corresponding to the maximum anomaly score in each cluster;

[0077] The judgment unit 240 is used to search for an arithmetic sequence in the impact index of the windowed signal according to the estimated impact period, and determine whether the period of the impact in the windowed signal is found by judging whether the ratio of the maximum length of the arithmetic sequence in the impact index of the windowed signal to the estimated number of impacts exceeds a specified predetermined ratio threshold.

[0078] The third embodiment of the present invention also provides a vibration signal impact period detection device based on envelope autocorrelation, which includes a memory and a processor, wherein the memory stores a computer program, and the computer program can be executed by the processor to implement the vibration signal impact period detection method based on envelope autocorrelation as described above.

[0079] The fourth embodiment of the present invention further provides a computer-readable storage medium storing a computer program, which can be executed by a processor of the device where the computer-readable storage medium is located to implement the vibration signal impact period detection method based on envelope autocorrelation as described above.

[0080] Illustratively, the above-mentioned various devices and various process steps can be implemented by a computer program. The computer program can be divided into one or more units. The one or more units are stored in the memory and executed by the processor to complete the present invention.

[0081] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0082] The memory can be used to store the computer programs and / or modules. The processor implements the various functions of the present invention by running or executing the computer programs and / or modules stored in the memory and accessing the data stored in the memory. The memory may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function (such as a sound playback function, an image playback function, etc.); the data storage area may store data generated based on the use of the mobile phone (such as audio data, a phone book, etc.). Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0083] If the integrated unit of the electronic device or printer is implemented as a software functional unit and sold or used as a standalone product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention can implement all or part of the process steps in the above-mentioned method embodiments by using a computer program to instruct the relevant hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium. It should be noted that the content of the computer-readable medium can be appropriately increased or decreased based on the requirements of legislation and patent practice in a jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media does not include electric carrier signals and telecommunication signals.

[0084] It should be noted that 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 displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement the present invention without inventive effort.

[0085] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for detecting the impact period of a vibration signal, characterized in that: The method comprises the following steps: S101, obtaining a vibration signal of a rotating mechanical device, performing window processing on the vibration signal to obtain a windowed signal, and extracting an upper envelope having a sample point length consistent with that of the windowed signal; wherein the time length of the windowed signal is determined according to the impact cycle detection requirement; S102: setting an allowable range of the estimated impact period according to the impact period detection requirement, calculating an unbiased autocorrelation coefficient sequence consistent with the sample point length of the windowed signal using an unbiased autocorrelation function on the upper envelope of the windowed signal, then determining the estimated impact period of the windowed signal using a biased autocorrelation function on the unbiased autocorrelation coefficient sequence, and obtaining an estimated number of impacts by dividing the sample point length of the windowed signal by the estimated impact period; S103, extracting the impulse sequence of the windowed signal from the upper envelope of the windowed signal, generating an extended feature sequence using the maximum-minimum normalized impulse sequence, calculating the anomaly score of each point in the impulse sequence using the PCA algorithm and the extended feature sequence to generate an anomaly score sequence, clustering the indices corresponding to all elements in the anomaly score sequence that exceed the anomaly score threshold using a preset anomaly score threshold, and determining the impulse index of the windowed signal based on the index corresponding to the maximum anomaly score in each cluster; S104, searching for an arithmetic sequence in the impact index of the windowed signal according to the estimated impact period, and determining whether the period of the impact in the windowed signal is found by judging whether the ratio of the maximum length of the arithmetic sequence in the impact index of the windowed signal to the estimated number of impacts exceeds a specified predetermined ratio threshold.

2. The method for detecting the impact period of a vibration signal according to claim 1, wherein: The impulse sequence is a sequence of the windowed signals obtained by subtracting a base extraction of the upper envelope of the windowed signal from an upper envelope of the windowed signal.

3. The method for detecting the impact period of a vibration signal according to claim 1, wherein: The maximum lag of the unbiased autocorrelation function is set to the number of sample points of the windowed signal minus 1; the maximum lag of the biased autocorrelation function is set to half the number of sample points of the windowed signal minus 1.

4. The method for detecting the impact period of a vibration signal according to claim 2, wherein: The expression of the biased autocorrelation function is: ， in , represents the unbiased autocorrelation coefficient sequence corresponding to the upper envelope of the current windowed signal, is a positive integer representing the hysteresis, Indicates the total number of sample points of the envelope of the current windowed signal, Dummy variable representing the summation operation.

5. The method for detecting the impact period of a vibration signal according to claim 2, wherein: The top-hat transform in morphology is used to extract the basis of the envelope of the windowed signal, and the window length of the top-hat transform is set according to the sampling rate of the windowed signal.

6. The method for detecting the impact period of a vibration signal according to claim 1, wherein: The square, cubic, fourth and fifth powers of each element in the maximum-minimum normalized impact sequence are calculated, and then together with the maximum-minimum normalized impact sequence, an extended feature sequence with a dimension of 5 is formed in index order. The PCA algorithm and the extended feature sequence are used to calculate the anomaly score of each point in the impact sequence to generate an anomaly score sequence. A clustering algorithm is used to cluster the indices corresponding to all elements in the anomaly score sequence that exceed the preset anomaly score threshold. The impact index in the windowed signal is determined according to the index corresponding to the maximum anomaly score in each cluster.

7. The method for detecting the impact period of a vibration signal according to claim 1, wherein: It also includes a preset proportional threshold, which uses the total number of sample points of the windowed signal divided by the estimated impact period to obtain the estimated number of impacts of the windowed signal, and determines whether the ratio of the length of the arithmetic sequence to the estimated number of impacts of the windowed signal exceeds the preset proportional threshold.

8. A device for detecting the impact period of a vibration signal, characterized in that: include: an envelope acquisition unit, configured to acquire a vibration signal of the rotating mechanical equipment, perform windowing processing on the vibration signal, select a time length of the windowed signal according to the impact period detection requirement, and extract an upper envelope having the same sample point length as the windowed signal; An estimated impact period unit is configured to set an allowable range of the estimated impact period according to the impact period detection requirement, calculate an unbiased autocorrelation coefficient sequence consistent with the sample point length of the windowed signal using an unbiased autocorrelation function on the upper envelope of the windowed signal, then determine the estimated impact period of the windowed signal using a biased autocorrelation function on the unbiased autocorrelation coefficient sequence, and obtain an estimated number of impacts by dividing the sample point length of the windowed signal by the estimated impact period; An impact index extraction unit is configured to extract an impact sequence of the windowed signal from the upper envelope of the windowed signal, generate an extended feature sequence using the impact sequence after maximum and minimum normalization, calculate an anomaly score for each point in the impact sequence using a PCA algorithm and the extended feature sequence to generate an anomaly score sequence, cluster indices corresponding to all elements in the anomaly score sequence that exceed the anomaly score threshold by presetting the anomaly score threshold, and determine an impact index of the windowed signal based on the index corresponding to the maximum anomaly score in each cluster; A judgment unit is used to search for an arithmetic sequence in the impact index of the windowed signal according to the estimated impact period, and determine whether the period of the impact in the windowed signal is found by judging whether the ratio of the maximum length of the arithmetic sequence in the impact index of the windowed signal to the estimated number of impacts exceeds a specified predetermined ratio threshold.

9. A device for detecting the impact period of a vibration signal, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program, and the computer program can be executed by the processor to implement the method for detecting the impact period of a vibration signal as described in any one of claims 1 to 7.

Images