Method, device and equipment for locating mutation signals based on PCA algorithm

Abstract

The present invention discloses a mutation signal positioning method, device and equipment based on PCA algorithm, the method comprises: obtaining the vibration time domain signal of the machine equipment, and taking the absolute value to obtain the positive signal; extracting the amplitude and energy of the windowed signal after windowing the positive signal, and obtaining the amplitude sequence and energy sequence; using the denoising algorithm to eliminate the mutation signal in the amplitude sequence and the energy sequence to obtain the smooth amplitude sequence and the smooth energy sequence, and then obtaining the amplitude envelope sequence and the energy envelope sequence; using the amplitude envelope sequence and the energy envelope sequence to splice into a feature sequence, using the PCA algorithm to obtain the abnormal score of each point in the feature sequence, and then determining the abnormal score threshold from the abnormal score sequence, and finally judging the mutation signal and its index position in the vibration time domain signal by comparing the size relationship between the abnormal score sequence and the abnormal score threshold. The present invention can more accurately extract the mutation signal characteristics and accurately locate the detection.

Description

Technical Field

[0001] The invention relates to the field of vibration signal analysis, and in particular discloses a mutation signal positioning method, device and equipment based on a PCA algorithm. Background Art

[0002] In vibration signal analysis and industrial equipment monitoring, processing and identifying mutation signals (such as impact, collision and other events) is a very important data preprocessing and pattern recognition task. Mutation signals often appear as spikes with high amplitude and uncertain duration. The amplitude of these spikes is often several times or even dozens of times higher than the maximum amplitude of the remaining part of the vibration signal. On the one hand, such spikes will affect the calculation results of statistical features, such as root mean square value, peak-to-peak value, etc., and will also destroy the structure of the frequency domain signal, resulting in changes in the flatness of the spectrum background noise, affecting the prominence of the frequency peak, etc. On the other hand, technicians sometimes also combine mutation signals to judge the safety and stability of equipment in the working state, and deal with the production safety hazards of the equipment in a timely manner according to the changes in mutation signals.

[0003] Existing mutation signal detection methods are often implemented through complex spectrum analysis technology, which not only has cumbersome processing logic, but also requires high experience of the staff and cannot meet actual usage requirements. Summary of the invention

[0004] The present invention provides a mutation signal positioning method, device and equipment based on PCA algorithm to at least partially improve the above problems.

[0005] The embodiment of the present invention provides a mutation signal positioning method based on the PCA algorithm, which includes:

[0006] S101, obtaining a vibration time domain signal of a machine device, and then taking an absolute value of the vibration time domain signal to obtain a positive signal;

[0007] S102, performing window processing on the positive signal to obtain windowed signals, extracting the amplitudes and energies of all windowed signals, and obtaining an amplitude sequence and an energy sequence;

[0008] S103, using a denoising algorithm to eliminate the sudden change signal in the amplitude sequence and the energy sequence to obtain a smoothed amplitude sequence and a smoothed energy sequence, and then obtaining an amplitude envelope sequence and an energy envelope sequence containing the sudden change signal according to the amplitude sequence, the energy sequence, the smoothed amplitude sequence, and the smoothed energy sequence;

[0009] S104, using the amplitude envelope sequence and the energy envelope sequence to splice into a feature sequence, using the PCA algorithm to obtain the anomaly score of each point in the feature sequence to obtain an anomaly score sequence, then determining an anomaly score threshold from the anomaly score sequence, and finally determining the mutation signal and its index position existing in the vibration time domain signal by comparing the size relationship between the anomaly score sequence and the anomaly score threshold.

[0010] Preferably, in step S102, the number of windows of data points within 1 second is preset to be , then the window length = , where f is the sampling rate, symbol Indicates floor operation.

[0011] Preferably, in step S103,

[0012] Subtraction is performed in an index alignment manner, using the amplitude sequence to subtract the smoothed amplitude sequence, and the energy sequence to subtract the smoothed energy sequence, to obtain an amplitude envelope sequence and an energy envelope sequence containing mutation signals, respectively.

[0013] Preferably, in step S103, a total variation denoising algorithm is used to eliminate the sudden change signals in the amplitude sequence and the energy sequence.

[0014] Preferably, the penalty coefficients of the total variation algorithm are set to the maximum values ​​of the amplitude sequence and the energy sequence respectively.

[0015] Preferably, it also includes:

[0016] A preset amplitude threshold is used to determine whether the amplitude envelope sequence contains a sudden change signal by comparing the amplitude envelope sequence with the preset amplitude threshold; the preset amplitude threshold is set to the maximum value of the smoothed amplitude sequence multiplied by a preset weight coefficient.

[0017] Preferably, in step S104, a triangular threshold method is used to obtain an anomaly score threshold corresponding to the anomaly score sequence.

[0018] The embodiment of the present invention further provides a mutation signal positioning device based on the PCA algorithm, which includes:

[0019] A signal acquisition unit, used to acquire a vibration time domain signal of a machine device, and then take an absolute value of the vibration time domain signal to obtain a positive signal;

[0020] A window processing unit, used for performing window processing on the positive signal to obtain window signals, extracting the amplitudes and energies of all window signals, and obtaining amplitude sequences and energy sequences;

[0021] A denoising unit, used for eliminating the sudden change signal in the amplitude sequence and the energy sequence by using a denoising algorithm to obtain a smoothed amplitude sequence and a smoothed energy sequence, and then obtaining an amplitude envelope sequence and an energy envelope sequence containing the sudden change signal according to the amplitude sequence, the energy sequence, the smoothed amplitude sequence and the smoothed energy sequence;

[0022] The positioning unit is used to use the amplitude envelope sequence and the energy envelope sequence to splice into a feature sequence, use the PCA algorithm to obtain the anomaly score of each point in the feature sequence to obtain an anomaly score sequence, then determine the anomaly score threshold from the anomaly score sequence, and finally determine the mutation signal and its index position existing in the vibration time domain signal by comparing the size relationship between the anomaly score sequence and the anomaly score threshold.

[0023] An embodiment of the present invention also provides a mutation signal locating device based on the PCA algorithm, 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 mutation signal locating method based on the PCA algorithm as described above.

[0024] In summary, this embodiment performs data preprocessing, windowing operation, and total variation denoising algorithm on the vibration time domain signal to obtain the amplitude envelope sequence and the energy envelope sequence, and then uses the PCA algorithm to obtain the anomaly score sequence. Finally, the triangular threshold method is used to obtain the corresponding anomaly score threshold, so that the position of the mutation signal in the vibration time domain signal can be determined more accurately. By removing the interference of the mutation signal, it is ensured that the statistical characteristics and spectrum changes reflect the actual operating status of the equipment, so as to avoid the influence of burst signals on the long-term trend of statistical characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A schematic diagram of a process of a mutation signal localization method based on a PCA algorithm provided in the first embodiment of the present invention;

[0026] Figure 2 A schematic diagram of the structure of a mutation signal locating device based on a PCA algorithm provided in the second embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0028] refer to Figure 1As shown, the first embodiment of the present invention discloses a mutation signal locating method based on the PCA algorithm, which can be executed by a mutation signal locating device based on the PCA algorithm (hereinafter referred to as the locating device), and in particular, executed by one or more processors in the locating device to implement the following method:

[0029] S101, obtaining a vibration time domain signal of a machine device, and then taking an absolute value of the vibration time domain signal to obtain a positive signal.

[0030] S102, performing window processing on the positive signal to obtain windowed signals, extracting the amplitudes and energies of all windowed signals, and obtaining an amplitude sequence and an energy sequence.

[0031] In this embodiment, the positioning device may be a computer, a server, a workstation, a gateway or other device with data processing capability, and the present invention does not make any specific limitation thereto.

[0032] In this embodiment, a vibration sensor is first installed at a suitable position of the machine equipment to be identified, and then according to the current scene requirements, the sensor sampling rate and acquisition time length of the vibration sensor are set to adapt to the scene requirements. For example, the sampling rate f of the vibration sensor is set to 50000 and the acquisition time length is 50 seconds.

[0033] Next, the vibration time domain signal of the machine equipment is preprocessed using a Butterworth filter with a set filter order and a frequency cutoff range, and a positive signal is obtained by taking an absolute value of the vibration time domain signal; wherein, in particular, the filter order is 34, and the frequency cutoff range is [0.1, 10000], but not limited thereto;

[0034] Then, the positive signal is subjected to window processing, the number of windows for data points within 1 second is preset to 10, and the window length is determined according to the sampling rate. , where the symbol Indicates floor operation;

[0035] Finally, let the window step size be equal to the window length, extract the amplitude and energy of all the window signals to form two sequences of equal length: amplitude sequence and energy sequence, where the lengths of the amplitude sequence and energy sequence are both , then the amplitude sequence and energy frequency sequence both start with 0 as the index, The ending index.

[0036] S103, using a denoising algorithm to eliminate the sudden change signals in the amplitude sequence and the energy sequence to obtain a smoothed amplitude sequence and a smoothed energy sequence, and then obtaining an amplitude envelope sequence and an energy envelope sequence containing the sudden change signals according to the amplitude sequence, the energy sequence, the smoothed amplitude sequence, and the smoothed energy sequence.

[0037] In this embodiment, specifically:

[0038] Firstly, a total variation denoising algorithm is used to eliminate the sudden change signals in the amplitude sequence and the energy sequence to obtain an amplitude sequence and a smoothed energy sequence. The penalty coefficients of the total variation algorithm can be set to the maximum values ​​of the amplitude sequence and the energy sequence, respectively.

[0039] Then, in an index alignment manner, the amplitude sequence and energy sequence are used to correspondingly subtract the smoothed amplitude sequence and the smoothed energy sequence to obtain an amplitude envelope sequence and an energy envelope sequence containing a sudden change signal.

[0040] S104, using the amplitude envelope sequence and the energy envelope sequence to splice into a feature sequence, using the PCA algorithm to obtain the anomaly score of each point in the feature sequence, and then determining the anomaly score threshold from the anomaly score sequence, and finally determining the mutation signal and its index position existing in the vibration time domain signal by comparing the size relationship between the anomaly score sequence and the anomaly score threshold.

[0041] In this embodiment, specifically:

[0042] First, from index 0 to , concatenate the data points at the corresponding index positions in the amplitude envelope sequence and the energy envelope sequence one by one to form a feature with a dimension of 2 and a length of The characteristic sequence of

[0043] Then, the PCA algorithm is used to extract the first principal component vector from the feature sequence, and then the Euclidean distance from each point in the feature sequence to the first principal component vector is calculated as the anomaly score to generate the length Anomaly score sequence of ;

[0044] Next, a triangular threshold method is used on the anomaly score sequence to obtain an anomaly score threshold. A weight coefficient may be set to adjust the size of the anomaly score threshold according to scenario requirements.

[0045] Finally, the magnitude relationship between each data point in the anomaly score sequence and the anomaly score threshold is compared one by one in index order, the index of the data point in the anomaly score sequence that is greater than the anomaly score threshold is recorded, and the index is multiplied by the window step size to obtain the index position of the mutation signal in the vibration data waveform.

[0046] In particular, between step S103 and step S104, the following may also be included:

[0047] A preset amplitude threshold is used to determine whether the amplitude envelope sequence contains a sudden change signal by comparing the amplitude envelope sequence with the preset amplitude threshold.

[0048] Specifically:

[0049] First, the preset amplitude threshold can be set to twice the maximum value of the smoothed amplitude sequence, and whether there are data points greater than the preset amplitude threshold in the amplitude envelope sequence is determined one by one in index order;

[0050] If it exists, go to S104;

[0051] If it does not exist, determine the next vibration time domain signal.

[0052] In summary, this embodiment performs data preprocessing, windowing operation, and total variation denoising algorithm on the vibration time domain signal to obtain the amplitude envelope sequence and the energy envelope sequence, and then uses the PCA algorithm to obtain the anomaly score sequence. Finally, the triangular threshold method is used to obtain the corresponding anomaly score threshold, so that the position of the mutation signal in the vibration time domain signal can be determined more accurately. By removing the interference of the mutation signal, it is ensured that the statistical characteristics and spectrum changes reflect the actual operating status of the equipment, so as to avoid the influence of burst signals on the long-term trend of statistical characteristics.

[0053] See also Figure 2 The second embodiment of the present invention further provides a mutation signal positioning device based on the PCA algorithm, characterized in that it includes:

[0054] The signal acquisition unit 210 is used to acquire a vibration time domain signal of the machine equipment, and then take an absolute value of the vibration time domain signal to obtain a positive signal;

[0055] A window processing unit 220, configured to perform window processing on the positive signal to obtain window signals, extract the amplitudes and energies of all window signals, and obtain an amplitude sequence and an energy sequence;

[0056] A denoising unit 230, configured to use a denoising algorithm to eliminate the sudden change signal in the amplitude sequence and the energy sequence to obtain a smoothed amplitude sequence and a smoothed energy sequence, and then obtain an amplitude envelope sequence and an energy envelope sequence containing the sudden change signal according to the amplitude sequence, the energy sequence, the smoothed amplitude sequence, and the smoothed energy sequence;

[0057] The positioning unit 240 is used to use the amplitude envelope sequence and the energy envelope sequence to splice into a feature sequence, use the PCA algorithm to obtain the anomaly score of each point in the feature sequence to obtain an anomaly score sequence, and then determine the anomaly score threshold from the anomaly score sequence, and finally determine the mutation signal and its index position existing in the vibration time domain signal by comparing the size relationship between the anomaly score sequence and the anomaly score threshold.

[0058] The third embodiment of the present invention also provides a mutation signal locating device based on the PCA algorithm, 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 mutation signal locating method based on the PCA algorithm as described above.

[0059] The fourth embodiment of the present invention further provides a computer-readable storage medium storing a computer program, wherein the computer program can be executed by a processor of a device where the computer-readable storage medium is located to implement the mutation signal localization method based on the PCA algorithm as described above.

[0060] Exemplarily, the above-mentioned devices and 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.

[0061] 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) or 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, etc.

[0062] The memory can be used to store the computer program and / or module, and the processor realizes various functions of the present invention by running or executing the computer program and / or module stored in the memory, and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0063] Wherein, if the electronic device or printer integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0064] It should be noted that the device embodiments described above are merely schematic, 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 on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying 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 may understand and implement it without paying any creative effort.

[0065] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle 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 mutation signal positioning method based on PCA algorithm, characterized in that: include: S101, obtaining a vibration time domain signal of a machine device, and then taking an absolute value of the vibration time domain signal to obtain a positive signal; S102, performing window processing on the positive signal to obtain window signals, extracting the amplitude and energy of all window signals, and obtaining an amplitude sequence and an energy sequence; wherein the number of windows for data points within 1 second is preset to be , then the window length = , f is the sampling rate, symbol Indicates rounding down operation; the window step size is equal to the window length; S103, using a denoising algorithm to eliminate mutation signals in the amplitude sequence and the energy sequence to obtain a smoothed amplitude sequence and a smoothed energy sequence, and then obtaining an amplitude envelope sequence and an energy envelope sequence containing mutation signals according to the amplitude sequence, the energy sequence, the smoothed amplitude sequence, and the smoothed energy sequence; wherein, performing a difference according to an index alignment method, using the amplitude sequence to subtract the smoothed amplitude sequence, and using the energy sequence to subtract the smoothed energy sequence, to obtain an amplitude envelope sequence and an energy envelope sequence containing mutation signals respectively; using a total variation denoising algorithm to eliminate mutation signals in the amplitude sequence and the energy sequence; S104, using the amplitude envelope sequence and the energy envelope sequence to splice into a feature sequence, using the PCA algorithm to obtain the anomaly score of each point in the feature sequence to obtain an anomaly score sequence, then determining an anomaly score threshold from the anomaly score sequence, and finally determining the mutation signal and its index position existing in the vibration time domain signal by comparing the size relationship between the anomaly score sequence and the anomaly score threshold; wherein the triangular threshold method is used to obtain the anomaly score threshold corresponding to the anomaly score sequence.

2. The mutation signal localization method based on PCA algorithm according to claim 1, characterized in that: The penalty coefficients of the total variation algorithm are respectively set to the maximum values ​​of the amplitude sequence and the energy sequence.

3. The mutation signal localization method based on PCA algorithm according to claim 1, characterized in that: Also includes: Preset an amplitude threshold, and determine whether the amplitude envelope sequence contains a sudden change signal by comparing the amplitude envelope sequence with the preset amplitude threshold; The preset amplitude threshold is set to be the maximum value of the smoothed amplitude sequence multiplied by a preset weight coefficient.

4. A mutation signal positioning device based on PCA algorithm, characterized in that: include: A signal acquisition unit, used to acquire a vibration time domain signal of a machine device, and then take an absolute value of the vibration time domain signal to obtain a positive signal; The window processing unit is used to perform window processing on the positive signal to obtain window signals, extract the amplitude and energy of all window signals, and obtain amplitude sequences and energy sequences; wherein the number of windows for data points within 1 second is preset to be , then the window length = , where f is the sampling rate, symbol Indicates rounding down operation; the window step size is equal to the window length; A denoising unit, used for eliminating mutation signals in the amplitude sequence and the energy sequence by using a denoising algorithm to obtain a smoothed amplitude sequence and a smoothed energy sequence, and then obtaining an amplitude envelope sequence and an energy envelope sequence containing mutation signals according to the amplitude sequence, the energy sequence, the smoothed amplitude sequence and the smoothed energy sequence; wherein, the amplitude envelope sequence and the energy envelope sequence containing mutation signals are obtained by performing a difference according to an index alignment method, using the amplitude sequence to subtract the smoothed amplitude sequence, and using the energy sequence to subtract the smoothed energy sequence; and using a total variation denoising algorithm to eliminate mutation signals in the amplitude sequence and the energy sequence; A positioning unit is used to use the amplitude envelope sequence and the energy envelope sequence to splice into a feature sequence, use the PCA algorithm to obtain the anomaly score of each point in the feature sequence to obtain an anomaly score sequence, and then determine the anomaly score threshold from the anomaly score sequence, and finally determine the mutation signal and its index position existing in the vibration time domain signal by comparing the size relationship between the anomaly score sequence and the anomaly score threshold; wherein the triangular threshold method is used to obtain the anomaly score threshold corresponding to the anomaly score sequence.

5. A mutation signal positioning device based on PCA algorithm, 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 mutation signal localization method based on the PCA algorithm as described in any one of claims 1 to 3.

Images