CN-116522226-B - Quantitative diagnosis method and system for mechanical faults

CN116522226BCN 116522226 BCN116522226 BCN 116522226BCN-116522226-B

Abstract

The invention provides a quantitative diagnosis method and a quantitative diagnosis system for mechanical faults, wherein the method comprises the steps of carrying out spectrum analysis on original vibration data of each component to obtain spectrum data, comparing preset fault characteristic frequency with the spectrum data, determining the mechanical fault type of each component according to a comparison result, extracting a first fault spectrum corresponding to each component according to a preset rule from the spectrum data according to the determined mechanical fault type of each component, carrying out reconstruction processing on the first fault spectrum to obtain a second fault spectrum, carrying out IFFT calculation on the second fault spectrum data to extract time domain data, further simplifying the time domain data, calculating corresponding time domain characteristic values, and carrying out quantitative diagnosis on the mechanical faults by taking the time domain characteristic values as quantitative indexes. The system has the same beneficial effects.

Inventors

GONG MIAO
TANG DEYAO
LI XIUWEN
ZHOU WEI
LI YANG
JIN YITAO

Assignees

北京唐智科技发展有限公司
唐智科技湖南发展有限公司

Dates

Publication Date: 20260508
Application Date: 20230428

Claims (12)

1. The quantitative diagnosis method for the mechanical faults is characterized by comprising the following steps of: performing spectrum analysis on the original vibration data of each component to obtain spectrum data; Determining the mechanical fault type of each component according to the frequency spectrum data and the preset mechanical fault characteristic frequency, wherein the preset mechanical fault characteristic frequency is a preset fault characteristic frequency corresponding to the mechanical fault type of each component; Extracting a first fault frequency spectrum corresponding to each component from the frequency spectrum data according to a preset rule according to the mechanical fault type of each component; reconstructing the first fault spectrum to obtain a second fault spectrum; Performing IFFT calculation on the second fault frequency spectrum to extract time domain data; after the time domain data is processed, calculating a time domain characteristic value; and quantitatively diagnosing the mechanical faults according to the time domain eigenvalue.
2. The quantitative diagnosis method for mechanical failure according to claim 1, wherein the determining of the failure type of each component based on the spectral data and a preset mechanical failure characteristic frequency comprises the steps of: presetting a mechanical failure characteristic frequency of an ith component corresponding to a mechanical failure type, wherein i is a component serial number and is used for traversing each component; judging whether an actual characteristic frequency which is the same as the preset mechanical failure characteristic frequency of the ith component exists in the frequency spectrum data or not; if so, judging that the ith component has the mechanical fault type corresponding to the mechanical fault characteristic frequency of the ith component; repeating the steps until the mechanical failure type of each component is determined.
3. The quantitative diagnosis method for mechanical failure according to claim 2, wherein the step of extracting the first failure spectrum corresponding to each component from the spectrum data according to a predetermined rule according to the failure type of each component comprises the steps of: Extracting a main spectrum of the ith component from the spectrum data according to the type of mechanical failure of the ith component; Repeating the steps until main spectrums corresponding to all the components are obtained, and taking the main spectrums corresponding to all the components as the first fault spectrums; The main frequency spectrum of the ith component is specifically the actual characteristic frequency which is the same as the preset mechanical fault characteristic frequency of the ith component and all higher orders of the actual characteristic frequency.
4. The quantitative diagnosis method for mechanical failure according to claim 3, wherein the reconstructing the first failure spectrum to obtain a second failure spectrum comprises the steps of: Constructing a remainder function according to an initial sampling frequency and the actual characteristic frequency in the first fault spectrum; judging whether the result of the residual function is equal to 0; If not, reconstructing a first sampling frequency in the whole period according to a first preset rule; Resampling the characteristic sample according to the first sampling frequency and an interpolation method to obtain the second fault frequency spectrum; The initial sampling frequency is specifically a sampling frequency set for acquiring the original vibration data of each component.
5. The quantitative diagnosis method of mechanical failure according to claim 4, wherein the first preset rule is specifically: ; Wherein, the For the first sampling frequency to be the same, As a practical characteristic frequency of the signal, Is the highest order of the order sequence of the actual characteristic frequency.
6. The quantitative diagnosis method for mechanical failure according to claim 2, wherein the step of extracting the first failure spectrum corresponding to each component from the spectrum data according to a predetermined rule according to the failure type of each component, further comprises the steps of: Extracting a main spectrum and a modulation spectrum of the ith component from the spectrum data according to the type of mechanical failure of the ith component; repeating the steps until main frequency spectrums corresponding to all the components are obtained, and taking the main frequency spectrums corresponding to all the components and the modulation frequency spectrums as the first fault frequency spectrums; The main frequency spectrum of the ith component is specifically the actual characteristic frequency which is the same as the preset mechanical fault characteristic frequency of the ith component and all higher orders of the actual characteristic frequency; The modulation spectrum specifically comprises a preset modulation frequency and all higher orders.
7. The quantitative diagnosis method for mechanical failure according to claim 6, wherein the reconstructing the first failure spectrum to obtain a second failure spectrum includes the steps of: Acquiring the modulation frequency and the least common multiple frequency of the actual characteristic frequency in the first fault frequency spectrum; Reconstructing a second sampling frequency in a whole period according to the least common multiple frequency, the initial sampling frequency and a second preset rule; resampling the characteristic sample according to the second sampling frequency and the interpolation method to obtain the second fault frequency spectrum; The initial sampling frequency is specifically a sampling frequency set for acquiring the original vibration data of each component.
8. The quantitative diagnosis method for mechanical failure according to claim 7, wherein the second preset rule is specifically: ; Wherein, the For the whole number of cycles, In order to round down the function, Is the initial sampling frequency; Is the least common multiple frequency; ; Wherein, the Is the second sampling frequency.
9. The quantitative diagnosis method for mechanical failure according to claim 1, wherein the calculating of the time domain eigenvalue after the processing of the time domain data comprises the steps of: Simplifying the time domain data according to the IFFT principle; and calculating and acquiring the time domain characteristic value according to the reduced time domain data.
10. The quantitative diagnosis method for mechanical failure according to claim 1, wherein the quantitative diagnosis for mechanical failure is performed based on the time domain feature value, further comprising the steps of: normalizing the time domain characteristic value; acquiring fault magnitude of each component according to the processed time domain characteristic value and the quantitative preset parameter of each component; and quantitatively diagnosing the mechanical faults of all the components according to the fault magnitude and a preset fault threshold.
11. The quantitative diagnosis method for mechanical failure according to claim 1, wherein the spectrum analysis is performed on the raw vibration data of each component to obtain spectrum data, further comprising the steps of: sequentially carrying out wavelet analysis, envelope demodulation/resonance demodulation on the original vibration data of each component to obtain an original vibration characteristic sample; and carrying out spectrum analysis on the original vibration characteristic sample to obtain the spectrum data.
12. A quantitative diagnosis system for mechanical failure, comprising: the frequency spectrum analysis module is used for carrying out frequency spectrum analysis on the original vibration data of each component to obtain frequency spectrum data; The fault type module is used for determining the mechanical fault type of each component according to the frequency spectrum data and the preset mechanical fault characteristic frequency, wherein the preset mechanical fault characteristic frequency is a preset fault characteristic frequency corresponding to the mechanical fault type of each component; the first fault frequency spectrum module is used for extracting a first fault frequency spectrum corresponding to each component from the frequency spectrum data according to a preset rule according to the mechanical fault type of each component; The reconstruction module is used for carrying out reconstruction processing on the first fault frequency spectrum to obtain a second fault frequency spectrum; The extraction module is used for performing IFFT calculation on the second fault frequency spectrum so as to extract time domain data; The processing module is used for calculating a time domain characteristic value after processing the time domain data; And the fault diagnosis module is used for quantitatively diagnosing the mechanical faults according to the time domain characteristic values.

Description

Quantitative diagnosis method and system for mechanical faults Technical Field The invention relates to the technical field of mechanical fault diagnosis, in particular to a quantitative mechanical fault diagnosis method and system. Background The machine is used as an important supporting component, the running state of the machine is directly the safety and reliability of industrial equipment, the machine belongs to a fault multiple component, the health state of the machine is very necessary to be monitored and quantitatively assessed by faults, and the fault location and the degree of the fault are quantitatively achieved through a certain monitoring means, so that the accurate maintenance and the service life prediction of the machine faults are better developed, and the important solution is needed at present. The existing mechanical fault diagnosis is mainly qualitative, and quantitative methods are not many, and are basically realized based on vibration signals. In the aspect of time domain analysis, vibration time domain characteristic values such as root mean square, kurtosis, pulse indexes, peak values, deviation indexes and the like are calculated, and the most suitable parameters are selected through entropy weight to serve as quantitative identification indexes, but besides fault signals, the vibration signals also contain strong noise and interference components, and the calculation results are directly influenced. In the aspect of frequency domain analysis, spectrum analysis is carried out on the vibration signals, the complexity of periodic components in the frequency spectrum is judged based on Lempe l _Ziv algorithm, the more the periodic signals are, the more faults are, and finally, no quantitative index exists. And the original vibration signal is subjected to envelope demodulation analysis to extract impact characteristics, the amplitude of the impact frequency is selected as a quantitative standard, but the mechanical failure frequency often presents multiple steps, and the final failure accuracy of the method is lower due to inaccuracy of only 1-order amplitude representation. In addition, the running state of the equipment is monitored by adopting a neural network, fuzzy recognition and other machine learning methods, and the method needs a large amount of fault sample accumulation, is seriously dependent on early data analysis experience, and is not suitable for on-line monitoring. Therefore, providing a quantitative diagnosis method and system for mechanical failure that can effectively solve the above technical problems is a problem that needs to be solved by those skilled in the art. Disclosure of Invention The invention aims to provide a quantitative diagnosis method and a quantitative diagnosis system for mechanical faults, wherein the method has clear logic, safety, effectiveness, reliability and simple and convenient operation, and can effectively improve the accuracy of the fault diagnosis of the rolling bearing on the premise of not accumulating a large number of fault samples. Based on the above purpose, the technical scheme provided by the invention is as follows: A quantitative diagnosis method for mechanical faults, comprising the following steps: performing spectrum analysis on the original vibration data of each component to obtain spectrum data; Determining the mechanical fault type of each component according to the frequency spectrum data and the preset mechanical fault characteristic frequency; Extracting a first fault frequency spectrum corresponding to each component from the frequency spectrum data according to a preset rule according to the mechanical fault type of each component; reconstructing the first fault spectrum to obtain a second fault spectrum; performing an I FFT calculation on the second failure spectrum to extract time domain data; after the time domain data is processed, calculating a time domain characteristic value; and quantitatively diagnosing the mechanical faults according to the time domain eigenvalue. Preferably, the determining the fault type of each component according to the spectrum data and the preset mechanical fault characteristic frequency comprises the following steps: Presetting a mechanical failure characteristic frequency of an ith component; judging whether an actual characteristic frequency which is the same as the preset mechanical failure characteristic frequency of the ith component exists in the frequency spectrum data or not; if yes, judging that the mechanical failure type of the ith component exists; repeating the steps until the mechanical failure type of each component is determined. Preferably, the extracting, according to the fault type of each component, a first fault spectrum corresponding to each component in the spectrum data according to a preset rule includes the following steps: Extracting a main spectrum of the ith component from the spectrum data according to the type of mechanical failure of the ith co