Search

CN-121977843-A - Bearing wear on-line monitoring device and early warning method based on acoustic envelope analysis

CN121977843ACN 121977843 ACN121977843 ACN 121977843ACN-121977843-A

Abstract

The invention provides an on-line bearing wear monitoring device and an early warning method based on acoustic envelope analysis, and belongs to the field of equipment state monitoring. The device comprises a non-contact acoustic sensing array, a signal preprocessing module, an edge calculation analysis unit, a communication module and a man-machine interface. The method comprises the steps of collecting high-frequency acoustic signals of the bearing, extracting envelope through self-adaptive band-pass filtering and Hilbert transformation, analyzing envelope spectrum, extracting multidimensional features, intelligently diagnosing abrasion state by using a machine learning model, and carrying out dynamic hierarchical early warning by combining trend and working condition. The invention realizes non-contact, high-sensitivity and intelligent online monitoring of early wear of the bearing and provides a high-efficiency and reliable solution for predictive maintenance of industrial equipment.

Inventors

  • LI DONG
  • LIU CHANGYANG
  • LIU KAI
  • LIU JIANFEI
  • SHANG WENBIN
  • XU ZIQI
  • ZHAO WEI
  • XIAO YONGXIANG

Assignees

  • 临沂恒源智能科技股份有限公司

Dates

Publication Date
20260505
Application Date
20260121

Claims (9)

  1. 1. Bearing wear on-line monitoring device based on acoustic envelope analysis, characterized by comprising: An acoustic sensor array module (101) for acquiring high-frequency acoustic signals generated by the operation of the bearing (107) in a non-contact or quasi-contact manner; The input end of the signal preprocessing and collecting module (102) is electrically connected with the output end of the acoustic sensing array module (101) and comprises a pre-amplifying circuit, a filtering circuit and an analog-to-digital converter, and the signal preprocessing and collecting module is used for amplifying, filtering and digitally collecting the acoustic signals; An edge calculation analysis unit (103) connected with the signal preprocessing and acquisition module (102) and used for running: The self-adaptive envelope analysis module is used for carrying out band-pass filtering, hilbert transformation and envelope extraction and envelope spectrum analysis on the digitized acoustic signals; the feature extraction and fusion module is used for extracting multidimensional feature vectors from the envelope spectrum and the time domain envelope signal; the intelligent diagnosis and early warning module is internally provided with a diagnosis model and is used for judging the abrasion state of the bearing (107) according to the characteristic vector and generating grading early warning information based on the state and the trend; The data communication module (104) is connected with the data interface of the edge calculation analysis unit (103) and is used for uploading data to the remote cloud platform (106); And the man-machine interaction and alarm module (105) is connected with the display and alarm interface of the edge calculation and analysis unit (103).
  2. 2. The on-line bearing wear monitoring device based on acoustic envelope analysis according to claim 1, characterized in that the acoustic sensor array module (101) comprises an ultrasonic microphone or an acoustic emission sensor, the operating band of which covers 20kHz to 200kHz.
  3. 3. The on-line bearing wear monitoring device based on acoustic envelope analysis according to claim 1, characterized in that the band pass filtering performed by the adaptive envelope analysis module has a band pass range dynamically set according to the theoretical failure frequency of the monitored bearing (107).
  4. 4. The device for on-line monitoring of bearing wear based on acoustic envelope analysis according to claim 1, wherein the feature vector extracted by the feature extraction and fusion module comprises a plurality of statistical features including the amplitude of the fault frequency of the bearing (107) and its harmonics in the envelope spectrum, and the time domain kurtosis, the root mean square value, the peak factor of the envelope signal.
  5. 5. The on-line bearing wear monitoring device based on the acoustic envelope analysis according to claim 1, wherein the diagnosis model in the intelligent diagnosis and early warning module is a classification model trained based on a machine learning algorithm and is used for outputting a state grade of bearing (107) wear, and the early warning module dynamically adjusts an early warning threshold value and generates a hierarchical early warning according to the state grade, a characteristic historical trend and a real-time working condition of equipment.
  6. 6. An on-line monitoring and early warning method for bearing wear based on the device of any one of claims 1 to 5, comprising the following steps: Step S1, deploying a device and configuring monitoring parameters; s2, acquiring and preprocessing acoustic signals of the bearing (107) in real time; s3, carrying out self-adaptive envelope analysis on the preprocessed signals to obtain envelope signals and frequency spectrums thereof; s4, extracting multidimensional feature vectors from the envelope spectrum and the time domain envelope signal; S5, inputting the feature vector into a diagnosis model to obtain a diagnosis result of the abrasion state of the bearing (107); Step S6, executing early warning logic by combining the diagnosis result, the characteristic historical trend and the working condition information, and generating and distributing grading early warning information; And S7, uploading data and continuously monitoring.
  7. 7. The method for on-line monitoring and early warning of bearing wear according to claim 6, wherein the self-adaptive envelope analysis in the step S3 specifically comprises setting a band-pass filter according to parameters of a bearing (107), performing Hilbert transform on a filtered signal to obtain an analysis signal, calculating the amplitude of the analysis signal as an envelope signal, and performing fast Fourier transform on the envelope signal to obtain an envelope spectrum.
  8. 8. The method for on-line monitoring and early warning of bearing wear according to claim 6, wherein the step S6 of hierarchical early warning includes at least three levels, namely, observation level early warning, early warning level alarm and emergency level alarm, which correspond to different bearing (107) wear severity and response requirements, respectively.
  9. 9. The method for on-line monitoring and early warning of bearing wear according to claim 6, further comprising the steps of converging a plurality of pieces of equipment data on a remote cloud platform (106), performing transverse comparison analysis, health status visualization and maintenance work order automatic generation.

Description

Bearing wear on-line monitoring device and early warning method based on acoustic envelope analysis Technical Field The invention relates to the technical field of predictive maintenance and state monitoring of industrial equipment, in particular to an on-line monitoring device and an intelligent early warning method for the abrasion state of a rotary mechanical bearing based on acoustic signal envelope analysis. Background Rolling bearings are the core components of rotary machines, the operating state of which is directly related to the reliability, safety and production efficiency of the whole plant. It is counted that about 30% -40% of the rotating machinery failures result from bearing failures. Therefore, the method and the device can monitor and early warn the abrasion state of the bearing accurately in real time, and have great significance for implementing predictive maintenance, avoiding unplanned shutdown and reducing maintenance cost. Existing bearing condition monitoring techniques rely primarily on vibration analysis. And an acceleration sensor is arranged on the bearing seat, vibration signals are collected and subjected to frequency spectrum analysis, so that characteristic frequencies (such as inner ring, outer ring and rolling body fault frequencies) of the bearing caused by local damage (such as pitting and peeling) are identified. However, vibration monitoring techniques have several inherent limitations in practical applications, including firstly that early, slight abrasion or weak impact signals tend to be submerged in strong background vibration noise, which is difficult to extract effectively, resulting in early warning delay, secondly that installation of vibration sensors usually requires machine shutdown modification (such as drilling and tapping) of equipment, and that installation positions and tightening moments have a great influence on signal quality, and furthermore, for low-speed heavy-duty or structurally complex equipment, failure impact energy is low and signal-to-noise ratio of vibration signals tends to be unsatisfactory. Acoustic monitoring (or acoustic emission monitoring) has begun to be of interest as a non-contact or quasi-contact method. When the bearing wears or microcracks, high-frequency stress waves are released, and the frequency range of the high-frequency stress waves is usually tens of kHz to hundreds of kHz, which is far higher than the vibration frequency of conventional equipment. This provides the possibility to separate the fault signal from the noise background. However, the original acoustic signal also contains a lot of environmental noise, and the direct analysis of the high frequency signal is computationally burdened. Envelope analysis techniques are effective tools for processing such amplitude modulated signals by demodulating and extracting the envelope of the impact event (i.e., the low frequency modulated signal) and then performing spectral analysis on the envelope, thereby greatly simplifying the analysis process and highlighting fault characteristics. In the prior art, a systematic solution for deeply fusing non-contact acoustic sensing, self-adaptive envelope analysis and processing with an intelligent early warning model is lacking. Most methods either only pay attention to single signal characteristics or require manual threshold setting, are difficult to adapt to complex and changeable industrial field working conditions, and cannot realize closed loops from data to executable early warning decisions. Disclosure of Invention The invention aims to provide an on-line bearing wear monitoring device and an on-line bearing wear early warning method based on acoustic envelope analysis, so as to solve the problems that the prior vibration monitoring technology is insensitive to early bearing wear detection and inconvenient to install, the prior acoustic monitoring method is lack of deep combination with an intelligent early warning model, and accurate and self-adaptive early warning is difficult to realize. The device comprises: The acoustic sensor array module comprises at least one high sensitivity, broadband (e.g. 20kHz-200 kHz) acoustic sensor, preferably an ultrasonic microphone or acoustic emission sensor. The module is fixed near the bearing seat in a non-contact mode (1-30 cm away from the bearing seat) or in a quasi-contact mode (through a magnetic attraction seat or a waveguide rod) and is used for collecting high-frequency acoustic signals generated by the operation of the bearing. The input end of the signal preprocessing and collecting module is electrically connected with the output end of the acoustic sensing array module, and the signal preprocessing and collecting module comprises a pre-amplifying circuit, a filtering circuit and an analog-to-digital converter and is used for amplifying, filtering and digitally collecting the acoustic signals and performing analog-to-digital conversion at a sampling rate not lower than 20