JPH02203221A - Abnormality diagnostic device for gear - Google Patents

Abstract

PURPOSE:To grasp quantitatively the eccentric abnormality of a gear by extracting the frequency components generated with the eccentric abnormality from the vibration waveform of a gear and calculating the abnormality degree number therefrom. CONSTITUTION:The abnormality for the gears 102-106 are diagnosed from the vibration waveforms which are detected by a vibration sensor 108 at the outside of a gear device 101. At this time, the waveforms are made to pass through a vibration frequency band caused by the gear by a band-pass filter 1, and the waveforms after passing through the filter 1 are repeatedly fetched in a waveform input part 2 in the manner of taking a signal worked 21 based on the pulses generating 22 from the gear revolution as a start command. Then, the waveforms fetched to the set number of times are averaged 3 and converted 4 into a frequency spectrum. The meshing and revolution frequency are calcu lated based on this frequency and the number of gear tooth, hence the noticed frequency is calculated 5. Based on the results, the size of necessary frequency components are extracted 6 from the frequency spectrum, then the abnormality degree and quantity are calculated 7 and displaied 8.

Description

[Detailed description of the invention] [Industrial application field]

The present invention relates to a gear abnormality diagnosing device that processes vibration waveforms captured for gear abnormality diagnosis to diagnose abnormalities due to eccentricity.

[Conventional technology]

In gear devices that house multiple gears that mesh and rotate in a case and are used for power transmission, rotation speed conversion, etc., abnormal eccentricity of the gears, including bending of the shaft, and dents on the tooth surfaces. Due to local abnormalities such as broken teeth, abnormal wear, etc.
Abnormal vibrations occur when these gears mesh with their mating gears. A gear abnormality diagnosis device detects this with a vibration sensor fixed to the case, analyzes the vibration waveform, and conversely identifies abnormalities in the gear. However, in the vibration waveform detected by fixing the vibration sensor to the case described above, vibrations other than gear abnormalities caused by rolling bearings that support the gear shaft and electromagnetic force of the electric motor are also overlapped, and the ratio of these vibrations is large. This obstructs gear abnormality diagnosis. For this reason, a rotation pulse is generated by a rotation mark provided on the gear shaft and a sensor that detects this, and vibration waveforms are repeatedly captured using this pulse as a reference, and the vibrations are averaged to reduce vibrations other than gear abnormalities. ing. Additive averaging refers to data whose timing (including phase, etc.) changes unstably with respect to the reference by repeating time domain data, i.e., vibration waveforms, taken with a certain point as a reference, and adding and averaging them many times. This method takes advantage of the fact that noise) approaches zero and disappears, leaving only data that is always at stable timing with respect to the reference point. However, gear abnormalities do not only occur in the gear attached to the shaft from which the rotational pulse is extracted, but also in the mating gear that meshes with the gear. In such a case, you can reattach the rotation mark to the shaft on which the gear you want to diagnose the abnormality is attached to and extract the rotation pulse of that shaft. However, due to the structure of the gear device, attaching the rotation mark is not possible. It is often possible. The applicant had previously applied for an abnormality diagnosis device for gears as Special No. 62-231734, but this device detects not only the rotation of the gear from which rotational pulses are extracted, but also the rotation of the mating gear meshing with this one-wheel. This is a gear abnormality diagnosing device that not only can obtain synchronized vibration waveforms, but also can capture the noticeable vibration waveforms generated by the meshing of both gears by arbitrarily adjusting them. Diagnosis was performed by converting the waveform after noise was removed by averaging the waveform into a frequency spectrum. That is, since the vibration components generated differ depending on the type of gear abnormality, it has been determined whether or not there is an abnormality by paying attention to changes in the spectrum.

[Problem to be solved by the invention]

In an actual gear device, a spectrum due to the gear attached to the shaft that meshes with the gear attached to the shaft to be diagnosed also appears, and the ratio also changes depending on the reduction ratio 9 and the increase ratio. Therefore, it is not one-to-one, and depending on the type of abnormality, many spectra such as the basic spectrum and integral multiples of the basic spectrum appear, so it is difficult to immediately tell which spectrum corresponds to the component caused by the abnormality of which gear. However, there was a problem in that abnormality could not be easily determined. In addition, in the early stages of an abnormality, the vibration component caused by the abnormality is small, and simply analyzing the averaged waveform will not work.
There was also the problem that the spectrum caused by the abnormality was buried in noise that could not be completely removed. FIG. 5 is a vibration waveform diagram of an eccentric gear determined from an experiment to detect eccentricity of a gear, and FIG. 6 is a waveform diagram of an external pulse signal used for abnormality diagnosis of the gear shown in FIG. In the vibration waveform, the horizontal axis is time and the vertical axis is vibration amplitude, and when compared with the waveform of the external pulse signal in FIG. 6, it can be seen that the amplitude increases and decreases at a rate of once per rotation. Furthermore, when the frequency of the increasing and decreasing vibrations was determined from this waveform, it was found that this frequency was the meshing wave number of the eccentric gear. The frequency spectrum in such a case is the rotational frequency of the gear f
It has been experimentally confirmed that, when f is the meshing frequency of gears, three frequency spectra of f.fr, fm, and fr-+r are generated. calculate numbers,
It is an object of the present invention to provide a gear abnormality diagnosing device for diagnosing abnormalities due to gear eccentricity.

[Means to solve the problem]

According to the present invention, the above object is to provide a band pass filter that passes a vibration frequency band caused by a gear, and a brush filter that passes through the band pass filter using an external pulse signal processed based on the rotation pulse of the gear as a starting command. a waveform input section that repeatedly takes in the waveform; a waveform averaging section that adds and averages the waveforms that have been taken in by the brush input section until the number of times set in the attached averaging number setting section is reached; and a waveform averaging section that converts the added and averaged waveform into a frequency spectrum. a Fourier transform unit; a frequency-of-interest calculation unit that calculates a meshing frequency and a rotational frequency based on the frequency of the rotational pulse and the number of teeth of the gear; and a frequency-of-interest calculation unit that calculates a frequency of interest based on the calculated values; a spectrum extraction section that extracts the magnitude of the frequency component necessary for diagnosis from the spectrum transformed by the Fourier transform section based on the frequency of interest calculated by the frequency calculation section; This is achieved by a gear abnormality diagnosing device that includes an abnormality degree calculation section that calculates the abnormality degree Y of the gear from its size, and a display section that displays the abnormality degree Y.

[For use]

f, −f, obtained through a bandpass filter. By extracting the frequency components of f, , f, +f, and averaging the extracted components in the waveform averaging section, the amplitude of the waveform becomes almost constant, and this averaged waveform is converted into a frequency spectrum in the Fourier transform section. , the frequency fEfF to be noted in the frequency-of-interest calculation section. Find f,, f, +f, and use the spectrum extractor to calculate the magnitudes of frequency components P(f, −f, ), P(
By extracting Lm), PCf@10f, ) and determining the abnormality degree number Y from these values, it is possible to diagnose gear eccentricity abnormality from the characteristics of the frequency spectrum caused by the eccentricity abnormality.

【Example】

The present invention will be described in detail below based on the drawings. FIG. 1 is a block diagram showing the configuration of a gear abnormality diagnosing device according to an embodiment of the present invention. In FIG. 1, the gear abnormality diagnosis device includes a bandpass filter 1 that passes a vibration frequency band caused by the gear;
a waveform input section 2 that repeatedly receives the waveform after passing through the bandpass filter 1 using an external pulse signal processed by the external pulse signal generation section 21 based on the rotation pulse generated from the rotation pulse generation section 22 of the gear as a starting command; a waveform averaging section 3 that adds and averages the waveforms input by the waveform input section 2 until the number of times set by the averaging number setting section 31 is reached; a Fourier transform section 4 that converts the averaged waveform into a frequency spectrum; Calculate the meshing frequency and rotation frequency based on the frequency of the rotation pulse and the number of gear teeth,
A frequency of interest calculation unit 5 calculates frequencies to be focused on based on the calculated values, and a spectrum converted by the Fourier transformation unit 4 based on the frequency of interest calculated by the frequency of interest calculation unit 5 is used to calculate necessary frequencies for diagnosis. a spectrum extracting unit 6 that extracts the magnitude of a frequency component; an abnormality degree calculation unit 7 that calculates a gear abnormality degree number Y from the magnitude of each frequency component extracted by the spectrum extraction unit 6; A display section 8 for displaying Y is provided. The gear device 101 includes a gear AlO2 and a gear B103 meshing with this gear AlO2. Gear Cl coaxial with this gear B103
It consists of a gear D105 meshing with a gear ClO4 and a gear ClO4. Next, the operation of this gear abnormality diagnosis device will be explained. b) First, in order to make the vibration component that appears due to eccentricity more clear, that is, in order to diagnose eccentricity abnormality, to remove the vibration component caused by local abnormality or wear that is considered as noise, outside the case of the gear device 101. A vibration waveform 109 taken out from a vibration sensor 108 fixed to is input to the bandpass filter 1. The passing frequency bandwidth of bandpass filter 1 is fl-nx
By setting r, to f, +nXf (n is about 1 to 3), other components can be removed as much as possible. Here, flI is the meshing frequency, meshing frequency = number of gear teeth x gear rotation speed + 6°f, is the gear rotation frequency, and gear rotation frequency = gear rotation speed + 60. Mouth) Next, in order to extract the eccentricity abnormality waveform of only the gear of interest, processing is performed based on the rotational pulse generated from the rotational pulse generation section 22, and the external pulse signal generated from the external pulse signal generation section 21 is used as a starting command to generate the above-mentioned band. The waveform after passing through the pass filter 1 is repeatedly input to the waveform input section 2. Here, the generation frequency of the external pulse signal is set to be the same as the rotational frequency of the gear of interest. The meshing frequencies of the meshing gears are exactly the same, and the bandpass filter 1 cannot determine which gear has eccentricity. FIG. 3 is a waveform diagram of an external pulse signal generated from the external pulse signal generator 21. c) The waveform input by the waveform input section 2 is averaged by the waveform averaging section 3 until the number of times set by the attached averaging number setting section 31 is reached. FIG. 2 shows vibration waveforms averaged by the averaging section 3. 2) The averaged waveform is converted into a frequency spectrum by the Fourier transform unit 4. FIG. 4 shows the spectrum of the vibration waveform obtained by the Fourier transform unit 4. e) Even if you perform the above waveform processing, there is a possibility that a spectrum of noise components that cannot be removed may still appear.
The frequency of interest calculation section 5 determines in advance the frequency to be read. The frequency of interest calculation unit 5 analyzes the frequency of the rotational pulse 22, calculates the meshing frequency fla+rotational frequency f1, etc. of the gear of interest based on the combination of the accurate frequency and the number of teeth of the gear input separately, and based on this, calculates the meshing frequency fla+rotation frequency f1 of the gear of interest. Should frequency f
, −f, or f* + fr. f) Output of the Fourier transform unit 4 and frequency of interest calculation unit 5
and the output thereof are input to the spectrum extracting section 6. In this spectrum extraction section 6, the magnitudes P(f, ) and P(f
, -f, ), P(f, +f, ) are extracted. g) Input the output of the spectrum extraction section 6 to the abnormality degree calculation section 7. This abnormality degree calculation unit 7 calculates the magnitude P(
f, −f, ), P(f, 10f, ), P(f,
) to calculate the abnormality degree Y. Here, Y=P(f, -f, ) xP(f, + f,)/[P(f,)/2] "The spectrum due to eccentricity abnormality always contains f, -, f. and f, + f, A frequency component appears.In an actual gear device, many gears are meshed, so it is possible that a spectrum may be created by chance at f, -f, or f, + f.However, the frequency component appears in both. Since the possibility is considered to be small, one of the values will be close to zero, so by calculating the multiplier, the value of Y will be close to zero and can be ignored.H) The above abnormality degree number The output of the calculation section 7 is input to the display section 8, and the abnormality degree number Y is displayed on the display section 8.In the example of the spectrum shown in FIG. The eccentricity abnormality of the gear device 101 is diagnosed from the abnormality degree number Y.

【Effect of the invention】

According to this invention, f is a frequency component generated from the gear vibration waveform based on eccentricity abnormality. −f r +f,,
By extracting the frequency components f, +f, and calculating the degree of abnormality from these values, it is possible to quantitatively understand the eccentricity abnormality of the gear.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a gear abnormality diagnosis device according to an embodiment of the present invention, FIG. 2 is a diagram showing vibration waveforms averaged by the waveform averaging section of FIG. Figure 4 is a waveform diagram of the external pulse signal of the gear abnormality diagnosis device, Figure 4 is a diagram showing the spectrum of the vibration waveform obtained by converting it in the Fourier transform section of Figure 1, and Figure 5 is a diagram showing the spectrum of the vibration waveform obtained by the gear eccentricity abnormality. The vibration waveform diagram in FIG. 6 is a diagram showing an external pulse signal used for abnormality diagnosis of the gear shown in FIG. 5. 1: Bandpass filter, 2: Waveform input section, 3: Waveform averaging section, 4: Fourier transform section, 5: Frequency of interest calculation section, 6
: Spectrum extraction section, 7; Abnormality degree calculation section, 8; Display section, 21: External pulse signal generation section, 22: Rotation pulse generation section, 31: Average number of times setting section, 101: Gear device, 10
6 double rotation mark, 107; sensor, 108: vibration sensor, 109: vibration waveform.

Claims (1)

[Claims] 1) A gear abnormality diagnosis device that diagnoses abnormalities in each gear from vibration waveforms detected by a vibration sensor attached to the outside of the case of a gear device that houses multiple gears that mesh and rotate. A bandpass filter that passes the vibration frequency band caused by the vibration, a waveform input section that repeatedly captures the waveform after passing through the bandpass filter using an external pulse signal processed based on the rotational pulse of the gear as a start command, and an attached averaging frequency setting. a waveform averaging section that adds and averages the waveforms captured by the waveform input section until the number of times set in the waveform input section is reached; a Fourier transform section that converts the averaged waveform into a frequency spectrum; and a frequency spectrum of the rotation pulse and gear teeth. a frequency-of-interest calculation unit that calculates the meshing frequency and rotational frequency based on the calculated values, and obtains a frequency of interest based on the calculated values; a spectrum extraction unit that extracts the magnitude of frequency components necessary for diagnosis from the spectrum converted by the conversion unit; and an abnormality degree number Y for determining the degree of abnormality of the gear from the magnitude of each frequency component extracted by the spectrum extraction unit. A gear abnormality diagnosing device comprising a calculation section and a display section for displaying the abnormality degree number Y.