CN102788695B - Identification method of rolling bearing abrasion - Google Patents

Abstract

The invention discloses an identification method of rolling bearing abrasion, which comprises the following steps: (1) characteristic frequencies of an intershaft bearing at different working stages are computed out; (2) short time Fourier transform is carried out on a bearing oscillation test signal to obtain a Fourier transform spectrogram; (3) a frequency range is set up, and the frequency range energy is obtained by an integration manner to be a calculated instant energy of the bearing characteristic frequency; (4) the actual instant energy and the instant energy under normal working situation in a time are respectively computed out; (5) the instant energy capacity under the normal working situation is taken as a baseline to set up a threshold value, and the abrasion trend and the bearing damage degree are compared and judged. The identification method can judge the abrasion of the rolling bearing according to the capacity of the instant energy, and has the characteristics of online real time and capability in measuring abrasion degree.

Description

A kind of recognition methods of rolling bearing wearing and tearing

Technical field

The invention belongs to the research field of mechanical Fault Monitoring of HV and diagnostic method, particularly a kind of recognition methods of carrying out bearing premature wear state based on instantaneous energy method.

Background technology

The outer ring of rolling bearing is placed in bearing saddle bore, and inner ring and transmission shaft are connected, and with axle, rotate.If itself has fault rolling bearing, when axle turns round with certain rotating speed and under certain load, the vibrational system of bearing and bearing seat or shell composition is produced to excitation, produce vibration.This vibration comprises the comprehensive excitation system of the internal motivation formation that external drive and bearing inner structure, installation, rigging error cause.In order effectively to isolate bearing operation troubles from vibration signal, need to first overcome the extraction of large noise background lower bearing Weak fault characteristic information, secondly different rotating speeds operating mode lower bearing characteristic frequency changes and respective bandwidth, and the method for failure message is effectively extracted in finally research.

The surface abrasion of high-speed overload rolling bearing aggravates gradually with the length of working time, is a kind of gradual failure, and the vibration producing after bearing surface wearing and tearing has individual features frequency, and the after vibration level of simultaneously wearing and tearing (amplitude) is apparently higher than normal bearing.Therefore the aggravation of vibration and transfinite also can principium identification bearing operation conditions.Bearing fault characteristics frequency is generally below low frequency (1kHz), but the high frequency proper vibration composition of the bearing arrangement that can bring out due to surface damage percussive action, and for example bearing outer ring radially bends proper vibration.

Instantaneous characteristic frequency, harmonic wave and ground unrest based on characteristic frequency and its frequency analysis formed the total characteristic of rolling body wearing and tearing, and suggestion adopts generation, development and the trend of instantaneous characteristic area harmonic spectrum ENERGY METHOD monitoring wearing and tearing for this reason.

The ground unrest amplitude size of wearing and tearing is relevant to rotating speed size, in addition also generation and the evolution of representative wearing and tearing.The fricative vibration noise of object of ground unrest source phase mutual friction and cause that because abrasion post gap increases whole system noise level increases, therefore we are referred to as subband to bring large-scale harmonic components, can differentiate the process of wearing and tearing by the wear and tear size of noise level within the scope of the subband at characteristic of correspondence frequency place of calculating.

The bearing of high-speed overload is as the intershaft bearing of aeromotor, and intershaft bearing is the parts of playing a supporting role between high pressure rotor and low pressure rotor, if intershaft bearing breaks down, the lighter's system vibration increases, severe one rotor locking engine flame-out in flight.The principle of work of intershaft bearing is to rely on the drawing force producing between active lasso and roller to drive rolling body and the running of retainer parts to realize to rotate.

Bearing, in wear process, there will be bearing fault frequency and harmonic wave thereof in succession on spectrogram.Along with wearing and tearing aggravation, in bearing fault frequency and harmonic wave place thereof, there will be side frequency, when bearing wear occurs and develops, distinctive frequency can be replaced by noise.Secondary lobe will become the uniquely tagged (in low speed situation) of these faults.The frequency band that side frequency and failure-frequency form forms mound, peak shape, is commonly called as " rick "." haystack effect " refers to that haystack can worsen and increase gradually along with fault, therefore can weigh fault degree by the energy of " haystack ".The present invention produces based on this kind of thinking.

Summary of the invention

Object of the present invention, is to provide a kind of recognition methods of rolling bearing wearing and tearing, and it can judge according to instantaneous energy size the wearing and tearing of rolling bearing, has online real-time, can weigh the feature of the degree of wear.

In order to reach above-mentioned purpose, solution of the present invention is:

A recognition methods for rolling bearing wearing and tearing, comprises the steps:

(1) calculation bearing is in the characteristic frequency in different operating stage;

(2) bear vibration test signal is carried out to Short Time Fourier Transform, obtain Fourier transform spectrogram;

(3) frequency range is set, uses integral method to ask for this band energy as the calculating instantaneous energy of bearing characteristic frequency;

(4) calculate respectively the instantaneous energy under interior actual instantaneous energy of a period of time and nominal situation;

(5) take the instantaneous energy size at nominal situation place is baseline, setting threshold, the relatively trend of judgement wearing and tearing and the degree of bearing damage.

In above-mentioned steps (1), described characteristic frequency comprises following parameter:

Rolling body rotation frequency:

$f_b=\frac{1}{2}\frac{D_m}{n}[1-{\left(\frac{d\cos \mathrm{\α}}{D_m}\right)}^2]\×f$

Rolling body passes through inner ring frequency:

$f_{\mathrm{bi}}=\frac{1}{2}[(1+\frac{d\cos \mathrm{\α}}{D_m})\×Z]\×f$

Rolling body is by the frequency of outer ring:

$f_{\mathrm{bo}}=\frac{1}{2}[(1-\frac{d\cos \mathrm{\α}}{D_m})\×Z]\×f$

Retainer frequency or rolling body revolution frequency:

$f_{\mathrm{bo}}=\frac{1}{2}(1-\frac{d\cos \mathrm{\α}}{D_m})\×f$

Wherein, the gyro frequency that f is rotating shaft, D _mfor the central diameter of bearing, d is rolling body diameter, and Z is rolling body number, and α is contact angle.

In above-mentioned steps (3), dividing frequency band,, there is harmonic components in the appearance along with wearing and tearing near this characteristic frequency:

$H\left(\mathrm{\ω}\right)=\left\{\begin{array}{cc}0& \mathrm{\ω}\∉({\mathrm{\ω}}_{\min },{\mathrm{\ω}}_{\max })\\ 1& \mathrm{\ω}\∈({\mathrm{\ω}}_{\min },{\mathrm{\ω}}_{\max })\end{array}\right.$

In formula, ω _min, ω _maxbe respectively lower limit and the upper limit of signal analysis frequency range, the upper and lower limit of described frequency refers near characteristic frequency, centered by characteristic frequency and do not comprise the suitable width of other characteristic frequencies.

In above-mentioned steps (4), the Fourier transform STFT to choosing with composition after filtering _hin Frequency Domain Integration, obtain instantaneous energy:

$E_t={\∫}_{{\mathrm{\ω}}_{\min }}^{{\mathrm{\ω}}_{\max }}\left({\mathrm{STFT}}_H\right)\mathrm{d\ω}$

Wherein, E _tbe the instantaneous energy that signal is relevant to characteristic frequency, ω _min, ω _maxbe respectively lower limit and the upper limit of signal analysis frequency range.

In above-mentioned steps (5), the establishing method of described threshold value is: the mean value of instantaneous energy size under nominal situation is added to its 2 times of variances obtain.

Adopt after such scheme, the present invention is based near the feature that there will be corresponding harmonic components in wear process of characteristic frequency, premature wear only has harmonic components not have characteristic frequency to occur, adopt Short Time Fourier Transform to calculate in real time the characteristic frequency of different operating modes and bearing different parts, adopt instantaneous energy to calculate simultaneously and obtain the threshold value of weighing degree of wear size, carry out the differentiation of bearing premature wear and development trend, there is online real-time, can weigh the feature of the degree of wear.

Accompanying drawing explanation

Fig. 1 is process flow diagram of the present invention;

Fig. 2 is near the instantaneous energy comparison diagram of the present invention certain operating mode feature frequency.

Embodiment

Below with reference to accompanying drawing, technical scheme of the present invention and beneficial effect are elaborated.

As shown in Figure 1, the invention provides a kind of recognition methods of rolling bearing wearing and tearing, comprise the steps:

(1) calculation bearing is in the characteristic frequency in different operating stage;

If the gyro frequency of rotating shaft is f, the central diameter of bearing is D _m, rolling body diameter is d, rolling body number is Z, contact angle α.

Rolling body rotation frequency:

$f_b=\frac{1}{2}\frac{D_m}{d}[1-{\left(\frac{d\cos \mathrm{\α}}{D_m}\right)}^2]\×f$

Rolling body passes through inner ring frequency:

$f_{\mathrm{bi}}=\frac{1}{2}[(1+\frac{d\cos \mathrm{\α}}{D_m})\×Z]\×f$

Rolling body is by the frequency of outer ring:

$f_{\mathrm{bo}}=\frac{1}{2}[(1-\frac{d\cos \mathrm{\α}}{D_m})\×Z]\×f$

Retainer frequency or rolling body revolution frequency:

$f_{\mathrm{bo}}=\frac{1}{2}(1-\frac{d\cos \mathrm{\α}}{D_m})\×f$

(2) whether bear vibration test signal is carried out to Short Time Fourier Transform simultaneously, by characteristic frequency in Fourier transform spectrogram monitoring step (1), occur and size, be calculative determination characteristic frequency and its scope of step (3) instantaneous energy;

Short Time Fourier Transform definition: the window function r (t) that a given time width is very short, make window slide, the STFT of signal s (t) is defined as:

$\mathrm{STFT}(\mathrm{\τ},f)={\∫}_{-\∞}^{\∞}\left[s\left(t\right)r^{*}(t-\mathrm{\τ})\right]e^{-j2\mathrm{\πft}}\mathrm{dt}---\left(1\right)$

Signal s (t) is exactly that signal s (t) is multiplied by " analysis window " r centered by τ at the STFT of time τ ^*the Fourier transform that (t-τ) does, is equivalent to and takes out near the local spectrum of signal t=τ.After bandpass filtering

STFT _H=(STFT(τ,f))H(ω)（2）

(3) frequency range is set, uses integral method to ask for this band energy as the calculating instantaneous energy of bearing characteristic frequency;

Dividing frequency band,, there is harmonic components near this characteristic frequency in the appearance along with wearing and tearing

$H\left(\mathrm{\ω}\right)=\left\{\begin{array}{cc}0& \mathrm{\ω}\∉({\mathrm{\ω}}_{\min },{\mathrm{\ω}}_{\max })\\ 1& \mathrm{\ω}\∈({\mathrm{\ω}}_{\min },{\mathrm{\ω}}_{\max })\end{array}\right.---\left(3\right)$

In formula, ω _min, ω _maxbe respectively lower limit and the upper limit of signal analysis frequency range, the upper and lower limit of described frequency refers near characteristic frequency, centered by characteristic frequency and do not comprise the suitable width of other characteristic frequencies (as each order frequency of rotor), need to specifically determine according to the spectrogram of monitoring and different unit.

(4) calculate respectively actual instantaneous energy in a period of time and the instantaneous energy under nominal situation;

Fourier transform STFT to choosing with composition after filtering _hin Frequency Domain Integration, obtain instantaneous energy.

$E_t={\∫}_{{\mathrm{\ω}}_{\min }}^{{\mathrm{\ω}}_{\max }}\left({\mathrm{STFT}}_H\right)\mathrm{d\ω}---\left(4\right)$

E _tbe the instantaneous energy that signal is relevant to characteristic frequency.E _treflected signal time-frequency distributions is instantaneous overall moment t's, also can regard signal frequency as at [ω _min, ω _max] composition in scope at τ the energy in the period.

(5) take the instantaneous energy size at nominal situation place is baseline, and setting threshold, can be that the mean value of nominal situation instantaneous energy size adds its 2 times of variances, surpasses this threshold value and judges wearing and tearing, relatively the trend of judgement wearing and tearing and the degree of bearing damage.

Shown in Fig. 2, represent to adopt the energy comparison diagram of the present invention 55-90Hz under August 12,14 days, 15 days, 17 days, 18 days MC operating mode, wherein Short Time Fourier Transform counts 4096, frequency range arranges: 55-90Hz(MC operating mode bearing fault frequency is 68Hz), as can be seen from the figure, increase bearing roller wearing and tearing are in time embodied in the growth of instantaneous energy, after being increased in gradually August 14, can differentiate bearing because of wear damage, because instantaneous vibration energy sharply strengthens.

Above embodiment only, for explanation technological thought of the present invention, can not limit protection scope of the present invention with this, every technological thought proposing according to the present invention, and any change of doing on technical scheme basis, within all falling into protection domain of the present invention.

Claims (3)

1. a recognition methods for rolling bearing wearing and tearing, is characterized in that comprising the steps:

(1) calculation bearing is in the characteristic frequency in different operating stage;

(2) bear vibration test signal is carried out to Short Time Fourier Transform, obtain Fourier transform spectrogram;

(3) frequency range is set, uses integral method to ask for this band energy as the calculating instantaneous energy of bearing characteristic frequency;

Specifically, dividing frequency band,, there is harmonic components in the appearance along with wearing and tearing near this characteristic frequency:

$H\left(\mathrm{\ω}\right)=\left\{\begin{array}{cc}0& \mathrm{\ω}\∉({\mathrm{\ω}}_{\min },{\mathrm{\ω}}_{\max })\\ 1& \mathrm{\ω}\∈({\mathrm{\ω}}_{\min },{\mathrm{\ω}}_{\max })\end{array}\right.$

In formula, ω _min, ω _maxbe respectively lower limit and the upper limit of signal analysis frequency range, the upper and lower limit of described frequency refers near characteristic frequency, centered by characteristic frequency and do not comprise the suitable width of other characteristic frequencies;

(4) calculate respectively the instantaneous energy under interior actual instantaneous energy of a period of time and nominal situation;

Specifically, the Fourier transform STFT with composition after filtering to choosing _hin Frequency Domain Integration, obtain instantaneous energy:

$E_t={\∫}_{{\mathrm{\ω}}_{\min }}^{{\mathrm{\ω}}_{\max }}\left({\mathrm{STFT}}_H\right)\mathrm{d\ω}$

Wherein, E _tbe the instantaneous energy that signal is relevant to characteristic frequency, ω _min, ω _maxbe respectively lower limit and the upper limit of signal analysis frequency range;

(5) take the instantaneous energy size at nominal situation place is baseline, setting threshold, the relatively trend of judgement wearing and tearing and the degree of bearing damage.

2. the recognition methods of a kind of rolling bearing wearing and tearing as claimed in claim 1, is characterized in that: in described step (1), described characteristic frequency comprises following parameter:

Rolling body rotation frequency:

$f_b=\frac{1}{2}\frac{D_m}{d}[1-{\left(\frac{d\cos \mathrm{\α}}{D_m}\right)}^2]\×f$

Rolling body passes through inner ring frequency:

$f_{\mathrm{bi}}=\frac{1}{2}[(1+\frac{d\cos \mathrm{\α}}{D_m})\×Z]\×f$

Rolling body is by the frequency of outer ring:

$f_{\mathrm{bo}}=\frac{1}{2}[(1-\frac{d\cos \mathrm{\α}}{D_m})\×Z]\×f$

Retainer frequency or rolling body revolution frequency:

$f_c=\frac{1}{2}(1-\frac{d\cos \mathrm{\α}}{D_m})\×f$

Wherein, the gyro frequency that f is rotating shaft, D _mfor the central diameter of bearing, d is rolling body diameter, and Z is rolling body number, and α is contact angle.

3. the recognition methods of a kind of rolling bearing wearing and tearing as claimed in claim 1, is characterized in that: in described step (5), the establishing method of described threshold value is: the mean value of instantaneous energy size under nominal situation is added to its 2 times of variances obtain.