CN111678698A - Rolling bearing fault detection method based on sound and vibration signal fusion - Google Patents

Rolling bearing fault detection method based on sound and vibration signal fusion Download PDF

Info

Publication number
CN111678698A
CN111678698A CN202010552045.5A CN202010552045A CN111678698A CN 111678698 A CN111678698 A CN 111678698A CN 202010552045 A CN202010552045 A CN 202010552045A CN 111678698 A CN111678698 A CN 111678698A
Authority
CN
China
Prior art keywords
signal
rolling bearing
fusion
representing
vibration
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010552045.5A
Other languages
Chinese (zh)
Other versions
CN111678698B (en
Inventor
石怀涛
李阳阳
白晓天
李献文
李思慧
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenyang Jianzhu University
Original Assignee
Shenyang Jianzhu University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shenyang Jianzhu University filed Critical Shenyang Jianzhu University
Priority to CN202010552045.5A priority Critical patent/CN111678698B/en
Publication of CN111678698A publication Critical patent/CN111678698A/en
Application granted granted Critical
Publication of CN111678698B publication Critical patent/CN111678698B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
    • G01M13/00—Testing of machine parts
    • G01M13/04—Bearings
    • G01M13/045—Acoustic or vibration analysis

Abstract

The invention provides a rolling bearing fault detection method based on sound and vibration signal fusion. Firstly, extracting a vibration signal and an acoustic signal of a rolling bearing to be detected, then establishing a sliding rectangular window function of the acoustic signal, then starting from a first signal point of the vibration signal to be detected, establishing a fusion signal after each movement by moving a rectangular window, solving a root mean square value and a uniform flexible value of the fusion signal to obtain a uniform sliding value, finding an optimal fusion signal, and finally diagnosing the fault of the bearing to be detected by judging a value range of approximate fault characteristic frequency where the optimal fusion signal is located; the invention solves the problems that the arrangement of the vibration sensor is limited, the amplitude at the fault characteristic frequency is low and the like; meanwhile, the problems that the signal-to-noise ratio is low and the like due to the fact that the acoustic signal is influenced by background noise are solved, the weak fault of the rolling bearing can be accurately identified, the diagnosis accuracy and the diagnosis efficiency are improved, and a good effect is achieved in the rolling bearing state monitoring.

Description

Rolling bearing fault detection method based on sound and vibration signal fusion
Technical Field
The invention relates to the technical field of bearing fault diagnosis, in particular to a rolling bearing fault detection method based on sound and vibration signal fusion.
Background
Rolling bearings are one of the most widely used components in rotary machines, and are widely used in important fields such as machining, metallurgy, chemical engineering, aviation and the like. As the rolling bearing is often in a high-temperature, high-speed and heavy-load working environment, the rolling bearing is easy to damage, and once the rolling bearing breaks down and is not detected in time, machine failure and unexpected halt can be caused, huge economic loss is caused, and even the personal safety of workers is threatened. Therefore, in the rotary machine, the failure diagnosis of the rolling bearing plays an important role.
The vibration signal of a rolling bearing contains a large amount of information and is extremely easy to acquire, and has been widely used in the past decades for fault diagnosis of the rolling bearing. However, the vibration signal is obtained by contact measurement, the arrangement of the vibration sensor is limited by the environment, the vibration sensor is not suitable for being installed in a high-temperature, high-humidity and high-corrosion working environment, the amplitude at the fault frequency is low, weak faults are not easy to detect, and the acoustic signal is obtained by non-contact measurement, so that the limitation of the contact measurement of the vibration signal can be improved, and the method becomes a new research method in fault diagnosis at present. However, in the process of acquiring the acoustic signal, the fault characteristic information is easily affected by background noise, so that the signal-to-noise ratio is low and the diagnosis precision is insufficient.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a rolling bearing fault detection method based on sound and vibration signal fusion, which comprises the following steps:
step 1: collecting vibration signals V 'and sound signals S' of a rolling bearing to be detected, and respectively storing the vibration signals V 'and the sound signals S' in a computer; sensor Q for collecting vibration signal V1Sensor Q mounted on bearing seat of rolling bearing to be tested and used for collecting acoustic signal S2Is arranged on a plane with a distance value d from the end surface of the bearing and satisfies the requirement of a sensor Q2Is located on the axis of the bearing;
step 2: extracting a segment containing N from the vibration signal VvThe vibration signal of each continuous signal point is used as the vibration signal V to be detected, namely V ═ V [ [ V ] ]u]|u＝1,2,…,Nv}; extracting a segment containing N from the acoustic signal SsThe acoustic signal of successive signal points is used as the acoustic signal S to be detected, i.e. S ═ S [ v ═ v]|v＝1,2,…,NsIn which N isvNumber of total signal points, N, representing vibration signal V to be detectedsRepresenting the total signal point number of the acoustic signal S to be detected;
and step 3: establishing a sliding rectangular window function aiming at an acoustic signal S to be detected, and setting the width and the moving step length of a rectangular window, wherein the width value of the rectangular window is a positive integral multiple of the rotating frequency of a rolling bearing to be detected, and defining the number of signal points of the rectangular window containing the acoustic signal as NO；
And 4, step 4: starting from a first signal point V1 of a vibration signal V to be detected, establishing a fusion signal of the vibration signal V to be detected and an acoustic signal S to be detected by moving a rectangular window;
and 5: calculating the fusion signal F using equation (1)jRoot mean square value of RMS (F)j)，
Figure BDA0002542915190000021
In the formula, FjRepresenting an acoustic signal SjAnd a vibration signal VjFusion signal of (D), Fj[i]Representing the fusion signal FjThe ith signal point of (1), NORepresenting the number of signal points contained in the rectangular window;
step 6: calculating the fusion signal F by using the formula (2) to the formula (3)jHarmonic signal SjMean softness value SM (F) betweenj,Sj)，
Figure BDA0002542915190000022
Figure BDA0002542915190000023
In the formula, Vj[i]Representing a vibration signal VjThe ith signal point of (1), Sj[i]Representing an acoustic signal SjThe (i) th signal point of (c),
Figure BDA0002542915190000024
representing the fusion signal FjIs determined by the average value of (a) of (b),
Figure BDA0002542915190000025
representing an acoustic signal SjAverage value of (d);
and 7: the fusion signal F is calculated by the formula (4)jAverage slip value RS ofj，
RSj＝RMS(Fj)×SM(Fj,Sj) (4)
And 8: after the rectangular window is moved for n-1 times, n uniform sliding values are obtained through calculation, and the maximum value RS in the n uniform sliding values is countedmaxTo RSmaxThe corresponding fusion signal is called the optimal fusion signal Fmax；
And step 9: finding the optimal fusion signal FmaxThe time domain signal of (a);
step 10: fourier transform is carried out on the time domain signal to obtain an optimal fusion signal FmaxThe frequency domain signal curve of (a);
step 11: calculating the outer ring fault characteristic frequency f of the rolling bearing to be measured by using the formula (5)oCalculating the inner ring fault characteristic frequency f of the rolling bearing to be tested by using the formula (6)iCalculating the rolling element fault characteristic frequency f of the rolling bearing to be measured by using the formula (7)bCalculating the characteristic frequency f of the retainer fault of the rolling bearing to be measured by using the formula (8)f，
Figure BDA0002542915190000031
Figure BDA0002542915190000032
Figure BDA0002542915190000033
Figure BDA0002542915190000034
Wherein z represents the number of rolling elements on the rolling bearing to be measured, dbIndicating the diameter of the rolling bodies on the rolling bearing to be measured, DmRepresenting the pitch circle diameter of the rolling bearing to be tested, α representing the contact angle of the rolling bearing to be tested, frRepresenting the rotation frequency of the rolling bearing to be tested;
step 12: at the optimal fusion signal FmaxFinding out the frequency value f corresponding to the maximum amplitude on the frequency domain signal curvemaxThen f ismaxNamely the fault characteristic frequency of the rolling bearing to be detected; when the frequency value fmaxSatisfy lambda1fo≤fmax≤λ2foThen, it is judged that the fault of the rolling bearing to be tested appears on the outer ring, wherein [ lambda ]1fo,λ2fo]Representing the value range of the approximate fault characteristic frequency of the outer ring; when the frequency value fmaxSatisfy lambda1fi≤fmax≤λ2fiThen, the fault of the rolling bearing to be tested is judged to appear on the inner ring, wherein [ lambda ]1fi,λ2fi]Representing the value range of the approximate fault characteristic frequency of the inner ring; when the frequency value fmaxSatisfy lambda1fb≤fmax≤λ2fbThen, it is determined that a fault of the rolling bearing to be tested occurs on the rolling element, wherein [ lambda ]1fb,λ2fb]Representing the value range of the approximate fault characteristic frequency of the rolling body; when the frequency value fmaxSatisfy lambda1ff≤fmax≤λ2ffThen, it is judged that a failure of the rolling bearing to be measured occurs on the cage, wherein [ lambda ]1ff,λ2ff]Representing the value range of the approximate fault characteristic frequency of the retainer; wherein λ1The value range is more than or equal to 0.9 lambda1≤0.95，λ2The value range is more than or equal to 1.05 lambda2≤1.1。
The step 4 specifically comprises the following steps: from a first signal point of an acoustic signal S to be detectedS[1]Firstly, sliding a rectangular window on the to-be-detected acoustic signal S according to the moving step length, and sequentially intercepting N on the to-be-detected vibration signal V when the rectangular window moves onceOUntil the rectangular window moves to the last moving step length, defining N intercepted after the jth movement of the rectangular windowOA section of vibration signal composed of signal points is Vj，Vj＝{Vj[i]|i＝1,2,…,NO}，Vj[i]Representing a vibration signal VjThe ith signal point of the rectangular window defines a section of acoustic signal corresponding to the jth movement of the rectangular window as Sj，Sj＝{Sj[i]|i＝1,2,…,NO}，Sj[i]Representing an acoustic signal SjThe ith signal point of (i), j is 0,1,2, …, n-1,
Figure BDA0002542915190000041
means not exceeding
Figure BDA0002542915190000042
Maximum positive integer of the ratio, NHNumber of signal points representing a predetermined step length of movement, defining an acoustic signal SjAnd a vibration signal VjThe fusion signal after the jth shift of the rectangular window is FjCalculating the acoustic signal S using equation (9)jAnd a vibration signal VjFusion signal F ofj，Fj[i]Representing the fusion signal FjThe ith signal point of (c);
Fj[i]＝Vj[i]+Sj[i]i＝1,2,…,NO(9)。
the invention has the beneficial effects that:
the invention combines the vibration signal and the running state information contained in the sound signal, and provides a rolling bearing fault detection method based on sound vibration signal fusion, wherein the sound vibration signal is fused, and the processed fusion signal is compared with the vibration signal to diagnose the fault of the rolling bearing, so that the signal-to-noise ratio and the sensitivity of fault diagnosis are effectively improved; moreover, the invention solves the problems that the arrangement of the vibration sensor is limited, the amplitude at the fault frequency is low and the like; meanwhile, the problems that the signal-to-noise ratio is low and the like due to the fact that the acoustic signal is influenced by background noise are solved, the weak fault of the rolling bearing can be accurately identified, the diagnosis accuracy and the diagnosis efficiency are improved, and a good effect is achieved in the rolling bearing state monitoring.
Drawings
Fig. 1 is a flowchart of a rolling bearing fault detection method based on acoustic-vibration signal fusion.
Fig. 2 is a frequency domain diagram of the envelope signal corresponding to the original vibration signal.
Fig. 3 is a frequency domain diagram of the original vibration signal corresponding to the fusion signal.
Detailed Description
The invention is further described with reference to the following figures and specific examples.
As shown in fig. 1, a rolling bearing fault detection method based on acoustic-vibration signal fusion includes the following steps:
step 1: collecting vibration signals V 'and sound signals S' of a rolling bearing to be detected, and respectively storing the vibration signals V 'and the sound signals S' in a computer; sensor Q for collecting vibration signal V1Sensor Q attached to bearing seat of rolling bearing to be measured and used for collecting acoustic signal S2Is arranged on a plane 420mm away from the end surface of the bearing and satisfies the sensor Q2Is located on the axis of the bearing;
step 2: extracting a segment containing N from the vibration signal VvThe vibration signal of a continuous signal point is used as the vibration signal V to be detected, namely V ═ V [ u [ u ] ]]|u＝1,2,…,Nv}; extracting a segment containing N from the acoustic signal SsThe acoustic signal of successive signal points is used as the acoustic signal S to be detected, i.e. S ═ S [ v ═ v]|v＝1,2,…,NsIn which N isvNumber of total signal points, N, representing vibration signal V to be detectedsNumber of total signal points, N, representing acoustic signals S to be detectedvValue 81920, NsThe value is 8192;
in order to highlight the impact component of the signal and make the fault characteristic frequency more obvious, the vibration signal and the acoustic signal are subjected to fusion processing, a sliding rectangular window function is established through MATLAB software, the appropriate rectangular window width and the appropriate moving distance are selected, and the fusion processing is carried out on the acoustic signalSliding; intercepting the number of signal points equal to the number of signal points contained in the sliding rectangular window on the vibration signal, and constructing an RSjAnd indexes are used for reconstructing the vibration signals and the acoustic signals so as to obtain the optimal fusion signals.
And step 3: establishing a sliding rectangular window function aiming at an acoustic signal S to be detected, and setting the width and the moving step length of a rectangular window, wherein the width value of the rectangular window is a positive integral multiple of the rotating frequency of a rolling bearing to be detected, and defining the number of signal points of the rectangular window containing the acoustic signal as NO，NOThe value is 4800;
and 4, step 4: starting from a first signal point V1 of a vibration signal V to be detected, establishing a fusion signal of the vibration signal V to be detected and an acoustic signal S to be detected by moving a rectangular window, specifically comprising the following steps:
from a first signal point S [1] of an acoustic signal S to be detected]Firstly, sliding a rectangular window on the to-be-detected acoustic signal S according to the moving step length, and sequentially intercepting N on the to-be-detected vibration signal V when the rectangular window moves onceOUntil the rectangular window moves to the last moving step length, defining N intercepted after the jth movement of the rectangular windowOA section of vibration signal composed of signal points is Vj，Vj＝{Vj[i]|i＝1,2,…,NO}，Vj[i]Representing a vibration signal VjThe ith signal point of the rectangular window defines a section of acoustic signal corresponding to the jth movement of the rectangular window as Sj，Sj＝{Sj[i]|i＝1,2,…,NO}，Sj[i]Representing an acoustic signal SjThe ith signal point of (i), j is 0,1,2, …, n-1,
Figure BDA0002542915190000051
means not exceeding
Figure BDA0002542915190000052
Maximum positive integer of the ratio, NHNumber of signal points representing a predetermined step length of movement, defining an acoustic signal SjAnd a vibration signal VjThe fusion signal after the jth shift of the rectangular window is FjCalculating the acoustic signal S using equation (9)jAnd vibration messageNumber VjFusion signal F ofj，Fj[i]Representing the fusion signal FjThe ith signal point of (c);
Fj[i]＝Vj[i]+Sj[i]i＝1,2,…,NO(9)。
there are many decision indicators in a stochastic resonance system (SR system for short), where a Root Mean Square Error (RMSE) and a smoothness indicator (SMO) increase with increasing noise intensity, show positive correlation, and are sensitive to noise, and a calculation formula of the root mean square error and the smooth motion indicator is as follows:
Figure BDA0002542915190000053
where k represents the signal length, x [ i ] represents the output signal, and Y [ i ] represents the input signal;
Figure BDA0002542915190000054
RMSE is used to express the difference degree between the output signal and the original input signal, and a new index RMS (F) is introduced for accurately calculating the RMS of the fusion signalj)；
And 5: calculating the fusion signal F using equation (1)jRoot mean square value of RMS (F)j)，
Figure BDA0002542915190000061
In the formula, FjRepresenting an acoustic signal SjAnd a vibration signal VjFusion signal of (D), Fj[i]Representing the fusion signal FjThe ith signal point of (1), NORepresenting the number of signal points contained in the rectangular window;
SMO is sensitive to noise, the average value is the centralized embodiment of the distribution of discrete signals, and in order to make the distribution of the discrete signals more centralized, a new index uniform soft value SM (F) is introduced on the basis of the SMOj,Sj)，
Step 6: using the formula (2) -formula (3) calculating the fusion signal FjHarmonic signal SjMean softness value SM (F) betweenj,Sj)，
Figure BDA0002542915190000062
Figure BDA0002542915190000063
In the formula, Vj[i]Representing a vibration signal VjThe ith signal point of (1), Sj[i]Representing an acoustic signal SjThe (i) th signal point of (c),
Figure BDA0002542915190000064
representing the fusion signal FjIs determined by the average value of (a) of (b),
Figure BDA0002542915190000065
representing an acoustic signal SjAverage value of (d);
therefore, in order to better highlight the positive correlation, make the evaluation index more sensitive to the noise intensity and correctly improve the fault characteristics, the invention uses RMS (F)j) And SM (F)j,Sj) Combining, constructing an evaluation index average slip value RSj，
And 7: the fusion signal F is calculated by the formula (4)jAverage slip value RS ofj，
RSj＝RMS(Fj)×SM(Fj,Sj) (4)
And 8: after the rectangular window is moved for n-1 times, n uniform sliding values are obtained through calculation, and the maximum value RS in the n uniform sliding values is countedmaxTo RSmaxThe corresponding fusion signal is called the optimal fusion signal Fmax；
And step 9: finding the optimal fusion signal FmaxThe time domain signal of (a);
step 10: fourier transform is carried out on the time domain signal to obtain an optimal fusion signal FmaxThe frequency domain signal curve of (a);
step 11: calculating the outer ring fault characteristic frequency f of the rolling bearing to be measured by using the formula (5)oCalculating the inner ring fault characteristic frequency f of the rolling bearing to be tested by using the formula (6)iCalculating the rolling element fault characteristic frequency f of the rolling bearing to be measured by using the formula (7)bCalculating the characteristic frequency f of the retainer fault of the rolling bearing to be measured by using the formula (8)f，
Figure BDA0002542915190000071
Figure BDA0002542915190000072
Figure BDA0002542915190000073
Figure BDA0002542915190000074
Wherein z represents the number of rolling elements on the rolling bearing to be measured, dbIndicating the diameter of the rolling bodies on the rolling bearing to be measured, DmRepresenting the pitch circle diameter of the rolling bearing to be tested, α representing the contact angle of the rolling bearing to be tested, frRepresenting the rotation frequency of the rolling bearing to be tested;
step 12: at the optimal fusion signal FmaxFinding out the frequency value f corresponding to the maximum amplitude on the frequency domain signal curvemaxThen f ismaxNamely the fault characteristic frequency of the rolling bearing to be detected; when the frequency value fmaxSatisfy lambda1fo≤fmax≤λ2foThen, it is judged that the fault of the rolling bearing to be tested appears on the outer ring, wherein [ lambda ]1fo,λ2fo]Representing the value range of the approximate fault characteristic frequency of the outer ring; when the frequency value fmaxSatisfy lambda1fi≤fmax≤λ2fiThen, the fault of the rolling bearing to be tested is judged to appear on the inner ring, wherein [ lambda ]1fi,λ2fi]Representing the value range of the approximate fault characteristic frequency of the inner ring; when the frequency value fmaxSatisfy lambda1fb≤fmax≤λ2fbThen, it is determined that a fault of the rolling bearing to be tested occurs on the rolling element, wherein [ lambda ]1fb,λ2fb]Representing the value range of the approximate fault characteristic frequency of the rolling body; when the frequency value fmaxSatisfy lambda1ff≤fmax≤λ2ffThen, it is judged that a failure of the rolling bearing to be measured occurs on the cage, wherein [ lambda ]1ff,λ2ff]Representing the value range of the approximate fault characteristic frequency of the retainer; wherein λ1The value range is more than or equal to 0.9 lambda1≤0.95，λ2The value range is more than or equal to 1.05 lambda2≤1.1。
The envelope processing is a common method for preprocessing an original signal, eliminates the noise of the original signal, enables the original signal to be smoother and softer, can improve the signal-to-noise ratio, carries out envelope processing on the vibration signal of the rolling bearing, and has the following envelope processing processes: inputting the collected vibration signals into Origin software, selecting the vibration signals, clicking an analysis function in Origin, selecting envelope processing, and obtaining vibration envelope signals.
Respectively inputting the envelope signal of the original vibration and the optimal fusion signal into an SR system to obtain a time domain signal curve; and respectively carrying out Fourier transform in the Origin system to obtain a frequency domain signal curve, and extracting the fault characteristics of the rolling bearing. A frequency domain signal curve corresponding to the envelope signal of the original vibration is shown in fig. 2, and a frequency domain signal curve of the optimal fusion signal passing through the SR system is shown in fig. 3.
After the envelope signal corresponding to the vibration signal and the optimal fusion signal are respectively input into the SR system, the signal-to-noise ratio of the envelope signal and the signal-to-noise ratio of the optimal fusion signal are obtained, and the signal-to-noise ratio of the optimal fusion signal is compared with the signal-to-noise ratio of the envelope signal, so that the sensitivity of fault diagnosis is remarkably improved.
According to the formulas (5) to (8) for calculating the fault characteristic frequency and the relevant parameters of the rolling bearing to be measured, the theoretical fault characteristic frequency values of different fault positions obtained in this embodiment are respectively: if the outer ring of the bearing to be tested breaks down, the theoretical fault characteristic frequency is 183.14Hz through calculation of a formula (5); if the inner ring of the bearing to be tested breaks down, the theoretical fault characteristic frequency is 296.86Hz through calculation of a formula (6); if the rolling body of the bearing to be tested fails, calculating by a formula (7) to obtain the theoretical failure characteristic frequency of 119.52 Hz; if the retainer of the bearing to be tested has a fault, the theoretical fault characteristic frequency is calculated to be 22.89Hz through a formula (8). The approximate fault characteristic frequency value range is defined as 95-105% of the theoretical fault characteristic frequency, namely lambda1＝0.95，λ2＝1.05。
In the embodiment, when the vibration signal of the rolling bearing is collected, the rotation frequency of the shaft is set to 60Hz, the amplitudes corresponding to the one-time frequency conversion and the two-time frequency conversion of the envelope signal in fig. 2 are obvious, and the amplitude corresponding to the fault characteristic frequency is weak, so that the fault of the detected bearing cannot be accurately diagnosed. Looking at the characteristic frequency corresponding to the higher amplitude of the optimal fusion signal in fig. 3, it can be seen that the fault characteristic frequency is 177.5Hz, and is near the theoretical outer ring fault characteristic frequency, so that the approximate fault characteristic frequency of the bearing outer ring is calculated, and ranges from 173.98Hz to 192.30Hz, and the fault characteristic frequency of the optimal fusion signal is in the range, so that the fault of the rolling bearing to be tested is diagnosed as the outer ring fault. Compared with the original vibration envelope signal, the signal fault characteristics obtained by the processing of the invention are obvious, the sensitivity and the signal-to-noise ratio of fault diagnosis are obviously improved, the fault of the bearing to be detected can be accurately diagnosed, and the accuracy and the diagnosis efficiency of diagnosis are improved.
The procedure was implemented in MATLAB as follows:
(1) login MATLAB interface
Storing the collected vibration signal of the rolling bearing to be tested in a document named as vibration signal txt, storing the sound signal in a document named as sound signal txt, and simultaneously storing two text documents in the same folder juxingchuang.m, wherein the specific programming is as follows:
function juxingchuang
clc
close all
a ═ load ('acoustic signal');
c ═ load ('vibration signal');
fs 16384; % sampling frequency
Ts＝1/Fs；
(2) Importing experimental data
The text document 'step 713 effective value 1.txt' is used for storing the acoustic signal of the rolling bearing to be tested, and the text document 'step 713 vibration 1.txt' is used for storing the vibration signal of the rolling bearing to be tested; and respectively dragging the text document 'step 713 effective value 1.txt' and 'step 713 vibration 1.txt' to the column 'current folder' on the left side of the MATLAB interface. In the juxingchuaang. m program, the 'step 713 effective value 1.txt' is used instead of the 'sound signal' and the 'step 713 vibration 1.txt' is used instead of the 'vibration signal', specifically programmed as follows:
function juxingchuang
clc
close all
a load ('step 713 valid value 1. txt'); % acoustic signal
c ═ load ('step 713 vibrate 1. txt'); % vibration signal
Fs 16384; % sampling frequency
Ts＝1/Fs；
(3) Running program
The width of the rectangular window is set to 4800 signal points, and the moving step length is set to 1 signal point number for more accurately obtaining the optimal fusion signal. The vibration signal contains 81920 signal points, the sound signal contains 8192 signal points, when the invention is specifically programmed, in order that the vibration signal and the sound signal contain the same signal points in each cycle, the vibration signal is averagely divided into 10 groups, the signal points in each group are the same as the signal points contained in the sound signal, and the RS is searched by cycling for 10 timesjTo obtain an optimal fusion signal. Clicking the "run" button in the "editor" column of the MATLAB SystemAnd (3) continuously calculating a key to obtain an optimal fusion signal, wherein the specific programming is as follows:
Figure BDA0002542915190000091
Figure BDA0002542915190000101
Figure BDA0002542915190000111
(4) inputting into SR system and calculating SNR
Additionally storing the time domain graph of the optimal fusion signal as u1.fig, wherein the file Untitled777.m can read the data of the x axis and the y axis of the picture; opening an Untittled777. m file in a folder, inputting u1. fig. in open ("), clicking a ' run ' key in the column of an MATLAB system ' editor ', obtaining an yc value of u1. fig. in the column of a ' working area ' on the right side of an MATLAB interface, and storing the yc value in a ' rectangular window y value ' txt ' text document, wherein the yc value refers to y-axis data, namely amplitude, of an optimal fusion signal time domain diagram, and the specific programming is as follows:
open('u1.fig')
lh＝findall(gca,'type','line')；
xc ═ get (lh, 'xdata'); % out x-axis data
yc ═ get (lh, 'ydata'); % fetch y-axis data
And carrying out envelope processing on the vibration signal, storing data into an envelope 1.txt 'text document, and dragging the envelope 1.txt' text document to a column of a current folder on the left side of an MATLAB interface. The files uh.m and sr111.m are used for SR processing of the original vibration envelope signal, the files uh.m and sr111.m in a folder are opened, an envelope 1.txt' is input in an s ═ load (") in a uh.m program, a key of" run "in a column of an MATLAB system" editor "is clicked, a time domain diagram of the original vibration envelope signal processed by the SR system is obtained, the time domain diagram is stored as u2.fig, and a signal-to-noise ratio of the envelope signal is obtained in a column of a" working area "on the right side of an MATLAB interface, and the specific programming is as follows:
Figure BDA0002542915190000121
Figure BDA0002542915190000131
and the file UHH.m is used for carrying out SR processing on the fusion signal, opening the UHH.m file in a folder, dragging the window function y value txt text document to the column of the current folder on the left side of the MATLAB interface, and inputting the window function y value txt' when s is load ("). Clicking a 'run' key in the 'editor' column of the MATLAB system to obtain a time domain graph of the fusion signal processed by the SR system, storing the time domain graph as u3.fig, and simultaneously obtaining the signal-to-noise ratio of the fusion signal in the 'working area' column on the right side of the MATLAB interface, wherein the specific programming is as follows:
calculating the signal-to-noise ratio function of function snr, mse ═ snr (I, IN)%
tn 4800; % signal length
Fs＝16384；
ts＝1/Fs；
t＝0:ts:(tn-1)*ts；
s-load ('window function y value txt');
D＝0.01；
white ═ sqrt (2 × D) × (1, tn); % Gaussian white noise production
signal ═ s' + white; superposition of a% windowed function processed signal and a noise signal
a＝0.1；b＝1；x0＝0.1；h＝0.1；
x＝sr111(x0,a,b,h,signal)；
x1＝17*x；
I ═ s; % original signal
IN ═ x 1; % denoised signal
Ps＝sum(sum((I-mean(mean(I))).^2))；
Pn＝sum(sum((I-IN').^2))；
snr＝10*log10(Ps/Pn)；
mse＝(Pn/length(I)).^0.5；
plot(t,x1)
end
(5) Separately Fourier transforming
Replacing 'u 1. fig' and 'u 2. fig' in the Untitled777.m program with 'u 2. fig' refers to a time domain diagram of an original vibration envelope signal processed by an SR system, clicking a 'run' key in the column of 'editor' of an MATLAB system to obtain xc and yc values of u2.fig in the column of 'working area' on the right side of an MATLAB interface, wherein the xc value refers to the time of the original vibration envelope signal, and the yc value refers to the amplitude of the original vibration envelope signal. The 'x 1.txt' text document is used to store the time of the original vibration envelope signal, the 'y 1.txt' text document is used to store the amplitude of the original vibration envelope signal, the xc value is stored in the 'x 1.txt' text document, and the yc value is stored in the 'y 1.txt' text document.
Similarly, 'u 3. fig' is used to replace 'u 2. fig' in the program of Untitled777.m, and u3.fig refers to the time domain diagram of the fusion signal processed by the SR system. Clicking a 'run' key in the 'editor' column of the MATLAB system to obtain the xc and yc values of u3. fig. in the 'working area' column on the right side of the MATLAB interface, wherein the xc value refers to the time of the fusion signal, and the yc value refers to the amplitude of the fusion signal. The 'x 2. txt' text document is used to store the time of the fusion signal, the 'y 2. txt' text document is used to store the amplitude of the fusion signal, the xc value is stored in the 'x 2. txt' text document, and the yc value is stored in the 'y 2. txt' text document.
The MATLAB system was turned off and the Origin system was turned on. Copying data in the text documents of 'x 1.txt' and 'y 1.txt' into Origin software, and performing Fourier change to obtain a frequency domain diagram of an original vibration envelope signal; similarly, the data in the text documents of 'x 2. txt' and 'y 2. txt' are copied into Origin software for Fourier change, and a frequency domain graph of the optimal fusion signal is obtained.
(6) Results
In the graph of the optimal fusion signal obtained by the invention, the amplitude at the fault characteristic frequency is obviously improved, the fault of the bearing to be tested can be accurately diagnosed, and compared with the envelope signal of the original vibration signal after envelope processing, the signal-to-noise ratio of the optimal fusion signal obtained by the invention is increased from-10.2059 dB to-3.9557 dB, so that the sensitivity of fault diagnosis is greatly increased, and the diagnosis efficiency is improved; experiments prove that the detection method provided by the invention has a good effect in fault diagnosis of the rolling bearing and has a certain help effect on monitoring the state of the rolling bearing.

Claims (2)

1. A rolling bearing fault detection method based on sound and vibration signal fusion is characterized by comprising the following steps:
step 1: collecting vibration signals V 'and sound signals S' of a rolling bearing to be detected, and respectively storing the vibration signals V 'and the sound signals S' in a computer; sensor Q for collecting vibration signal V1Sensor Q mounted on bearing seat of rolling bearing to be tested and used for collecting acoustic signal S2Is arranged on a plane with a distance value d from the end surface of the bearing and satisfies the requirement of a sensor Q2Is located on the axis of the bearing;
step 2: extracting a segment containing N from the vibration signal VvThe vibration signal of a continuous signal point is used as the vibration signal V to be detected, namely V ═ V [ u [ u ] ]]|u＝1,2,…,Nv}; extracting a segment containing N from the acoustic signal SsThe acoustic signal of successive signal points is used as the acoustic signal S to be detected, i.e. S ═ S [ v ═ v]|v＝1,2,…,NsIn which N isvNumber of total signal points, N, representing vibration signal V to be detectedsRepresenting the total signal point number of the acoustic signal S to be detected;
and step 3: establishing a sliding rectangular window function aiming at an acoustic signal S to be detected, and setting the width and the moving step length of a rectangular window, wherein the width value of the rectangular window is a positive integral multiple of the rotating frequency of a rolling bearing to be detected, and defining the number of signal points of the rectangular window containing the acoustic signal as NO；
And 4, step 4: starting from a first signal point V1 of a vibration signal V to be detected, establishing a fusion signal of the vibration signal V to be detected and an acoustic signal S to be detected by moving a rectangular window;
and 5: calculating the fusion signal F using equation (1)jRoot mean square value of RMS (F)j)，
Figure FDA0002542915180000011
In the formula, FjRepresenting an acoustic signal SjAnd a vibration signal VjFusion signal of (D), Fj[i]Representing the fusion signal FjThe ith signal point of (1), NORepresenting the number of signal points contained in the rectangular window;
step 6: calculating the fusion signal F by using the formula (2) to the formula (3)jHarmonic signal SjMean softness value SM (F) betweenj,Sj)，
Figure FDA0002542915180000012
Figure FDA0002542915180000013
In the formula, Vj[i]Representing a vibration signal VjThe ith signal point of (1), Sj[i]Representing an acoustic signal SjThe (i) th signal point of (c),
Figure FDA0002542915180000014
representing the fusion signal FjIs determined by the average value of (a) of (b),
Figure FDA0002542915180000015
representing an acoustic signal SjAverage value of (d);
and 7: the fusion signal F is calculated by the formula (4)jAverage slip value RS ofj，
RSj＝RMS(Fj)×SM(Fj,Sj) (4)
And 8: after the rectangular window is moved for n-1 times, n uniform sliding values are obtained through calculation, and the maximum value RS in the n uniform sliding values is countedmaxTo RSmaxThe corresponding fusion signal is called the optimal fusion signal Fmax；
And step 9: finding the optimal fusion signal FmaxThe time domain signal of (a);
step 10: fourier transform is carried out on the time domain signal to obtain an optimal fusion signal FmaxThe frequency domain signal curve of (a);
step 11: calculating the outer ring fault characteristic frequency f of the rolling bearing to be measured by using the formula (5)oCalculating the inner ring fault characteristic frequency f of the rolling bearing to be tested by using the formula (6)iCalculating the rolling element fault characteristic frequency f of the rolling bearing to be measured by using the formula (7)bCalculating the characteristic frequency f of the retainer fault of the rolling bearing to be measured by using the formula (8)f，
Figure FDA0002542915180000021
Figure FDA0002542915180000022
Figure FDA0002542915180000023
Figure FDA0002542915180000024
Wherein z represents the number of rolling elements on the rolling bearing to be measured, dbIndicating the diameter of the rolling bodies on the rolling bearing to be measured, DmRepresenting the pitch circle diameter of the rolling bearing to be tested, α representing the contact angle of the rolling bearing to be tested, frRepresenting the rotation frequency of the rolling bearing to be tested;
step 12: at the optimal fusion signal FmaxFinding out the frequency value f corresponding to the maximum amplitude on the frequency domain signal curvemaxThen f ismaxNamely the fault characteristic frequency of the rolling bearing to be detected; when the frequency value fmaxSatisfy lambda1fo≤fmax≤λ2foThen, it is judged that the fault of the rolling bearing to be tested appears on the outer ring, wherein [ lambda ]1fo,λ2fo]Representing the value range of the approximate fault characteristic frequency of the outer ring; when the frequency value fmaxSatisfy lambda1fi≤fmax≤λ2fiThen, the fault of the rolling bearing to be tested is judged to appear on the inner ring, wherein [ lambda ]1fi,λ2fi]Representing the value range of the approximate fault characteristic frequency of the inner ring; when the frequency value fmaxSatisfy lambda1fb≤fmax≤λ2fbThen, it is determined that a fault of the rolling bearing to be tested occurs on the rolling element, wherein [ lambda ]1fb,λ2fb]Representing the value range of the approximate fault characteristic frequency of the rolling body; when the frequency value fmaxSatisfy lambda1ff≤fmax≤λ2ffThen, it is judged that a failure of the rolling bearing to be measured occurs on the cage, wherein [ lambda ]1ff,λ2ff]Representing the value range of the approximate fault characteristic frequency of the retainer; wherein λ1The value range is more than or equal to 0.9 lambda1≤0.95，λ2The value range is more than or equal to 1.05 lambda2≤1.1。
2. The rolling bearing fault detection method based on sound and vibration signal fusion as claimed in claim 1, wherein the step 4 specifically comprises: from a first signal point S [1] of an acoustic signal S to be detected]Firstly, sliding a rectangular window on the to-be-detected acoustic signal S according to the moving step length, and sequentially intercepting N on the to-be-detected vibration signal V when the rectangular window moves onceOUntil the rectangular window moves to the last moving step length, defining N intercepted after the jth movement of the rectangular windowOA section of vibration signal composed of signal points is Vj，Vj＝{Vj[i]|i＝1,2,…,NO}，Vj[i]Representing a vibration signal VjThe ith signal point of the rectangular window defines a section of acoustic signal corresponding to the jth movement of the rectangular window as Sj，Sj＝{Sj[i]|i＝1,2,…,NO}，Sj[i]Representing an acoustic signal SjThe ith ofThe signal point, j-0, 1,2, …, n-1,
Figure FDA0002542915180000031
means not exceeding
Figure FDA0002542915180000032
Maximum positive integer of the ratio, NHNumber of signal points representing a predetermined step length of movement, defining an acoustic signal SjAnd a vibration signal VjThe fusion signal after the jth shift of the rectangular window is FjCalculating the acoustic signal S using equation (9)jAnd a vibration signal VjFusion signal F ofj，Fj[i]Representing the fusion signal FjThe ith signal point of (c);
Fj[i]＝Vj[i]+Sj[i]i＝1,2,…,NO(9)。
CN202010552045.5A 2020-06-17 2020-06-17 Rolling bearing fault detection method based on sound and vibration signal fusion Active CN111678698B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010552045.5A CN111678698B (en) 2020-06-17 2020-06-17 Rolling bearing fault detection method based on sound and vibration signal fusion

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010552045.5A CN111678698B (en) 2020-06-17 2020-06-17 Rolling bearing fault detection method based on sound and vibration signal fusion

Publications (2)

Publication Number Publication Date
CN111678698A true CN111678698A (en) 2020-09-18
CN111678698B CN111678698B (en) 2022-03-04

Family

ID=72455305

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010552045.5A Active CN111678698B (en) 2020-06-17 2020-06-17 Rolling bearing fault detection method based on sound and vibration signal fusion

Country Status (1)

Country Link
CN (1) CN111678698B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112362368A (en) * 2021-01-14 2021-02-12 西门子交通技术(北京)有限公司 Fault diagnosis method, device and system for train traction motor and readable medium
CN112926014A (en) * 2021-01-19 2021-06-08 北京化工大学 Rolling bearing acoustic signal multiband fusion fault diagnosis method based on RLS and RSSD
CN114496218A (en) * 2022-01-07 2022-05-13 西南交通大学 Structural state non-contact diagnosis method and system based on visual perception
CN114509158A (en) * 2022-01-04 2022-05-17 东南大学 Acoustic-vibration fused blade crack fault detection method and application
WO2024045387A1 (en) * 2022-08-31 2024-03-07 南方电网调峰调频发电有限公司储能科研院 "listening-vibration" combined gas leakage monitoring method for sealed device in power plant

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104951082A (en) * 2015-07-09 2015-09-30 浙江大学 Brain-computer interface method for intensifying EEG (electroencephalogram) signals through stochastic resonance
CN105303181A (en) * 2015-11-04 2016-02-03 燕山大学 Stochastic resonance weak impact feature enhancement extraction method on the basis of sliding window
CN105987809A (en) * 2015-02-10 2016-10-05 沈阳透平机械股份有限公司 Centrifugal-compressor semi-open-type impeller crack detection method based on random resonance
CN107506710A (en) * 2017-08-15 2017-12-22 河北建设集团股份有限公司 A kind of rolling bearing combined failure extracting method
GB2543522B (en) * 2015-10-20 2018-06-20 Skf Ab Method and data processing device for detecting bearing defects
CN109855874A (en) * 2018-12-13 2019-06-07 安徽大学 A kind of accidental resonance filter of sound ancillary vibration small-signal enhancing detection
CN111178327A (en) * 2020-01-16 2020-05-19 佛山科学技术学院 Deep learning-based bearing state identification method and system

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105987809A (en) * 2015-02-10 2016-10-05 沈阳透平机械股份有限公司 Centrifugal-compressor semi-open-type impeller crack detection method based on random resonance
CN104951082A (en) * 2015-07-09 2015-09-30 浙江大学 Brain-computer interface method for intensifying EEG (electroencephalogram) signals through stochastic resonance
GB2543522B (en) * 2015-10-20 2018-06-20 Skf Ab Method and data processing device for detecting bearing defects
CN105303181A (en) * 2015-11-04 2016-02-03 燕山大学 Stochastic resonance weak impact feature enhancement extraction method on the basis of sliding window
CN107506710A (en) * 2017-08-15 2017-12-22 河北建设集团股份有限公司 A kind of rolling bearing combined failure extracting method
CN109855874A (en) * 2018-12-13 2019-06-07 安徽大学 A kind of accidental resonance filter of sound ancillary vibration small-signal enhancing detection
CN111178327A (en) * 2020-01-16 2020-05-19 佛山科学技术学院 Deep learning-based bearing state identification method and system

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
郑平: "基于随机共振微弱信号检测的滚动轴承故障诊断方法研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112362368A (en) * 2021-01-14 2021-02-12 西门子交通技术(北京)有限公司 Fault diagnosis method, device and system for train traction motor and readable medium
CN112926014A (en) * 2021-01-19 2021-06-08 北京化工大学 Rolling bearing acoustic signal multiband fusion fault diagnosis method based on RLS and RSSD
CN112926014B (en) * 2021-01-19 2023-08-29 北京化工大学 Rolling bearing acoustic signal multiband fusion fault diagnosis method based on RLS and RSSD
CN114509158A (en) * 2022-01-04 2022-05-17 东南大学 Acoustic-vibration fused blade crack fault detection method and application
CN114496218A (en) * 2022-01-07 2022-05-13 西南交通大学 Structural state non-contact diagnosis method and system based on visual perception
CN114496218B (en) * 2022-01-07 2022-11-18 西南交通大学 Structural state non-contact diagnosis method and system based on visual perception
WO2024045387A1 (en) * 2022-08-31 2024-03-07 南方电网调峰调频发电有限公司储能科研院 "listening-vibration" combined gas leakage monitoring method for sealed device in power plant

Also Published As

Publication number Publication date
CN111678698B (en) 2022-03-04

Similar Documents

Publication Publication Date Title
CN111678698B (en) Rolling bearing fault detection method based on sound and vibration signal fusion
Wang et al. A novel weighted sparse representation classification strategy based on dictionary learning for rotating machinery
Huo et al. Incipient fault diagnosis of roller bearing using optimized wavelet transform based multi-speed vibration signatures
CN107228766B (en) Based on the Fault Diagnosis of Roller Bearings for improving multiple dimensioned fuzzy entropy
CN108982107B (en) It is a kind of based on morphology and it is multiple dimensioned arrangement entropy mean value bearing fault quantify trend diagnosis method
Dou et al. A rule-based intelligent method for fault diagnosis of rotating machinery
JP6792746B2 (en) State identification method by segment feature analysis of frequency domain
Popescu et al. Fault detection of rolling element bearings using optimal segmentation of vibrating signals
CN111504645B (en) Rolling bearing fault diagnosis method based on frequency domain multipoint kurtosis
CN109932179A (en) A kind of rolling bearing fault testing method based on the reconstruct of DS Adaptive spectra
Zheng et al. An improved empirical wavelet transform and refined composite multiscale dispersion entropy-based fault diagnosis method for rolling bearing
CN104132967A (en) Low-concentration gas detection method based on double threshold stochastic resonance
CN109492708A (en) Missing data interpolating method is detected in a kind of pipe leakage based on LS-KNN
CN108444715A (en) Bearing state diagnostic method, device, storage medium and electronic equipment
CN114755017B (en) Variable-speed bearing fault diagnosis method of cross-domain data driving unsupervised field shared network
CN1120366C (en) Fault detecting and diagnosing method based on non-linear spectral analysis
Li et al. Refined composite variable-step multiscale multimapping dispersion entropy: A nonlinear dynamical index
Zhang et al. Complementary ensemble adaptive local iterative filtering and its application to rolling bearing fault diagnosis
Huo et al. Edge permutation entropy: An improved entropy measure for time-series analysis
Akcan et al. Diagnosing bearing fault location, size, and rotational speed with entropy variables using extreme learning machine
Han et al. Fault diagnosis of rolling bearings using dual-tree complex wavelet packet transform and time-shifted multiscale range entropy
CN111947929B (en) Method for analyzing working condition sensitivity of vibration characteristic evaluation index of rotary machine
Chen et al. Performance degradation assessment of rotary machinery based on a multiscale Tsallis permutation entropy method
Puchalski et al. Data-driven monitoring of the gearbox using multifractal analysis and machine learning methods
Li et al. A novel scheme based on modified hierarchical time-shift multi-scale amplitude-aware permutation entropy for rolling bearing condition assessment and fault recognition

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant