Department of Go, change the world

RV College of Electrical & Electronics Engineering

Engineering

DSP EL PHASE-1

VIBRATION ANALYSIS OF ROTATING MACHINERY

S
Under the Guidance of,
Mrs. Jyothi R
Asst. Professor
EEE department
RVCE SUBMITTED BY:
Lathesh Shetty K K-1RV20EE402
Mallaraddy -1RV20EE403
Rida Arfain A -1RV20EE404
Sudarshan M J -1RV20EE405
 Go, change the world
RV College of
Engineering® TABLE OF CONTENTS
PHASE 1
• Introduction
• Problem Statement
• Methodology
• References
PHASE 2
• Accelerometers
• Gear faults
Time domain
Frequency domain
Time synchronous averaging
• Conclusion
• References
 Go, change the world
INTRODUCTION
RV College of
Engineering®

• Rotating machinery are used for industrial use in

manufacturing and processes, nuclear power station,
automobile etc.

• Bearing faults play a major role in failure of machine

subjected to fatigue and catastrophic failure.

• Condition monitoring vibration technique is widely

used to accurate assessment of the health condition of
the rotating machinery.
 RV College of Go, change the world
Engineering®

Why vibration analysis?

• Early detection of faults

• Reduce economical losses to production and equipment
• To note the stability
• To diagnosis faults

Possible faults:
1.Unbalance 6.Excessive Wear
7.Overload
2.Axial movement
8.Improper fit or
3.Bend shaft bad assembly.
4.Mechanical lossenes
5.Bad bearing
 RV College of Go, change the world
Engineering®

PROBLEM STATEMENT
● The aim is to analyze vibration signals from a gearbox using time-synchronous
averaging and envelope spectra.
 RV College of Go, change the world
Engineering® WHAT IS TIME SYNCHRONOUS AVERAGING?

• With rotating machinery, a single rotation period encapsulates the full gamut of machinery component
interaction.
• Averaging over uniform rotation angles or complete rotations, rather than averaging over more arbitrary
time segments, rejects any noise, disturbance, or periodic signal content that is not coherent with the
rotation.
• This rotation-based averaging is called time-synchronous averaging
• It is used to greatly reduce the effects of unwanted
noise in the measurement
The following syntax is used in MATLAB for TSA

ta = tsa(x,fs,tp)
ta = tsa(x,t,tp)
ta = tsa(xt,tp)
 Go, change the world
RV College of
Engineering® METHODOLOGY OUTLINE
• Generates and analyzes vibration data for a gearbox whose shafts rotate at a fixed
speed.
• Using MATLAB and Simulink platform to simulate and analyze.
 RV College of Go, change the world
Engineering®

REFERENCES

• [1]
https://www.ijert.org/research/a-review-of-vibration-analysis-techniques-for-rotating-machines-IJERTV4IS03082
3.pdf
• [2] https://www.sciencedirect.com/science/article/pii/S2351978918300696
• [3] https://www.globalspec.com/learnmore/motion_controls/power_transmission/gears/rack_pinion_gears
• [4]
https://www.erbessd-instruments.com/vibration-analyzers/?gclid=Cj0KCQiA5aWOBhDMARIsAIXLlkeBMeENb8
M3FcOEj5JT9iFx7l7nW89i25CtI5UfxOBbWDM5z2lQ5k8aAj2aEALw_wcB
• [5] https://power-mi.com/content/vibration-analysis-gearboxes
RV College of Go, change the world
Engineering®

PHASE 2
 Go, change the world
RV College of
Engineering® ACCELEROMETERS

• A1 and A2 accelerometers must be used for measuring vibration signals

from the pinion and bearing
A1 pinion singnals
A2 bearing signals

• Types of accelerometers are piezo electric, piezo resistive & capactive

• Digital accelerometers (3 axial or single axial)

• The accelerometers operate at a sample rate of 20 kHz.

machines produce complex vibration signals in which inherent operating

frequencies, random noise, and other failure frequencies can be combined,
both "standard" and specific failures associated with gear failure modes
 RV College of
DETEMINING GEAR AND MESH FREQUENCY Go, change the world
Engineering®

• The pinion rotates at a rate fPinion = 22.5 Hz or 1350

rpm. fs = 20E3; % Sample Rate (Hz)

Np = 13; % Number of teeth on pinion

The rotating speed of the gear and output shaft is Ng = 35; % Number of teeth on gear

fGear= fPinion x Number of pinion teeth (Np) fPin = 22.5; % Pinion (Input) shaft frequency (Hz)
Number of gear teeth (Ng )
fGear = fPin*Np/Ng; % Gear (Output) shaft frequency
The tooth-mesh frequency, also called gear-mesh (Hz)
frequency, is the rate at which gear and pinion teeth
fMesh = fPin*Np; % Gear Mesh frequency (Hz)
periodically engage:

fMesh= fPinion×Np = fGear×Ng .
 RV College of
WAVEFORM GENERATION Go, change the world
Engineering®

1. To Generate vibration waveforms for the pinion and the gear.

2. It is Modeled to sinusoids at primary shaft gear mesh frequencies.
3. 20 seconds of vibration data is analysed .

t = 0:1/fs:20-1/fs;

vfIn = 0.4*sin(2*pi*fPin*t); % Pinion waveform

vfOut = 0.2*sin(2*pi*fGear*t); % Gear waveform

vMesh = sin(2*pi*fMesh*t); % Gear-mesh wave

plot(t, vfIn + vfOut + vMesh)

xlim([0 0.25])
xlabel('Time (s)')
ylabel('Acceleration')
 RV College of
ON FAULT ON GEAR TOOTH Go, change the world
Engineering®

• A dent on the tooth surface of the gear generates

high-frequency oscillations over the duration of
the impact.

• The frequency of impact is dependent on

gearbox component properties and its natural
frequencies.

• In this example, it is arbitrarily assumed that the

impact causes a 2 kHz vibration signal and
occurs over a duration of about 8% of 1/fMesh,
or 0.25 milliseconds.

• The impact repeats once per rotation of the gear.

 RV College of Go, change the world
Engineering®

ipf = fGear;
fImpact = 2000;

tImpact = 0:1/fs:2.5e-4-1/fs;
xImpact = sin(2*pi*fImpact*tImpact)/3;

Make the impact periodic by convolving it with a comb function.

xComb = zeros(size(t));

Ind = (0.25*fs/fMesh):(fs/ipf):length(t);
Ind = round(Ind);
xComb(Ind) = 1;
xPer = 2*conv(xComb,xImpact,'same’);

Add the fault signal xPer to the shaft signal. Add white Gaussian noise to the output signals for both the fault-free and
the faulty gear to model the output from A2 

vNoFault = vfIn + vfOut + vMesh;

vFault = vNoFault + xPer;

vNoFaultNoisy = vNoFault + randn(size(t))/5;

vFaultNoisy = vFault + randn(size(t))/5;
 RV College of Go, change the world
Engineering®

subplot(2,1,1)
plot(t,vNoFaultNoisy)
xlabel('Time (s)')
ylabel('Acceleration')
xlim([0.0 0.3])
ylim([-2.5 2.5])
title('Noisy Signal for Healthy Gear')

subplot(2,1,2)
plot(t,vFaultNoisy)
xlabel('Time (s)')
ylabel('Acceleration')
xlim([0.0 0.3])
ylim([-2.5 2.5])
title('Noisy Signal for Faulty Gear')
hold on
MarkX = t(Ind(1:3));
MarkY = 2.5;
plot(MarkX,MarkY,'rv','MarkerFaceColor','red')
hold off
RV College of Go, change the world
Engineering®
 RV College of Go, change the world
Engineering®

Calculate the spectrum of the healthy and faulty gears. Specify a frequency range that includes the shaft frequencies at
8.35 Hz and 22.5 Hz and the gear-mesh frequency at 292.5 Hz.

[Spect,f] = pspectrum([vFaultNoisy' vNoFaultNoisy'],fs,'FrequencyResolution',0.2,'FrequencyLimits',[0 500]);

Plot the spectra. Because the fault is on the gear and not the pinion, sidebands are expected to appear at fsideband,Gear
and spaced fGear apart on the spectra. The spectra show the expected peaks at fGear, fPin, and fMesh. However, the
presence of noise in the signal makes the sideband peaks at fsideband,Gear indistinguishable.

figure
plot(f,10*log10(Spect(:,1)),f,10*log10(Spect(:,2)),':')
xlabel('Frequency (Hz)')
ylabel('Power Spectrum (dB)')

hold on
plot(fGear,0,'rv','MarkerFaceColor','red')
plot(fPin,0,'gv','MarkerFaceColor','green')
plot(fMesh,0,'bv','MarkerFaceColor','blue')
hold off

legend('Faulty','Healthy','f_{Gear}','f_{Pinion}','f_{Mesh}')
RV College of Go, change the world
Engineering®
 RV College of Go, change the world
Engineering®

Use the function TSA to generate time-synchronized waveforms for both the pinion and the gear.

Specify time-synchronized pulses for the pinion. Calculate the time-synchronous average for 10 rotations of the
pinion.

tPulseIn = 0:1/fPin:max(t);
taPin = tsa(vFaultNoisy,fs,tPulseIn,'NumRotations',10);

Specify time-synchronized pulses for the gear. Calculate the time-synchronous average for 10 rotations of the
gear.

tPulseOut = 0:1/fGear:max(t);
taGear = tsa(vFaultNoisy,fs,tPulseOut,'NumRotations',10);

Visualize the time-synchronized signals for a single rotation. The impact is comparatively easier to see on the
time-synchronous averaged signal for the gear, while it is averaged out for the pinion shaft. The location of the
impact, indicated on the plot with a marker, has a higher amplitude than neighboring gear-mesh peaks.

TheTSA function without output arguments plots the time-synchronous average signal and the time-domain
signals corresponding to each signal segment in the current figure.
 RV College of Go, change the world
Engineering®

figure

subplot(2,1,1)
tsa(vFaultNoisy,fs,tPulseIn,'NumRotations',10)
xlim([0.5 1.5])
ylim([-2 2])
title('TSA Signal for Pinion')

subplot(2,1,2)
tsa(vFaultNoisy,fs,tPulseOut,'NumRotations',10)
xlim([0.5 1.5])
ylim([-2 2])
title('TSA Signal for Gear')
hold on
plot(1.006,2,'rv','MarkerFaceColor','red')
hold off
RV College of Go, change the world
Engineering®
 Go, change the world
RV College of
Engineering® CONCLUSION

• Vibration signals from machinery part like gear , pinion and mesh can be monitored
and transformational tools like time to frequency domain transformation is used to
detect the faults and to check on the health. Time-synchronous averaging technique is
used to reduce the noise ,harmonics and to have a clear picture of the the vibration
signals and to attenuate the effects of small changes in shaft speed.
 Go, change the world
RV College of
Engineering® REFERENCES

[1]. Scheffer, Cornelius, and Paresh Girdhar. Practical Machinery Vibration Analysis and Predictive
Maintenance. Amsterdam: Elsevier, 2004.

[2]. Randall, Robert Bond. Vibration Based Condition Monitoring: Industrial, Aerospace and Automotive
Applications. Chichester, UK: John Wiley and Sons, 2011.

[3]. Lacey, S. J. An Overview of Bearing Vibration Analysis. (From:

http://www.maintenanceonline.co.uk/maintenanceonline/content_images/p32-42%20Lacey%20paper
%20M&AM.pdf)

[4]. Brandt, Anders. Noise and Vibration Analysis: Signal Analysis and Experimental Procedures.
Chichester, UK: John Wiley and Sons, 2011.
RV College of Go, change the world
Engineering®