6-10 July 2008, Daejeon, Korea LASER DOPPLER VIBROMETER...
Transcript of 6-10 July 2008, Daejeon, Korea LASER DOPPLER VIBROMETER...
LASER DOPPLER VIBROMETER AND IMPULSE SIGNAL PHASE DEMODULATION IN ROTATION UNIFORMITY MEASUREMENTS
Jiri TUMA VSB Technical University of OstravaCzech Republic
6-10 July 2008, Daejeon, Korea
ICSV 15, 6-10 July 2008, Korea 2
Czech Republic
ICSV 15, 6-10 July 2008, Korea 3
Outline
Terminology noteMotivationTransducers & Signal Processing Methods Principle of phase demodulation using Hilbert transformAccuracy of Incremental Rotary EncodersComparison of Measurements Done with Laser and IRCConclusion
ICSV 15, 6-10 July 2008, Korea 4
Terminology Note: Nominal vs. Actual Angle of Rotation
Uniform rotation Non-uniform (irregular) rotation
360°360° 0°0°
360° 360°
0°0°
360°0°
Nominal angle of rotation, deg Nominal angle of rotation, deg
Actual angle of rotation, deg Actual angle of rotation, deg
Angle variation
Nominal angle of rotation, deg
It is assumed rotation at the steady-state rotational speed. In fact it is a time axis.
ϕ
ϕΔ
ϕΔ
ϕ
ICSV 15, 6-10 July 2008, Korea 5
Motivation
Angular vibration of a gear versus linear vibration on the gearbox housing (geared axis systems)Transmission error measurement of a gear trainMeasuring and analyzing car engine crankshaft rotational irregularities, caused by dynamic changes in torque
Angular vibration as one of the important machine vibration and noise source.
ICSV 15, 6-10 July 2008, Korea 6
Gear Angular VibrationTime : Time (Enhanced Time(Encoder))
-0,0016
0,0000
0,0016
0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0Nominal Revolution
deg
Time : Time (Time (Enhanced Time(Encoder))) - 0 to 100 ord
-60000-30000
03000060000
0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0Nominal Revolution
deg/
s2̂
Time : Order Analyzer : Enhanced Time(Vibrace H)
-10-505
10
0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0Nominal Revolution
m/s
^2
Linear acceleration onthe gearbox housing
Angular acceleration
Difference between actual and nominal rotation angle
Double differentiation
ICSV 15, 6-10 July 2008, Korea 7
Test Stand for Transmission Error (TE) Measurements
211
22 r
nnTE ⎟⎟
⎠
⎞⎜⎜⎝
⎛Θ−Θ=
-4-2024
0,0 0,2 0,4 0,6 0,8 1,0Nominal tooth pitch rotation
TE in
mic
rons
Back-to-back test rig
Simple gear train, tooth numbers n1, n2
2 incremental rotary encoders
Pinion
Wheel
2Θ1Θ
r2
Measurement range: ± 10 μm (microns)
( ) 211
22 r
nnmTE ⎟⎟
⎠
⎞⎜⎜⎝
⎛Θ−Θ=
ICSV 15, 6-10 July 2008, Korea 8
Rotational irregularities of a car engine evaluated using an impulse signal for ECU
-200
-150
-100
-50
0
50
100
150
200
250
300
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2
Revolution
rad/s2
-
+
Two complete revolutions
Compression strokes
Power strokes
Crankshaft angular acceleration
Engine idle speed: 800 RPM
ICSV 15, 6-10 July 2008, Korea 9
Transducers & Signal Processing Methods for Angular Vibration Measurements
Tangentially mounted accelerometersTorsiografLaser Torsional Vibration Meter (Doppler effect)Incremental rotary encoders(several hundreds of pulses per revolution)
Time interval length measurementsSample number & InterpolationHigh frequency oscillator (10 GHz) & Impulse counter (Rotec)
Quadrature mixingPhase demodulation
Transducers How to process impulse signals?
ICSV 15, 6-10 July 2008, Korea 10
Transducers
Dual-beam laser transducer based on the Doppler effectLaser: Ga-Al-As diode, 780 nm lightInstantaneous changes in angular velocityMeasurement ranges: 10, 100, 1000, 10000°/s Frequency range: 0.3 to 1000 Hz Accuracy: ±1% of full scale
Laser Torsional Vibration Meter, Brüel & Kjær Type 2523
Heidenhain Incremental Rotary Encoders
ERN 460-500 type, 500 impulses per revolution
ERN 460-1024 type, 1024 impulses per revolution
The maximum directional deviation is within ± 1/20 grating period.
ICSV 15, 6-10 July 2008, Korea 11
Signal Analyzers for impulse signals
PULSE, the BK Signal Analyzers
Frequency range FR:
25.6 kHz (65536 kHz samp freq)
102.4 kHz (256 kHz sampl freq)
Phase demodulation, including possible signal resampling for order analysis, requires 5 or 6 samples per impulse period at least.
ICSV 15, 6-10 July 2008, Korea 12
Principle of Phase Demodulation
-1,5-1,0-0,50,00,51,01,5
0,0 0,2 0,4 0,6 0,8 1,0Time, s
yi
How to evaluate the instantaneous phase or frequency of a phase modulated harmonic signal?
( )iii Eyarcsin=ϕ
Instantaneous phase may be evaluated for given sample yi and envelope Ei using a formula
Solution: The creation of an analytic signal to evaluate phase and envelope.
Problems: estimation of instantaneous envelope, measurement noise.
ICSV 15, 6-10 July 2008, Korea 13
Analytic Signal
fP π=ω 2
Real harmonic signal Complex analytic signal
(vanishing XN and multiplication XP by 2 )
Analytic signalHilbert transform =+ jTime signal
( ) ( ) ( )22 tytxtE +=Envelope
( ) ( ) ( )( )txtyt atan=ϕWrapped Phase
π+
π−
t+
Unwrapped Phase
Wrapped Phase( )tϕ
ICSV 15, 6-10 July 2008, Korea 14
Phase demodulation based on the Hilbert transform
Using FFT and Inverse FFT Using digital filters as the Hilbert transformer
2π
2π Hilbert
Transformer
x(t)x(t)
y(t)
Real part
Imaginary part
( ) ( ){ }kxFFTjX =ω
( ) ( ){ }ω= jYIFFTky
Impulse Response
-1,0-0,50,00,51,0
-16 -12 -8 -4 0 4 8 12 16Index n
Delay
ICSV 15, 6-10 July 2008, Korea 15
Accuracy of Incremental Rotary Encoders
Two incremental rotary encoders
PinionWheel
2Θ1Θ
r2
Gear train transmission error measurement
Required error is less than
10-4 deg (0.0001 deg) for a mesh cycle
ICSV 15, 6-10 July 2008, Korea 16
Comparison of two encoders of the Heidenhain ERN 460-500 type
-0,6
-0,4
-0,2
0,0
0,2
0,4
0,6
0,0000 0,0002 0,0004 0,0006 0,0008
Time, s
[V]
-100
-80
-60
-40
-20
0
0 10000 20000Frequency, Hz
RM
S dB
/ref 1
VE1E2
E1E2
500 impulses per revolution
Fs = 65536 Hz
104010441048105210561060
0,0 0,2 0,4 0,6 0,8Time, s
RPM
Both encoders rotate at the same rotational speed
Global uniformity of rotation
ICSV 15, 6-10 July 2008, Korea 17
Order of Averaging Operation and Phase Demodulation
Impulse time signals
Averaging in time domain
Phase demodulation
Frequency spectrum
Phase difference
Impulse time signals
Averaging in time domain
Phase demodulation
Frequency spectrum
Phase difference
Phase difference
Averaging in freq domain
Frequency spectrum
Inst frequency spectrum
BA
C
D E F
(Signal Enhancement in BK Signal Analyzers)
ICSV 15, 6-10 July 2008, Korea 18
Phase difference frequency spectra
-0,03-0,02-0,010,000,010,020,03
0,0 0,2 0,4 0,6 0,8 1,0
Nominal revolution
deg
-0,03-0,02-0,010,000,010,020,03
0,0 0,2 0,4 0,6 0,8 1,0
Nominal revolution
deg
AveragedPhaseDifferencesAveraged TimeSignals
Phase difference
1E-61E-51E-41E-31E-21E-1
1 10 100 1000
Order
RM
S d
eg
Averaged spectrumAveraged time signal
Phase difference
1E-61E-51E-41E-31E-21E-1
1 10 100 1000
Order
RMS
deg
Spectrum of the averaged impulse signals (D)
Averaged spectrum of phase differences (F) spectrum of averaged phase differences (E)
16 phase differences
Averaging of impulse signals (A) and phase differences (B)
C
A
B
FE
D
ICSV 15, 6-10 July 2008, Korea 19
Phase Difference between Two Encoders of the ERN 460-500 Type
E1E2
Phase difference
0,000001
0,000010
0,000100
0,001000
0,010000
0,100000
1,000000
1 10 100 1000Order
RM
S de
g 1040 RPM
Phase difference
-0,04
-0,03
-0,02
-0,01
0,00
0,01
0,02
0,03
0,04
0,0 0,2 0,4 0,6 0,8 1,0
Nominal Revolution
deg
Maximum directional deviation
360° 36° 3.6° 0.36°
Tooth-by-tooth rotation teethn
0360=
Wave length
Teeth
Averaged spectrum of phase differences
ICSV 15, 6-10 July 2008, Korea 20
Comparison of Measurements Done with Laser Torsional Vibration Meter and Incremental Rotary Encoder
Laser Torsional Vibration Meter Brüel & Kjær Type 2523 Incremental Rotary EncoderHeidenhain ERN 460-1024 typePULSE, the Brüel & KjærSignal AnalyzerSampling frequency 65536 Hz
Laser
Encoder Instrumentation
ICSV 15, 6-10 July 2008, Korea 21
Signal Processing Procedure
Band pass filtration with the centre frequency equal to the mean frequency of the impulse signal Computation of the Hilbert transform of the phase modulated harmonic signal using the Hilbert TransformerComputation of the wrapped phase using ATAN function for real and imaginary coordinates in the complex plane Unwrapping phaseComputation of the first derivative of the phase with respect to time to obtain a signal proportional to the instantaneous angular frequency
Laser Torsional Vibration MeterSensor output signal is proportional to the instantaneous angular velocity
Incremental rotary encoder
ICSV 15, 6-10 July 2008, Korea 22
Band Pass Filtration
-4
-2
0
2
4
0,000 0,001 0,001 0,002 0,002Time, s
V
-10123456
0,000 0,001 0,001 0,002 0,002Time, s
V
-100
-80
-60
-40
-20
0
20
0 5000 10000 15000 20000 25000Frequency, Hz
RM
S dB
/ref 1
V
Impulse signal Phase modulated harmonic signal
-100
-80
-60
-40
-20
0
20
0 5000 10000 15000 20000 25000Frequency, Hz
RM
S dB
/ref 1
V
ICSV 15, 6-10 July 2008, Korea 23
Differentiation with Respect to Time
Impulse Response : Ideal Diff FIR Filter
-150-100-50
050
100150
0,0000 0,0005 0,0010 0,0015 0,0020Time, s
gi
Phase demodulation
Differentiation with respect to time
Phase (Angle)
Angular velocity
Frequency Response : Ideal Diff FIR Filter
20304050607080
10 100 1000 10000Frequency, Hz
Mag
nitu
de in
dB
Impulse signal
Band pass filtration
Harmonic signal
Impulse response
ICSV 15, 6-10 July 2008, Korea 24
Averaged Time Signal
-200
-100
0
100
200
300
0,0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9Nominal revolution
deg/
s
LaserEncoder
-200
-100
0
100
200
300
0,00 0,01 0,02 0,03 0,04 0,05 0,06 0,07 0,08 0,09 0,10Nominal revolution
deg/
s
LaserEncoder
390 RPM6.5 Hz40.8 rad/s
2340 deg/s
Mean rotational speed
15.4 ms
ICSV 15, 6-10 July 2008, Korea 25
Averaged Frequency Spectrum
-6-3036
0 20 40 60 80 100 120 140 160Order
RM
S de
g/s
Difference
0
10
20
30
40
50
60
0 20 40 60 80 100 120 140 160Order
RM
S de
g/s
EncoderLaser
63 ord
1 kHz
ICSV 15, 6-10 July 2008, Korea 26
Conclusion
The paper describes two methods for angular vibration measurement during rotation The first one is based on the phase demodulation of impulse signals and the second one employs a two-beam Doppler laser vibrometer. The measurement method, which is based on using the phase demodulation, was demonstrated on the incremental rotary encoderaccuracy testing and measurement of the hand drill rotation uniformity. The second presented measurement demonstrates employing the laser vibrometer. Both angular vibration measurements result in almost the same time history and frequency spectrum.
ICSV 15, 6-10 July 2008, Korea 27
Thank you for your attention