Impact Testing and Frequency Response Functions
Monitoring and Pass / Fail Testing
Waterfall Plot with Order
Tracking
Analyze Only a Portion of Data
Acquired
Double Integration to Obtain Velocity and Displacement from
Measured Acceleration
Octave Band Analysis
Nyquist + Real & Imaginary Plots
Orbit Plot, Lissajous Plot, X-Y Plot,
Hysteresis
Impact Testing and Frequency Response Functions
For various impact tests, such as modal and dynamic stiffness, the upper
display indicates whether the hammer hit was valid. Checks include looking for double
hits, and whether the force or response values were too high or too low. If the impact was
not valid, the upper display shows red, and the bad data is not averaged to calculate the
frequency response function. Plots include magnitude and phase of the frequency response
function, as well as coherence.
Monitoring
& Pass / Fail
Testing
Plot shows reference template of acceptable values for frequency spectrum. Second
waveform shows newly acquired spectrum. Digital meter turns red if the newly acquired data
exceeds the template and shows the amplitude and frequency where the template was
exceeded.
Waterfall plot of amplitude vs. order and RPM on the Y, X
and Z scales respectively. The 53rd order is dominant in the middle of the
plot. Also, clearly seen is the second harmonic (106th order) and near the
front of the plot the 26.5 order is seen. Speed is increasing toward the
front of the plot and is the most recent data. Resonances curve to the
right moving toward the back of the plot.
Spectrogram shows same data as waterfall plot above.
Advantage of spectrogram (also called Campbell diagram) is that more data
can be seen on the plot compared to a waterfall plot and there are no
hidden peaks. It is clearly seen that when and order and a resonance
intersect, the amplitude increases as designated by the brighter colors
representing higher amplitudes.
Spectra can be overplotted to see variation over RPM.
Think of this as a poor man's waterfall plot. Meters read RPM and dB level
respectively.
Top right shows variation of amplitude with revolutions.
This data was acquired in the revolution (angle domain), not the
time domain. Performing a real-time FFT produces the lower plots. Lower
left is over entire order range and bottom right is zoomed around the 53rd
order to show modulation effect of the operating speed. Traditional
Cepstrum analysis is shown on top right. See next plot for more intuitive
Cepstrum analysis.
This instrument replays data from a previously recorded data file (left plot) and
analyzes its spectrum (right plot). User defines the region of data to analyze with
markers to analyze only that specific range of data.
Acceleration (red) is the measured waveform. By
integrating acceleration, velocity was derived (green). By double
integrating acceleration, displacement was calculated. As expected
displacement is 180 degrees out of phase with acceleration (peaks when
acceleration has valley), and velocity is 90 degrees out of phase with
acceleration (has a zero crossing when acceleration peaks).
Double integration is tricky because as the frequency
approaches 0 Hz, any noise becomes magnified. Therefore, high-pass
filtering is required.
Octave Band Analysis
1/3 Octave Plot of Sound Pressure Level.
Values are given in dB as an option.
Nyquist Plot is in the upper right.
The real and imaginary plots are on the top left of the Nyquist plot.
The lower plots are the magnitude and phase for
comparison.
Orbit Plot with Changes in Amplitude and Phase between the two
measurement points that are physically located 90 degrees apart. Orbit
plot on right is an ellipse because the amplitudes are not identical. The
ellipse is not horizontal (long axis) because the phase angle between the
two inputs is not 90 out of phase, as expected from the transducer
positions. They are actually 150 degrees out of phase.
An
X-Y plot or Lissajous pattern are general plots of one measurement vs.
another. The orbit plot has the two similar sensors 90 degrees apart so
the ideal plot is a circle. A hysteresis plot has two similar sensor with
no initial phase shift except what is measured.
Contact HEM Data for more
examples.
|
