|
Acceleration
enveloping
Acceleration
Enveloping or demodulation is a signal processing technique which greatly
enhances an analyst's ability to determine the condition of rotating
equipment. Resident in the Data Collectors/Analysers and Continuous
Monitoring units used and supplied by the
aptGroup, the technique enables
the detection of impulse fault signals such as bearing defects much
earlier than traditional analysis techniques permit.
Very
simply, the envelope's signal process converts the higher frequency
harmonics of the defect frequency into frequency components within the FFT
spectrum range. Every vibration signal generates harmonics. If the dynamic
range of the Data Collector is high enough these signals can be seen for
several orders of the fundamental.
If
the dynamic range is low, the harmonic signals are down in the noise floor
(sometimes referred to as being down in the dirt). Even with a high dynamic
range, the harmonics still disappear within a short span and cannot be
seen. The key to detecting bearing faults is to capture the low amplitude
bearing defect harmonics without including the high amplitude rotational
vibration signals. To accomplish this, the user selects from up to four
"band pass" filters and two envelope measurement methods"
enveeloped acceleration and enveloped velocity.
Band
pass filters are selected based upon the frequency and vibration units of
interest. The filters are; 5 to 100 Hz, 50 to 1 kHz, 500 to 10 kHz and
5kHz to 40 kHz.
Why
filter?
The
purpose of the band pass filter is to reject the high amplitude, low
frequency rotational signals such as 1X, 2X, 3X and only amplify the high
frequency harmonic components of the repetitive impulse signals. Excluding
these higher amplitude rotation components from the signal path results in
a significant improvement of the signal to noise ratio, allowing these
very small harmonic components to be detected and converted by the
envelope process down to the fundamental defect frequency and its
associated harmonics.
Selecting
the Band Pass Filter
The
selection of the Band Pass Filter is related mainly to the shaft speed.
Depending on the application and transducer location, the best filter
selection is a choice between rejection of low frequency rotational
signals, wide pass band of harmonic bearing defect components and
rejection of unstable higher frequency noise. Although the low frequency
corner limit of the envelope filter should be at least ten times higher
than rotational speed, the optimum selection might be the filter with the
next highest corner that would pass more defect harmonic components and
still adequately suppress higher frequency noise.
Essentially
the best envelope filter passes the highest number of signal components
related to the bearing defect and rejects all other non-related signals.
In summary, the envelope/filter process enhances the measurement of
impulsive signals while suppressing the higher amplitude rotational signals
as well as broad band random noise. For example, assume an 1800 RPM motor
with a bearing that has a BPFO (Ball Pass Frequency Outer Race) of 107.6
Hz. Since the speed is 30 Hz (1800 RPM) the filter low limit should be at
least 300 Hz.
Although
this analysis technique works very well for bearings it also provides
early warning of other types of faults such as those in gear boxes, as
well as misalignment and looseness. It also enables detection and analysis
of very low frequency defects on low RPM machines using standard
equipment. Any type of repetitive impulse vibration that generates higher
harmonics will be shown as a repetitive harmonic series in the fundamental
spectrum display.
What
is Enveloping?
Envelope
detection is a method for intensifying the repetitive components of a
dynamic signal to provide an early warning of deteriorating mechanical
condition. Common applications are concerned with rolling element bearings
and gear-mesh fault analyses.
The
vibration frequency caused by a race or anti-friction roller bearing
defect depends upon how often the defect strikes another part of the
bearing. The rolling element impacts the defect and the repetitive impulse
depends on number of balls, bearing geometry, and defect location. For
example, if there is a chip on an outer race, each roller will strike it
as it goes by and cause a vibration signal. This signal can often be
identified as some multiple of shaft rotational frequency. The multiple is
estimated by knowing a bearing's geometry and number of rollers.
A
vibration signal from a defective bearing is made up of low frequency
signals from rotational components, defect impulse signals, and machine
noise. Often bearing fault signals are of very short duration which
translate in the frequency domain as small harmonic amplitudes spread over
a wide frequency range and buried in machine noise. Machine noise masks
the early stages of bearing faults making spectrum analysis alone a
difficult diagnostic tool.
Enveloping
analysis first filters out the low frequency rotational components from
the complex signal. The high frequency repetitive components are enhanced
and converted down to the bearing spectrum range while machine noise is
reduced by a significant signal-to-noise factor. If vibration amplitudes
appear in the envelope spectrum that is related to bearing defect
frequencies it can be deduced that an incipient bearing defect is in
progress.
Envelope
analysis techniques permit an earlier prognosis of an eventual bearing
failure by reducing masking noise and by enhancing the significant
spectral components relating to bearing performance.
The
output of the envelope detector can be measured as RMS average (ENV AVE)
which signifies the vibration energy of the measurement, or as envelope
peak (ENV PEAK) which is mainly influenced by the signal crest factor
(true peak value divided by true RMS value). Initial bearing wear probably
shows more change in ENV PEAK when viewed in a trend comparison. As a
bearing defect broadens, ENV AVE will be the better diagnostic trend
indicator.
For
more information, read our Frequently
Asked Questions (FAQ) about Enveloping
Top
of page
]
|
