White Paper

April 2019 Machinery Health

The Evolution of PeakVue™ Processing –

PeakVue Plus Analytics
PeakVue has long been recognized as the premiere technology for detecting rolling element bearing
defects. Since its release in 1995, PeakVue applications have continually expanded to cover new
areas of machinery health monitoring and analysis. PeakVue Plus Analytics take PeakVue a step
further to provide prescriptive analytics.
White Paper
April 2019 Machinery Health

Introduction
The peak value analysis (PeakVue) methodology introduced by Emerson for the analysis of impacting on
machines has long been recognized as the premiere tool for identifying bearing defects.

The Evolution of PeakVue

The Expert’s Tool: First released in 1995, PeakVue processing quickly became the expert’s best tool for
diagnosing rolling element bearing defects. It had the ability to visualize the impacts caused by bearing
defects in the frequency spectrum more clearly than ever before. In the example in Figure 1, despite the
relatively low level amplitude of impacting, the peaks are
clearly visible in the PeakVue spectrum. In the
corresponding demodulation spectrum, the peaks
associated with the defect would be lost in the “noise”
providing little or no indication of an actual defect being
present.

In fact, PeakVue was able to identify a defect during the

initial stage of a developing fault when there was only
sub-surface deterioration – or weakening - of the bearing
metal, but without any visible sign of bearing wear to the
human eye. This forced end users to delay their Figure 1 - PeakVue Spectrum showing clear
recommendation for bearing maintenance. indication of bearing defect frequencies

At this point in time, however, PeakVue remained “the expert’s tool” as it took special knowledge and
training to properly set up the acquisition parameters

Automated Set-up: To address the complexity of setting up a PeakVue measurement and to increase
accessibility to this innovative new technology, an automated configuration module was implemented as
part of the Analysis Experts (Figure 2). First introduced
in the CSI 2120, the “Bearing and Gear Analysis”
application is still available as part of the AMS 2140.
This feature allows end users to execute complex
measurements using PeakVue processing with little or
no training in advanced vibration techniques. At the
touch of a button, the unit automatically configures
the acquisition parameters, opening up access to the
power of PeakVue processing to all end users.

Figure 2: PeakVue Analysis Expert

White Paper
April 2019 Machinery Health

PeakVue Waveform: While PeakVue provided superior results in the frequency spectrum, its usefulness
was not limited to the vibration spectrum. PeakVue also provides valuable diagnostic information in the
waveform. This made it unique and more powerful than other techniques as some defects simply can’t be
detected in the spectrum (e.g. very low
speed bearing faults, broken gear tooth,
etc.).

Figure 3 shows impacting from an inner

race defect that occurred a shaft turning
at slower than 1 RPM. This type of
defect cannot be effectively identified in
the spectrum due to the long collection Figure 3 - PeakVue waveform showing inner race defect
time and the relative fluctuations in the on shaft turning at 1/2 RPM
turning speed. Further, the amplitude of the
impacting (less than 0.1 g’s) would be further diminished in the spectrum, so that it would fall below the
signal to noise threshold. Instead, we have a clear picture of the actual signal caused by the inner race
damage. The impacting increases significantly (10X) as the defect passes through the load zone, and then
fall back down to extremely low levels.

PeakVue as Trend Parameter: In 2010, as PeakVue readings were first broadcast into the control room, it
became clear that the PeakVue “overall” provided one of the most valuable trend parameters for
monitoring machine health. The
trend of the peak value in the
PeakVue waveform correlates
directly to the health of the
bearing. Figure 4 shows a trend of
the overall vibration (blue line) on
a machine as the bearing
progresses to failure. The blue
Figure 4 – PeakVue trend (red) show clear indication of a developing fault as a bearing line appears to fluctuate
progresses to failure, while Overall Vibration (blue) provides no indication. randomly without providing any
indication of the impending
production shutdown. In comparison, the PeakVue trend (red line) indicates the onset of the failure around
Nov. 28th as it lifts off its baseline and shows a steady increase in the trend value – reaching about 40 g’s
before the bearing fails on Dec. 11th.

As an initial premise for operators to interpret the PeakVue trend, we can assert that the PeakVue value on
a good machine - properly installed and well lubricated - should be at or close to zero. This became known
as the “Zero Principle”, and it allowed an operator to verify easily that the bearing and lubrication on a
machine were operating within normal limits.
White Paper
April 2019 Machinery Health

For most process equipment (e.g. machines with turning speeds between 900 and 4000 RPM), once the
PeakVue value has reached a level of 10 g’s, it is an indication that there is an abnormal situation on the
bearing. Furthermore, the PeakVue value increases steadily as the bearing condition deteriorates. This led
to the development of the “Rule of 10’s” for monitoring most process equipment, which is summarized in
Figure 5. It states that when the PeakVue value measured in g’s Peak reaches the level of 10, that here is an

Figure 5 - Rule of 10's

abnormal situation on the bearing. If it doubles to 20, then it becomes a serious abnormal situation. If it
continues to double to 40, then it has become a critical abnormal situation. Apply the “Zero Principle” and
the “Rule of 10’s”, an operator with no vibration training is now able to identify the following aspects of
machine condition:

 Know when the machine is healthy

 Detect the presence of a machine defect
 Monitor the severity of the defect

PeakVue Plus Analytics: The latest evolution of this ground-breaking PeakVue technology is referred to as
“PeakVue Plus Analytics”. This innovative approach mimics the process that an analyst would apply to
determine the nature of an abnormal situation, once it has been identified on a machine based on the
PeakVue level. At a high level, machine faults can be categorized as mechanical or non-mechanical.
Mechanical faults are caused by defects in rolling element bearings or gears, and the signals they generate
are highly periodic. In contrast, non-mechanical signals are typically the result of under-lubrication of the
bearing (or potentially cavitation on pumps). The signals generated by under-lubrication are non-periodic –
or random – in nature. A trained analyst would employ a technique called “Autocorrelation” to distinguish
between these two types of defects. In a similar manner, PeakVue Plus Analytics applies autocorrelation
and series of sophisticated algorithms to determine the root cause of the defect on a given machine.

Autocorrelation: The statistical technique called autocorrelation is a means of identifying the amount of
periodicity in a given signal. While this value is not tremendously helpful when applied to traditional
vibration, it is extremely useful when analyzing machine impacting as measured by PeakVue technology.
The autocorrelation of a signal will typically return values between zero and one. Values around zero
indicate a high level of randomness in the signal, while values above 0.25 start to indicate periodic or
mechanical impacting.
White Paper
April 2019 Machinery Health

Figure 6 shows the correlation of 2 different signals. Both signals contain significant impacting, such that a
recommended course of action is in order. The autocorrelation values for the signal on the left hover

Figure 6 - Autocorrelation of a lubrication defect (left) and a bearing fault (right).

around zero and shows no visible pattern. This is the result that is expected when the root cause is
insufficient lubrication. Conversely, the autocorrelation values for the signal on the right show a clear
repeating pattern, and the amplitudes extend up above 0.5. This indicates a high level of periodicity as
would be expected with a rolling element bearing defect.

Combining this principle with the ability of

PeakVue to identify an abnormal condition has
resulted in the new capability referred to as
“PeakVue Plus”. This can be easily employed to
determine the root cause of the defect at the
touch of a button. This technology has been
implemented in the AMS 2140 as a new Analysis
Expert as shown in Figure 7.

To establish the severity of a defect, it is important

to first establish the turning speed of the shaft.
This is the key value that is used to determine the
appropriate limits for the PeakVue amplitudes. In
general, the higher the turning speed, the higher Figure 7 - New PeakVue Plus Analysis Expert
the alert limit.
White Paper
April 2019 Machinery Health
When launching PeakVue Plus Analytics from a Route point in the AMS 2140, the turning speed of the
route point will be used as basis for the alert limit calculation. When launching the from the Analyze
application, however, it is necessary to input or
measure the turning speed of the machine before
you can proceed with the analysis.

The algorithm uses the turning speed to determine

the alert level for the PeakVue impacting on the
machine, and then it uses autocorrelation together
with a sophisticated algorithm to automatically
identify the root cause of the impacting. These
values are combined to determine the likelihood of
a mechanical or lubrication defect on the machine.
Figure 8 shows the output of the PeakVue Plus
Analytics for a machine with a bad bearing.

Figure 8 - Results from PeakVue Plus Analytics in the AMS 2140

A simplified representation of the logic used to

determine the nature and extent of a machine defect is shown in Figure 9. In general terms, as the
amplitude of the PeakVue impacting increases, if the Periodicity is low, then it indicates a high likelihood of
a lubrication defect on the machine. Conversely, if both
the amplitude and the periodicity of the PeakVue
impacting are high, then this indicates a high likelihood
of a mechanical defect such as bearing or gear wear.

When there is a significant increase in periodicity

without a significant increase in the PeakVue impacting,
this is an early sign of mechanical wear.

For a good machine with no defects, both the PeakVue

impacting and the Periodicity of the PeakVue signals
should be very low.

By identify the nature and severity of the defect,

Figure 9 - Simplified Model for
PeakVue Plus Analytics
PeakVue Plus Analytics also prescribes the appropriate
action to address the defect. In this way, PeakVue Plus
Analytics takes the AMS 2140 beyond Predictive and Proactive Maintenance to Prescriptive Maintenance.
White Paper
April 2019 Machinery Health

Additional Considerations: As with any system for automated diagnostics, it is important to consider
additional factors to verify the result. The implementation of PeakVue Plus Analytics in the AMS 2140 offers
on-board help as follows:

Mechanical Defects: Aside from a rolling element bearing defect, other potential conditions that could be
flagged as a mechanical defect include:

 If the machine train includes a gear, check for possible gear wear.
 On variable speed equipment, ensure that the RPM is correct. An incorrect entry for the RPM of the
machine may lead to an incorrect result.
 If there is significant vibration on a neighboring machine, it may be transferred through the
foundation so that it registers at the actual measurement location on the suspect machine. This
may in certain situations be identified as a mechanical defect. Repeating the analytics on
neighboring machines prior to scheduling maintenance will eliminate this possibility. It is also an
option to repeat the analytics on the suspect machine after shutting it down for maintenance. If
the result remains the same, then the defect is actually based on the vibration from the
neighboring machine.
Lubrication Defects: Random impacting can also occur with pump cavitation. If the suspect machine is a
pump, please consider the possibility that the actual defect could be cavitation. When dealing with an
actual lubrication defect, the result will remain constant regardless of the process configuration. In
comparison, when a pump in cavitating, the result of the PeakVue Plus Analytics may change drastically
together with changes in the process.

When used as intended, PeakVue Plus Analytics can be an effective part of a Prescriptive Maintenance
program. For more information on the AMS 2140 or other Emerson products, please consult our website at
www.emerson.com.