 |
 EMR
will publish, online, reprints of technical articles by staff members.
The following article, Condition
Monitoring/Predictive Maintenance and Building Commissioning,
is the first of this series. It was originally presented at the Fourth
National Conference on Building Commissioning, April 29 through May
1, 1996, at St. Petersburg Beach, Florida.
|
|
...Condition
Monitoring/Predixtive Maintenance and
...Building
Commissioning
...By
EMR Staff
|
Synopsis
Reliability Centered Maintenance (RCM)
was developed in the airline industry in the late 1960's and has migrated
to the military, utilities, and, more recently, into industry. One
element of RCM, Condition Monitoring, is the use of inspection techniques
to determine the condition of systems and equipment, and is directly
applicable to building systems. Condition Monitoring is also known
as Predictive Maintenance (PdM). These inspection techniques have
proven to be a cost-effective means to monitor systems condition,
identify when a failure pattern has started, and predict remaining
life. In addition to monitoring the condition of in-service systems
and equipment, Condition Monitoring techniques are highly effective
in inspecting and accepting new systems. Often, hidden defects can
be identified and documented, providing irrefutable evidence of required
rework before the "warranty" period ends.
|
Contents
|
What
is Condition Monitoring/Predictive Maintenance?
Simply stated, Condition Monitoring,
also known as Predictive Maintenance, is the continuous or periodic
monitoring and diagnosis of equipment and components in order to forecast
failures. Typical Condition Monitoring technologies and approaches
include flow and process analysis, vibration analysis, oil analysis,
infrared thermography, ultrasonic noise detection, and electrical
circuit analysis.
Condition Monitoring has been emerging
within the last few years as more people have become aware of its
benefits and as the costs associated with monitoring the condition
of systems and equipment have decreased. Prior to the 1950''s, there
was little discussion regarding the maintenance of systems and equipment.
Most maintenance was intuitive or break down (reactive) and there
was little examination of the relationship between failures and maintenance.
In the 1960's, the airline industry set out to improve the effectiveness
of maintenance in order to increase aircraft availability, reduce
costs, and increase reliability, without sacrificing safety. The book,
Reliability-Centered Maintenance by Stanley Nowlan and Howard Heap,
was the first detailed discussion on the subject and is the basis
for modern Reliability Centered Maintenance (RCM) programs. A key
element of RCM is the understanding that time-based maintenance is
sometimes not the most effective. Time-based maintenance may introduce
problems into otherwise healthy machines or, in extremes, result in
premature overhaul or replacement. Based upon this understanding,
it became apparent that, when possible, a time-based inspection of
systems and equipment would result in more effective utilization of
maintenance resources. In the 80's, advances in technology, and proven
results in the aerospace, military, utilities, and process industries,
raised the awareness of time-based inspections, which have become
known as Condition Monitoring.
For the remainder of this paper I will
use the term Condition Monitoring to describe the process of inspection,
monitoring and diagnosis of equipment condition. I prefer Condition
Monitoring, instead of Predictive Maintenance, because it best describes
what is being done. Predictive implies positive forecasting of machinery
condition. This simply is not always the case for individual machines.
However Condition Monitoring will improve the overall health of machines,
including extending life and diagnosing most impending failures. There
are other names for the approach in various industry or agencies;
for example, NASA uses the term Predictive Testing & Inspection
(PT&I).
|
System/Machine
Failure
In order to understand how to develop
effective maintenance, the airline industry examined all aspects of
maintenance, including when in their life-cycle the systems failed.
They concluded that there are six age-reliability patterns associated
with machines. Of interest to owner of a new system- approximately
78% of the systems exhibited an age-reliability pattern known as infant
mortality. In other words, there was a high probability of system
failure immediately after manufacture or overhaul. We confirm this
conclusion on a routine basis in our day-to-day contact with consumer
products. It therefore stands to reason that we want to identify damaged,
degraded and failing systems before a catastrophic failure or expiration
of the supplier s warranty. More interestingly, there was a relatively
constant probability of failure with age, for approximately 89% of
the systems. In other words there was little or no wear-out region
directly related to operating age. In those cases, an interval-based
replacement or overhaul will have little effect on reducing the probability
of failure, and will, in fact, increase the probability of failure
due to infant mortality. Follow-on studies in Sweden in 1973 and by
the U.S. Navy in 1983 produced similar results.
A failure is defined as a loss of function.
A fan may still be turning, but if it is not performing within the system
specification, it has failed. Often times, especially on the component
level, functional failures must be sought out. This is because the system
or machine appears to be operating satisfactorily even though some components
have functionally failed. In other cases, the function is only performed
at intermittent times and failure will have to be uncovered through
testing. This is often the case for safety devices, such as electric
power circuit breakers.
|
Condition Monitoring Technologies
Key to any effective maintenance program
is economics. At what cost will you maintain function? Areas that
must be considered are effects on safety, operation (reason for being),
and capital costs. When all items are considered, Condition Monitoring
often offers a cost effective means of forecasting, and therefore
preventing, failure. The other maintenance approaches are Run-To-Failure
and Interval Based Maintenance (the traditional Preventive Maintenance).
Both of these are appropriate choices in some circumstances.
The Condition Monitoring technologies
discussed in this paper are only some of the approaches used in various
industry. You must ask yourself if you could improve the reliability
of a system through the use of Condition Monitoring, and then determine
if there is a means to performing the monitoring.
When we think of Condition Monitoring,
we usually think of vibration analysis. The technology and techniques
have been developing for over 30 years and over 78% of all manufacturing
or processing plants use vibration analysis. Vibration analysis
of rotating machines such as motors, pumps, fans, and gears is widely
accepted as a viable technique to identify changing conditions.
Reduced costs of test equipment and data management (primarily computers),
availability of training, and development of computer-based expert
systems are all contributing to this acceptance.
The technique measures machinery movement
(vibration), typically through the use of an accelerometer, and
examines the vibration spectrum to identify and trend frequencies
of interest. Some frequencies are associated with the machine design,
regardless of its condition. For example, a healthy fan or rotary
compressor may have a frequency that is equal to the machine speed
times the number of fan blades. The vibration analysts may monitor
this frequency to note changes in the amplitude indicating a degrading
condition. Other frequencies, for example, those associated with
rolling element bearings, may be a sign of bearing damage and will
alert the analysts to the start of bearing failure. It is common
for electric motor problems, such as broken rotor bars or stator
eccentricity, to be seen in vibration associated with electrical
line frequency. In new equipment, vibration analysis can identify
defective bearings and confirm proper alignment and balance at installation.
The vibration data is usually collected
with a portable device for periodic monitoring, or a continuous
monitoring system may be installed for costly or critical systems.
Analysis of the vibration data requires a detailed understanding
of machinery operations and of vibration analysis techniques. Vibration
data is usually collected and analyzed on a monthly to quarterly
basis (on continuously-running equipment). Costs vary due to machinery
locations (the more spread out, the higher the cost) but typically
are $30 to $50/machine/inspection.
Infrared Thermography (IRT) is the
application of infrared detection instruments to identify pictures
of temperature differences (thermogram). The test instruments used
are non-contact, line-of-sight, thermal measurement and imaging
systems. Because IRT is a non-contact technique, it is especially
attractive for identifying hot/cold spots in energized electrical
equipment, large surface areas such as roofs and building walls,
and other areas where "stand off" temperature measurement
is necessary.
IRT inspections are identified as
either qualitative or quantitative. The quantitative inspection
is interested in the accurate measurement of the temperature of
the item of interest. The qualitative inspection identifies relative
differences, hot and cold spots, and deviations from normal or expected
temperature ranges. Qualitative inspections are significantly less
time-consuming than quantitative because the thermographer is not
concerned with highly-accurate temperature measurement. What the
thermographer does identify is highly accurate temperature differences
( 'D T) between like components. For example, a typical motor control
center will supply three-phase power, through a circuit breaker
and controller, to a motor. Current flow through the three-phase
circuit should be uniform, which means that the components within
the circuit should have similar temperatures, one to the other.
Any uneven heating (perhaps due to dirty or loose connections) would
quickly be identified with the IRT imaging system.
IRT can be utilized to identify degrading
conditions in electrical systems such as transformers, motor control
centers, switchgear, switchyards, or power lines. In mechanical
systems, IRT can identify blocked flow conditions in heat exchanges,
condensers, transformer cooling radiators, and pipes. It can also
be used to verify fluid level in large containers, such as fuel
storage tanks, and identify degraded refractory in boilers and furnaces.
Inspections are performed on an annual or biennial basis and cost
about $1,000/day. Roof inspections, useful in identifying water
penetration of the roof membrane, has special conditions (evening
hours following a sunny day, low wind, no condensation) to be effective
and can require several days to accomplish.
Lubricating oil analysis is performed
for three reasons: to determine the machine mechanical wear condition,
to determine the lubricant condition, and to determine if the lubricant
has become contaminated. There are a wide variety of tests that
will provide information regarding one or more of these areas. The
test used will depend on the test results sensitivity and accuracy,
the cost, and the machine construction and application. Note that
the three areas are not unrelated, as changes in lubricant condition
and contamination, if not corrected, will lead to machine wear.
Because of the important relationships, commercial analysis laboratories
will often group several tests in cost-effective "packages"
that provide information about all three areas. Cost vary, but a
typical analysis package is about $30/sample. Oil should be analyzed
on a quarterly to semiannual basis.
Oil analysis is performed on in-service
machines to monitor and trend emerging conditions, confirm problems
identified through other means such as vibration, and to troubleshoot
known problems. Lube condition trending, such as depletion of additives,
can identify when the oil should be changed. Material, such as metal
or seal particles, can identify machine damage before catastrophic
failure, allowing for less costly repair. In new buildings, oil
analysis will confirm that the specified lubricants are being used
and that the system is free of construction contamination.
A relatively inexpensive device called
an ultrasonic noise detector can be used to locate liquid and gas
(pressure and vacuum) leaks. When a fluid moves from a high pressure
region to a low pressure region it produces ultrasonic noise due
to turbulent flow. The detector translates the ultrasonic noise
to the audible range, allowing an inspector to identify the source
of the leak. In-service fluid systems, such as compressed air and
steam, are usually inspected semiannually or annually. The period
of time between inspections is usually adjusted depending on the
fluid system and the number of leaks routinely found.
In addition, an ultrasonic noise detector
can detect arcing, tracking, and corona in electrical systems. Often
the detector is used in conjunction with the IRT inspection.
-
Electrical
Circuit Analysis
Electrical testing is used to measure
the complex impedance and insulation resistance of electrical conductors,
starters, and motors. It detects faults, such as broken windings,
broken motor rotor bars, voltage imbalances, and cable faults. Most
motor failures are bearing-related; however, a close second is stator-related,
due to overheating and insulation breakdown. It is estimated that
voltage imbalances as small as 2% to 3% will increase the temperature
rise in the motor 20% to 25%. It follows then that the conditions
to monitor on electric motors should include the motor insulation
and the voltage balance. Techniques include megohmmeter testing,
conductor complex impedance, and motor current spectrum analysis.
For new motors only, a high potential test (high-pot) and surge
test can be performed. These tests can be destructive and are not
used on in-service motors or circuits.
-
Flow
and Process Analysis
Flow and process analysis often will
provide insight into the overall condition of a complex system.
For example, changing fluid flow rates and pump discharge pressures
can indicate pump degradation. By monitoring the condition, the
pump can be overhauled based on the condition as opposed to a set
interval. Differential temperature can be measured in heat exchangers
to determine heat transfer coefficients and to indicate possible
fouling of tubes.
Trend Analysis: This technique typically uses two-dimensional graphs
in order to visually display trends. The graph may have alert or
alarm levels to aid in the analysis.
Pattern Recognition: Often
times machines exhibit recognizable operation patterns. Deviations
from the pattern or norm are indications of changes that may identify
the onset of failure. For example, the infrared thermography inspections
discussed earlier are looking for unexpected thermal patterns.
Test Against Limits or
Ranges: For parameters
or conditions that do not follow continuous trends or repeatable
patterns. Useful in instrument calibration.
Relative Comparison of
Data: Look for
change as related to earlier data or from another baseline (such
as similar equipment). Need stable building/system conditions.
Statistical Process Analysis
(Also called Parameter Control monitoring): Generally uses process or maintenance data that already
exists or is collected. Applies statistical techniques to process
or maintenance data looking for deviation from the norm. Correlation
Analysis: The most powerful technique is the one that uses
data from multiple sources, related technologies, or different analysts.
|
References
- National Aeronautics and Space Administration,
NASA Facilities Maintenance and Energy Management Handbook (NHB
8831.2A), Washington, DC.
- Nowlan, F. and H. Heap, 1978, Reliability-Centered
Maintenance. Report AD/AO66 579, available from National Technical
Information Service
- Nicholas, J., 1994, Practical Electrical
Predictive Test and Inspection. PdMA Corporation, Tampa, FL
- Toms, L., Machinery Oil Analysis,
1995, Pensacola, FL
|
Copyright © 2003 Enviro-Management & Research, Inc.
5500 Cherokee Avenue, Suite 200, Alexandria, VA 22312
Tel. 703/766-2670 Fax 703/766-2674
E-mail Webmaster at getinfo@emrinc.com
|
