Professional Documents
Culture Documents
Condition Monitoring Program
Condition Monitoring Program
Condition Monitoring Program
The common theme for all these definitions is that condition monitoring is
solely a maintenance tool. Because of these misconceptions, the majority of
established programs have not been able to achieve a marked decrease in
maintenance, cost of repairs, or a measurable improvement in overall plant
performance.
1
We must remember that the output of a condition- monitoring program is
data. Until this data is communicated and acted upon to resolve the problems
revealed by the program, plant performance cannot be improve. What this
means is that a commitment from senior management to ensure full co-
corporation from all the plant areas is imperative for the program to succeed.
Why then, do so many condition- monitoring programs fail? The fact is that
predictive technology is not adopted as a core operating philosophy by
senior management of an organization and only becomes the pet project of
one or more local engineers.
2
The condition-monitoring program can be invaluable to production
management by the information it can provide in being able to increase
production capacity, product quality, and overall effectiveness of the
production function. Remember condition-monitoring can provide the data
needed to achieve optimum and consistent reliability.
3
As we have noted throughout this section senior management support is
crucial for the success of any condition-monitoring program.
Once a commitment from management has been gained the first step would
be to determine what machines will be monitored and what monitoring
techniques will be employed to monitor these machines. Below is an
overview of typical monitoring techniques.
4
In determining what machines to monitor generally start small rather than
going plant wide, in essence “get the runs on the board”. Starting small
minimizes the drain on resources, especially in the beginning where data
collection techniques and procedures are being fine tuned. The initial list of
machines may have a higher critical rating or machines that have chronic
problems. Remember the term critical may relate to machinery that would
limit production capability, as well as equipment with high initial or
replacement cost. Using a small number of critical machines with chronic or
serious problems are most likely to show immediate benefits. These benefits
can be used to build support during the program implementation.
We must remember that we have not limited the size of the machine list, but
rather put together a sizeable list that will show the benefit condition
monitoring quickly without tying up valuable resources. Once the program
has been established to increase the database size will be relatively easy.
Once the initial machine list has been established the appropriate monitoring
Technique is applied. Arguably the simplest and most cost-effective method
of condition monitoring is the visual inspection. The main advantage of
visual inspection is that it is cheap, and can be carried out by any operator.
No matter what type of conditioning monitoring method, the use of all
human senses should be an integral part of any program. The negative side
of using visual inspections is the subjective nature of the observations.
5
Spectral analysis: uses the Fast Fourier Transform (FFT) to convert the raw
signal (waveform) into its individual frequencies and positive amplitudes.
The FFT uses a long mathematical algorithm for the calculations. Spectral
has the advantage of being able to pin point faults using the these individual
frequencies (bearing, gear, structural, and process faults). Spectral analysis
involves the use of software and hardware. The software is used to build the
database as well as store and analyze the data. The hardware consists of a
data collector, which after the template (route) has been downloaded into the
data collector, it is used to collect the data as per the machine and point
description in the route. Once the data collection has been completed, the
data is then uploaded to the software for analysis and reporting.
Oil analysis and wear debris analysis: in contrast, are restricted to circulatory
arrangements such as lubrication or hydraulic power systems. One of the
major financial benefits of oil analysis is monitoring of the physical
condition of the oil, this allows a condition-based oil changing policy to be
implemented according to additive depletion, contaminate build up and/or
other changes in its physical state.
6
One of advantages of wear debris monitoring over vibration monitoring is
that it can provide a much clearer indication of the source of a fault.
Whereas vibration caused by a fault can sometimes be masked by vibration
emanating from other components, wear debris analysis will only show
excessive amounts of a particular material when there is a fault present.
Once we have determined our initial machine list and type of condition
monitoring technique that will be implemented, the appropriate monitoring
frequency must be determined. In some part the frequency will be dictated
by previous failures or more precise, the time between failures, criticality of
the machine, duty cycle of the equipment and actual speed of the rotating
components.
Previous failures and the time between similar failures will help determine
the monitoring frequency. There is a number of studies into machine failure
modes and the interval between failure starting and actual failure.
One such method the ‘P-F interval’ uses the potential failure and functional
failure to determine the lead time to failure.
7
The P-F interval is also known as the warning period or the lead time to
failure. The frequency of checking must be significantly less than the P-F
interval if we to detect the potential failure before it becomes a functional
failure. Determining the P-F interval for age-related failure modes is fairly
as simple as the final stages of deterioration are linear. However, in the case
of random failures, we don’t know when the next failure mode is going to
start on its way down the P-F curve. So we if don’t even know where the P-F
curve is going to start, how can we go about finding out how long it is?
The most accurate and cost effective way is based on judgment and
experience. Utilizing personnel who have knowledge in the machinery type,
previous experience in monitoring particular type of equipment and the
operator/trades people who maintain the equipment will be able to establish
a realistic P-F interval. Remember we are trying to determine how quickly
the item fails rather than how long it lasts. The actual speed of the rotating
equipment will be factor in determining the survey schedules. Specifically in
slower machines, it would in most cases extend the P-F interval and thus
allow a larger duration in time between survey schedules.
When determining the monitoring frequency we must above all factor the
account, not one factor alone can dictate the frequency. The last factor that
8
we must consider, and one that may over ride all the above resources. If the
resources are hard to come by, which seems to be problem in industry at the
moment, attempting to monitor a complete plant on long monitoring
schedules may be worthless when compared to monitoring a smaller amount
of machines using more realistic schedule, and actually picking faults before
failure.
Experience shows that whatever the limits that are set initially these will
need to be tailored further down the track. The main point here is not to
become
overly worried on setting accurate alert and alarm limits. Although
manufacturers will give tolerance limits, these will normally be excessive
when compared to the actual machine vibration resulting in miss of critical
fault. On the other hand many new comers set limits to conservatively
resulting in the flood of maintenance work orders and causing frustration
when no faults are found. Purely relying on alert and alarm limits can be
dangerous and can even lead to the failure of a program. Machines reported
with faults and no faults found is as detrimental to the program as faults
missed, both will incur a cost and depending on the repair may cost more
than a run to failure. Unfortunately that the risks we must factor in, they can
however be minimized through training, familiarization with the equipment,
external support from experts and comparing like equipment.
There are number of severity tables published using machine sized and
running speed to determine acceptable and fault limits. These are ideal as a
basis to start with, however they should not be relied on 100%. Ideally the
start of a condition monitoring program should have the survey frequency,
set frequently say no more than monthly, after the first three months the
frequency is re-evaluated and the alert and alarm limits can have their first
pass of tailoring. Tailoring should then occur at the twelve-month mark after
9
there is sufficient data. Tailoring of the alert and alarm limits should not
become stagnate rather an ongoing process in pursuit too developing an
efficient and reliable database.
At this stage we have the program up and running, data is being collected
and analyzed and faults are reported and repaired. What is crucial during this
first year is the feedback from the operators, maintenance personnel and
contractors. For the program to grow and mature, the feedback from these
personnel regarding repairs, failures, modifications and process alterations is
essential to correlate the data from the program to what is actually
happening out in the field. Faults found through the condition monitoring
program should be followed through and correlated with the actual repairs
done.
Writing the case histories and getting them circulated throughout the
company are perhaps the most important, and yet underused tools available.
These perform three essential functions:
10
The first year will be the foundation of the following years to come. Time
invested in training personnel, ensuring adequate resources, setting
procedures and ensuring the continuation of management commitment will
help the program succeed. To many programs fail in the first year due to
these four factors not to be addressed resulting in the equipment being
mothballed.
It is important to look forward past the next 1-2 year window, to the next 3-5
years. Emphasis need to be put on succession planning, as manning the
group will continue to be an ongoing issue. Using specialists or trained
personnel will keep the Condition Monitoring Team fixed cost low and will
allow key activities to be completed. There is a need to market our results to
show the Condition Monitoring function will and will still adds value.
Conclusion
Having a clear purpose and vision by which the priorities of activities can be
set, is a way that the Condition Monitoring function can still move forward.
In a time where downsizing the work force has become necessary for
companies to be viable, condition monitoring should be on the lips of most
senior managers, in an effort to minimize down time and increase plant
efficiency, thereby increasing plant productivity.
11
TYPICAL MONITORING TECHNIQUE
12