“Predictive Maintenance (PdM)” or “Preventive Maintenance (PM)”
The proponents of PM and PdM have
been debating for long over which is more beneficial and effective at reduction
in maintenance costs while fulfilling the main goal of maintaining equipment
reliability. Let us look at some of the advantages:
Reduction in unplanned outages
Reduction in failure rates
Saving in maintenance costs (mainly outage
These benefits are not marketing
jargons but has time and again been proved and several statistics are already
available over different surveys conducted by eminent bodies (eg. CIGRE, IEEE
etc.). However, theyseem to accrue from both types of maintenance.
Predictive Maintenance (PdM) is generally taken to mean “Assessing
machine condition through acquiring/analysing data on key parameters, during
its operating condition and anticipating a failure”. So, the key in above
definition is “during its operating condition” and most of the time this is to
mean “Online” assessment while the equipment is running, and you will find several
articles on the Net, where PdM is mostly about vibration and noise,
thermography and possibly extending into lube oil analysis.
Primarily this is the most
“visible” area in PdM which the maintenance guys can relate to. Alignment with
load, soft foot corrections, bearing replacements are most common maintenance
activities and hence vibration and noise are well understood. So, many plants
have their own in-house teams for such PdM programs (vibration analysis). These
teams conduct measurements on a periodic basis and survey a wide range of
mechanical equipment. Also, now with better development in wireless and
internet technology/IoT, most industries either have upgraded or are
contemplating upgradation to a continuous monitoring system with mobile alarms/alerts.
So, the role of PdM then is to
highlight machines or components that show a defect that can manifest into a
failure. But PdM by itself will not help with above benefits. It is just an
observation and a recommendation! In most scenarios, unfortunately, these
recommendations occupy the shelves in Managers cabin. In good scenario, the
recommendations are used to determine a PM schedule. In best scenario, these go
into AMPs (Asset Management Program) or CMMS (Computerized Maintenance
Management System)where there are deeper learnings, knowledge sharing about
equipment or the systemand is linked
to business objectives, spare procurement /inventory management, generating
purchase requirements/work orders etc. Importantly, they can quantify all these
efforts into monetary value.
What is ignored is that PdM can
also tell you which machines are in good condition and does not merit any
service. What does this mean? It means “Money saved”. But ironically many
maintenance managers fail to realize the importance of a report that says,
“Your machine is absolutely in good condition”. They feel sceptical - all that
money spent on PdM has been down the drain – the results should have pointed
out some defect – ain’t that an objective of PdM?
It’s like going to a doctor
saying, “I kind of feel low, give me some medicine” and the doctor inspects you
(even takes a good fee) and says, “There’s nothing wrong with you. Just go home
and sleep and you are fit”. So, the mind says, “What! If I am good, why did I
pay that fee for?”. It’s time to rejoice! Your machine is in good condition.
Not only you defer its maintenance and save on cost, you can use that budget
for other priorities. You can showcase “money saved” by not taking an outage or
not having failures. You can highlight that your maintenance staff is very
effective and deserve a bonus! Ok if not bonus, at least a pat on the back.
Preventive Maintenance (PM) is performed while the
equipment is under normal operation to avoid unexpected breakdowns and the
associated downtime and costs. During the specified date and time, equipment is
shut down and PM tasks on that piece of equipment are performed. Maintenance is
set on a schedule based on calendar dates or usage, often at the recommendation
of the manufacturer.
So, the important aspect is the
warranty issues, spare or service support issues with manufacturer or
legalities of the insurance in the sad circumstances of failure backed up by
proof that PM was prescribed by the manual, but was not conducted in the given
schedule – blasphemy! So, if the bearing needs to be changed after x hours,
change it. The new bearing will definitely outlast the old one, which may be in
good condition presently. This may be a good rationale and justified too, but
definitely, it is a factor that makes PM a costlier proposition.
Also, PM advocates claim that
“Condition Monitoring is all gimmick and fancy” or “Flavour of the month”. These
sentiments are often nurtured as some of the technologies are far too advanced
and complex –Fourier Transform, Demodulation, Neural networks, fuzzy logic, AI,
Machine Learning. Wow! The reports, however graphicalthey may be, are beyond
the realms of the person who changes bearing or does alignment. Many
maintenance personnel argue, “What is the use of incurring additional costs in
PdM when the recommendations will be to change the bearings or overhaul the
motor. We were going to do the same anyway”. How many of you just said
PdM advocates say that it has an
edge over PM as the costs are much lower than PM while the benefits are further
fine-tuned. For e.g. If the condition of component or asset is good, it need
not be repaired/replaced, resulting in more savings. “If ain’t broken, why fix
it?”. The other common sentiment is “Machine was running nicely but now after
assembly, something has been messed up and machine has higher
vibrations/noise”. Now how many of you said “Bingo!”
But on a happy note, everyone
does agree that “Breakdown Maintenance” is a strict NO, except for small value
assets where replacement is much easier and less costly option than
When it comes to electrical equipment
maintenance – say a motor, generator or transformer, PdM with only
vibration/noise scope is only a part of the story. Consider a scenario, where
the main motor fails after having all the relevant processes, tools, manpower
(read that as investment in PdM) in place. To the top management, the failure
has resulted in loss of production plus losses in present batch – sudden
unexpected outflow of cash, lost revenue and profitability - even after
allocation of maintenance budget. To the credit of Maintenance Manager – he had
conducted the measurements and analysis at correct schedules and there was no
cause of alarm from the reports. On further investigation, one sees a Stator
bar failure! So obviously, this may not reflect in vibration signals.
So, would a PM help in this case
For such insulation diagnostics,
PdM is available in form of conventional IR-PI checks, which if done as spot
measurements are not very effective. Tan Delta & Capacitance measurements
with tip-up calculations, partial discharge measurements, surge comparison,
balance of resistance/impedance, DC HV Step voltage tests, Charge-Discharge
Current Profiling (better than IR/PI) are some tests that can give effective
assessment as a package of tests and it is not prudent to skip one due to
reasons like lack of budget.
But these tests are mainly outage
based which further intensifies the debate.
Is it wise to plan an outage for
doing these tests and later another planned outage for PM? For plants that have
standby motors, it is possible to do a PdM and maintain a tab, especially as
condition of machine can be affected even when it is not operational.
Combining the power of PdM and PM
I guess is the answer and a
change of perspective. If PdM, by itself, cannot increase reliability, why is
it considered as a separate maintenance approach? Shouldn’t PdM be considered
as an integral part of or as a subset of PM?
This works wonders for a plant
that has several assets but limited outage times and budget per year. So, while
a PM activity must be scheduled, do it around the first time as per the
manufacturers recommendation. Also do the offline PdM assessments like test
mentioned earlier or say a visual inspectionin same outage. Conduct the online
assessment before and/or after the outage. Consolidate the results to determine
the frequency of next PdM schedule and base the next PM schedule based on the
condition of equipment or the component, instead of time.
So, if a motor is say “clean” and
bearing condition is “good” (as confirmed from reports) defer the PM activity
for that motor to later than that indicated by time based maintenance.
Conversely, if the results indicate any anomaly, pre-pone the PM activities.
The other advantage of doing the
PdM with PM i.e. say doing insulation diagnostics in offline condition after
the cleaning activity, can have multiple benefits too.
It will help assess if the PM activities were
done effectively. E.g. Say a drop in PD levels can help assess that the
overhang clean-up was effective.
Any defect shrouded by previous condition can be
revealed during PM or with PdM measurements. For e.g.
Often ageing conditions is difficult to diagnose for a contaminated machine.
So, after removal of contaminants, the measurements can help in correct
assessment of insulation degradation.
The reverification by PdM may be several times
less costly than opening the motor for cleaning only to find that it is clean.
So, while the PM activity saves
cost by way of reducing failure, the PdM integrated with PM further fine tunes
or optimises the maintenance costs by focusing only on machines or components
that need attention.
Thus, PdM and PM are not rivals but pitched together can make a winning