1

PREDICTIVE MAINTENANCE: RELIABILITY AND QUALITY.

INTRODUCTION :

In recent years, predictive maintenance management has been widely discussed.

defined a variety of techniques ranging from vibration monitoring to imaging in
infrared. Predictive maintenance has been recognized as an effective technique for
maintenance management.
Other terminologies have emerged as maintenance management tools, these new
terms - RCM, reliability-centered maintenance; TPM, total productive maintenance; and JIT,
Just-in-Time maintenance - is presented as a substitute for predictive maintenance and the solution
definitive due to its high maintenance costs.
This article aims to explain the basic knowledge necessary for selection and
implementation of a comprehensive and cost-effective maintenance management program
in your factory.
Since most manufacturing and processing plants rely on equipment
mechanics for most of their processes, vibration-based predictive maintenance is the
dominant technique used for most maintenance management programs. However, the
ability to monitor all critical machines, equipment, and systems in a plant
typical industrial cannot be limited to a single technique.
The monitoring techniques in predictive maintenance, that is, condition-based, include: analysis
vibration, ultrasound, ferrography, tribology, process monitoring, visual inspection, and others
non-destructive analysis techniques. The combination of these monitoring and analysis techniques
it offers the means for direct monitoring of all critical equipment and systems in its
factory.
Maintenance costs account for the main part of the total operating costs of
all manufacturing and production industrial plants. Depending on the specific industry, the
maintenance costs can represent between 15% to 30% of the cost of goods produced. By
example, in food industries, average maintenance costs can represent about
15% of the cost of goods produced; while in the steel, paper, and cellulose industries,
In other heavy industries, maintenance can represent up to 30% of total production costs.
Recent research on the effectiveness of maintenance management indicates that one third of all
the maintenance costs are wasted as a result of unnecessary maintenance or
inadequately performed. When you consider that the American Industry spends more than 200
billions of dollars every year on factory equipment and facilities maintenance, the impact
about the productivity and profit represented by the maintenance operation becomes clear.
The result of ineffective maintenance management represents a loss of over 60 billion.
of dollars every year. Perhaps more importantly is the fact that our ineffective management of maintenance
it has a dramatic impact on our ability to manufacture quality products that are
competitive in the global market. The loss of production time and product quality, which
The result of inadequate maintenance management has had a dramatic impact on our condition.
to compete with Japan and other countries that have implemented more advanced philosophies of
manufacturing and maintenance management.
The dominant reason for this ineffective management is the lack of factual data that quantifies the
real need for repair or maintenance of machinery, equipment, and plant systems
industrial. The maintenance schedule has been, and in many cases is, planned based on data from
statistical trend or in the actual failure of equipment in the industrial plant.
 2
Until recently, middle and corporate management had ignored the impact of
operation of maintenance on product quality, production costs and, more importantly, in
basic profit. The general opinion has been that 'Maintenance is a necessary evil', or 'Nothing can'
to be made to improve maintenance costs." Perhaps these were true statements.
or 20 years ago.
Meanwhile, the development of the microprocessor and other instruments based on
computers used to monitor the operating condition of factory equipment, machinery, and
systems have provided means to manage maintenance operations. They have enabled the
people reducing or eliminating unnecessary repairs, avoiding catastrophic machine failures, and reducing
the negative impact of maintenance operations on the performance of the industrial plant
manufacturing and production.
To understand predictive maintenance management programs, one must consider
first the traditional management techniques. Industrial and process plants typically use
two types of maintenance management: corrective maintenance (run until failure) or maintenance
preventive.

CORRECTIVE MAINTENANCE:

The logic of management in corrective maintenance is simple and straightforward: when a machine
break it, fix it. This method ("If it isn't broken, don't fix it") of maintenance of
Industrial machinery has represented a large part of the plant's maintenance operations.
industrial, since the first factory was built and, on top of that, it seems reasonable. A plant
industrial using management for corrective maintenance does not spend any money on maintenance,
until a machine or system fails to operate.
Corrective maintenance is a reactive management technique that waits for machine failure.
or equipment, before any maintenance action is taken. It is also the most
maintenance management caro.
Few industrial plants use a true maintenance management philosophy.
corrective. In almost all cases, industrial plants perform basic preventive tasks, such as
lubrication and adjustments of the machine, even in a corrective maintenance environment. However,
in this type of management, the machines and other equipment of the industrial plant are not inspected and
No major repairs are made until the equipment fails in its operation.
The greatest costs associated with this type of maintenance management are: high costs of
spare parts stocks, high extra labor costs, increased downtime of
machine, and low production availability.
Since there is no attempt to anticipate the maintenance requirements, a plant
industries that use absolute corrective maintenance management must be able to react to all
possible failures within the factory. This reactive management method forces the department to
maintenance to keep expensive spare parts stocks that include backup machines or, by
less, all the main components for all critical equipment in the factory. The
the alternative is to rely on equipment vendors that can provide immediate delivery of
all requested spare parts.
Even if the last one is possible, the rewards for prompt delivery increase.
substantially the costs of repairing parts and the downtime needed to fix the
machine failures. To minimize the impact on production caused by unexpected failures of the
machines, the maintenance staff must also be prepared to respond immediately to all failures
of the machine. The net result of this reactive type of maintenance management is a higher cost of
maintenance and lower availability of process machinery. The analysis of maintenance costs
It indicates that a repair carried out in corrective-reactive mode will have an average cost of about 3 times.
greater than when the same repair is done within a scheduled or preventive mode. The
 3
Repair scheduling ensures the ability to minimize repair time and associated costs.
of labor. It also ensures the means to reduce the negative impact of expedited remittances and
lost production.

PREVENTIVE MAINTENANCE:

There are many definitions of preventive maintenance. However, all programs of

preventive maintenance management is triggered by time. In other words, the tasks of
Maintenance is based on time spent or operational hours. The well-known average time curve
for failure (CTMF) or the "bathtub," indicates that a new machine has a high probability of
failure, due to installation problems, during the first weeks of operation. After this
In the initial period, the probability of failure is relatively low for an extended period of time.
After this normal operating period of the machine, the probability of failure increases sharply with the
elapsed time. In the management of preventive maintenance, repairs or refurbishments of
machines are programmed based on CTMF statistics.
The implementation of real preventive maintenance varies quite a bit. Some programs are
extremely limited and consist of lubrication and minor adjustments. The programs more
comprehensive preventive maintenance schedules repairs, lubrication, adjustments, and
Reconditioning of machines for all critical machinery in the industrial plant. The denominator
common to all these preventive maintenance programs is the planning of maintenance x
time.
All preventive maintenance management programs assume that the machines
they will degrade with a timeline typical of their particular classification. For example, a
horizontal, single-stage centrifugal pump will typically run for 18 months before it needs to
to be overhauled. Using preventive management techniques, the pump would be taken out of service and overhauled
after 17 months of operation.
The problem with this approach is that the mode of operation and specific variables of the plant
industrial or system failures directly affect the normal operational life of the machinery. The time
the mean time between failures (MTBF) will not be the same for a pump that is working with water and
a pumping abrasive pulp of ore. The normal result of using the TMF statistic for
schedule maintenance or it is an unnecessary repair or a catastrophic failure. In the example, the
the pump may not need to be refurbished after 17 months. Therefore, the labor and materials
the materials used for the repair were wasted. The second scenario of preventive maintenance is
even more expensive. If the pump fails before 17 months, we are forced to repair it using techniques
corrective. The analysis of maintenance costs has shown that a repair done in a way
Reactive (that is, after the failure) will typically be three times more expensive than the same repair done
on a programmed basis, for the reasons mentioned earlier.
The old adage that machines will break down at the worst possible time is a very
real of industrial plant maintenance. Typically, the breakdown will occur when the demands of
production were the highest. The maintenance team must then respond to the unexpected failure. In this
in reactive maintenance mode, the machine is disassembled and inspected to determine the repairs
specifics required to return it to service. If the repair parts are not in stock,
they must be ordered at market costs, and expedited shipping should be requested.
Even when the repair parts are already in stock at the plant, the lead time
of work for repair and the cost are much higher in this type of reactive maintenance. The staff of
maintenance must disassemble the entire machine to locate the source of the problem or problems that
they forced the failure. Assuming they correctly identify the problem, the time required to
dismantling, repairing, and reassembling the machine would have been, at least, greater than what would have been required by
a planned repair.
In predictive maintenance programs, the specific mode of failure (that is, the problem)
can be identified before the failure. Therefore, the correct parts for repair, tools, and
 4
workforce skills may be available to fix the machine problem before the
occurrence of catastrophic failure.
Perhaps the most important difference between reactive and predictive maintenance is the ability to
to schedule the repair when it will have the least impact on production. The production time
lost as a result of reactive maintenance is substantial and rarely can be recovered. The
Most industrial plants operate 24 hours a day during peak production periods.
Therefore, the lost production time cannot be recovered.

Predictive Maintenance :

Like preventive maintenance, predictive maintenance has many definitions. For the
mechanics, predictive maintenance monitors the vibration of rotating machinery in an attempt to
detect emerging problems and prevent catastrophic failure. For electricians, it is the monitoring of
infrared images of circuits, electrical switches, motors, and other electrical equipment
to detect problems in development.
A common premise of predictive maintenance is that regular monitoring of the condition
real mechanics, operational efficiency, and other indicators of the operational condition of the machines and
process systems will provide the necessary data to ensure the maximum interval between the
repairs. It would also minimize the number and costs of unscheduled downtime caused by
machine failures.
Predictive maintenance is much more. It is a means of improving productivity.
the quality of the product, the profit, and the overall effectiveness of our manufacturing plants
of production. Predictive maintenance is not merely vibration monitoring or analysis of
lubricating oil or thermal imaging or any of the other non-destructive testing techniques that
have been marked as predictive maintenance tools. Predictive maintenance is a
philosophy or attitude that uses the actual operational condition of the equipment and systems of the plant
industrial to optimize the overall operation of the industrial plant. A comprehensive program of
predictive maintenance management uses a combination of the most cost-effective tools
to obtain the real operational condition of critical systems of the industrial plant and, based on these
real data, all maintenance activities are scheduled on a certain basis 'as per
necessary.
Predictive maintenance is a preventive maintenance program triggered by conditions.
Instead of relying on average lifespan statistics in the industrial plant or manufacturing (e.g., time
(medium for failure) to schedule maintenance activities, predictive maintenance uses
direct monitoring of mechanical conditions, system performance, and other indicators for
determine the average time to real failure or loss of performance for each machine and system in
industrial plant. At best, traditional time-driven methods ensure
a guide to 'normal' machine life spans.
In preventive or corrective programs, the final decision about the repair or
Reconditioning is based on the intuition and personal experience of the maintenance manager. The addition
a comprehensive predictive management program can provide data on mechanical condition
real of each machine and the operational efficiency of each process system. This data
will enable the maintenance manager to schedule maintenance activities much more effectively
in terms of cost.
A predictive maintenance program can minimize the number of breakdowns of all
mechanical equipment of the industrial plant and ensure that the repaired equipment is in
acceptable mechanical conditions. It can identify machine problems before they become
serious since most mechanical problems can be minimized if they are detected and
repaired in advance. The normal modes of mechanical failure degrade at a rate
 5
directly proportional to its severity; therefore, when a problem is detected early,
Usually, major repairs can be avoided.
There are five non-destructive techniques that are commonly used for management of
predictive maintenance: vibration monitoring (with electrical current spectra)
monitoring of process parameters, thermography, tribology, and visual inspection. Each technique
there is a unique set of data that will assist the maintenance manager in determining the
real need for maintenance.
The predictive maintenance that uses vibration signature analysis is based on two
Basic facts: (1) all common failure modes have distinct frequency components of
vibrations that can be isolated and identified, and (2) the amplitude of each distinct component of
vibration will remain constant unless there is a change in the operational dynamics of
machine.
Predictive maintenance that uses process yield, heat loss, or others
non-destructive techniques can quantify the operational yield of equipment or systems
non-mechanical plant industrial. These techniques, used in conjunction with vibration analysis
can provide the maintenance manager or engineer of the industrial plant with factual information
that will enable you to achieve optimal reliability and availability from your plant.
How do you determine which technique or techniques are necessary in your industrial plant? How
do you determine the best method to implement each of the technologies? If you listen to the
sellers or sales managers who provide predictive maintenance systems, theirs is the only one
solution for your problem. How do you separate the good from the bad?
The most comprehensive predictive maintenance programs will use vibration analysis as
primary tool associated with current spectra, which usually comes associated in a
same data collection instrument. Since most of the normal equipment of the plant
Industrial are mechanical (driven by electric motors), the vibration monitoring will provide
the best tool for routine collection and identification of emerging problems. However,
only vibration analysis will not provide with high reliability the required data about
electrical equipment (the spectrums of the electric current that supplies it should also be used)
motor), areas of heat loss, condition of the lubricating oil, or other parameters that must
to be included in your program. Therefore, a total predictive maintenance program for the plant
industry must include various techniques, each designed to provide specific information
about industrial plant equipment, to obtain the benefits that this type of management of
maintenance can offer.
The specific techniques will depend on the type of plant equipment, its impact on the
production and other key parameters of the industrial plant operation, and the objectives that one wishes to achieve
that the predictive maintenance program achieves.

Márcio Tadeu de Almeida. D.Eng.

Professor at the Federal School of Engineering of Itajubá.
Consultant in Machine Monitoring by MTA.
ITAJUBÁ - MG.