|
The Power
of Failure Analysis to Eliminate Process Interruptions
|
presented by
Charles J. Latino, President
& Founder of Reliability Center, Inc.
Paper Industry Maintenance
Conference
October 20-24, 1997
|
|
Our job is to run our plants with the absolute minimum of planned outages
without unplanned shutdowns producing high quality products at the lowest
possible cost. If we accomplish this, we are fulfilled. We know that we
have done the job. |
|
To produce with an absolute minimum of planned outages and no unplanned
outages, it follows that our operating and support staff have to be focused
on the right priorities. Let us review the reliability formula: |
|
|
Since maintainability is a small factor compared to reliability, we have
to conclude that our priority has to be Reliability or those actions that
will eliminate the causes of downtime. This does not suggest that it is
inappropriate to determine means to predict failures and thus limit downtime.
What it says, is that given limited resources you will get more for your
expense dollar by eliminating failures than limiting damage after primary
or component failure has occurred. |
|
The next question has to be how do we eliminate failures? We have two means
to accomplish this, one at the front end of the run cycle and one at the
back end. On the front end, we need to institute a precision paradigm.
One that says we will strive to do every task once and do it right the
first time. If we limit this concept to machinery it will be very difficult
to establish it as a paradigm. It is in our best interest to have precision
in administrative activities and in our manufacturing processing as well
as building it in when we work on machinery. We have all seen what happens
when crime labs do not act with precision. Criminal cases are put in jeopardy.
Lack of precision in industrial labs might have more dire results, certainly
for us. Procedures that are not thoroughly thought out can cause disasters,
loss of quality, and drain money depending on their application. |
|
Respecting machinery, Mobil Corporation ran a precision built BMW with
their synthetic oil providing precision maintenance, as delineated in the
owner's manual. One million miles were run before the car was dismantled
and all the engine parts measured for wear. This is the equivalent of 66
years of normal driving. They found that the vast majority of the parts
showed no wear and a few had very slight wear. So much for the statement
that we often hear that bearings have a very limited life. Indeed, the
very concept of running a bearing is that except for that split second
at startup, there should be no metal to metal contact if clearances are
precision designed and built. |
|
On the other end of the run cycle, we generally
run into failure of one
sort or another. Many will be mechanical or process related. Some will
be failures to obtain sales and provide orders and others will be failures
to build in precision. This latter failure reduces our confidence that
our machine can run for a long period of time making it necessary for us
to take planned outages. All of these failures can be resolved, and most
often eliminated, by using failure analysis. |
|
To get the maximum amount of leverage from failure analysis, it must be
applied on two tracks. On one track, Root Cause Failure Analysis is practiced
to eliminate and/or mitigate those failures that represent 80% of a facilities
losses in one focused parameter such as uptime or quality. These are called
the Significant Few Failures because they generally represent less than
2O% of a facility?s total focused failures in that parameter. |
|
It may seem obvious which failures represent the Significant Few, but experience
has demonstrated that they are usually not readily apparent. To help the
process of identification, it is suggested that facilities employ a Modified
Failure Modes and Effects Analysis. When the application of Failure Modes
and Effects Analysis (F\MEA) utilizes historical information on failures
instead of perspective or potential failure probabilities, the analysis
tool becomes simple to use and a much more powerful failure identification
tool. |
|
Many believe that the most significant failures are those sporadic events
that result when a plant experiences major trauma like a fire, explosion
or the destruction of major component in key machinery. As dramatic as
these can be, they seldom compare to the losses experienced by chronic
mechanical and process failures occurring hourly, daily, weekly and monthly
in our plants refineries and mills. Certainly, a $10,000,000 machinery
breakdown failure that occurs once every 10 to 20 years is not a small
incident. However, when amortized over 10 to 20 years it pales when compared
to the chronic failures that keep occurring with a frequency that is so
high that many of them become accepted as a cost of doing business. |
|
Consider a mining operation where conveyors carrying ore to surface treatment
plants stopped an average of four times in an eight hour shift. This was
not unusual as the belts are equipped with safety trip lines to protect
personnel and ore occasionally falls from the belts on the trip lines.
Since mining operations have miles and miles of conveyors, vehicles are
equipped with radios to direct drivers to tripped belts. In the vast majority
of trips, the drivers simply restart the belts. Downtime experienced is
an average of 15 minutes/trip. This was generally accepted as a cost of
doing business. However, when it was pointed out that the trips represented
12.5% of available uptime, management looked at the problem differently. |
|
Once the candidates for Root Cause Failure Analysis are delineated, the
process of analysis begins. Teams are carefully selected, success is defined
and clear criteria for recommendations are solicited from management. Careful
and comprehensive work is planned and executed to secure failure information.
These preliminary steps are vital to successful analysis. |
|
A cigarette manufacturer experiencing millions in dollar losses due to
cigarette rod breaks had to define the characteristics of each break to
determine break patterns. Six months of observation and high speed photography
identified 5 distinctive breaks each associated with different areas in
the machinery. |
|
It is to be remembered that when performing Root Cause Failure Ana1ysis,
we are working on 80% of our losses. Usually, we will save millions of
dollars. It is certainly worth the time and money to neutralize these losses. |
|
Analysis is secured through a reiterative process of defining hypotheses,
through deductive and inductive thinking, followed by verification. Most
of the time, plant experts have a difficult time solving chronic failures.
They feel they do not have the time to devote to a single problem so they
find themselves adapting solutions without proper analysis. We have found
that 50% of the time, the quick fix does not work. Prove this to yourself
by merely asking what your maintenance force does each day. Most of them
are doing what they did yesterday and last week and last month. |
|
Most of the time chronic failures are accepted as a cost of doing business
because each incident is usual1y not very dramatic in terms of cost or
time lost. What is, of course, missing, in this thinking, is frequency.
While one incident may have a small impact the combination of several similar
incidents can be very large. This is why Failure Modes and Effects Analysis
is so important. It takes into account frequency of occurrence as well
as Impact. |
|
Plant experts generally shy away from verifying hypothesis as they work
their way down on a logic tree. They feel it is not necessary because they
already know the answer. When we are engaged to guide an analysis the experts
report that they learn a great deal through verification. It was verification
that established the path that leads the way to root cases. |
|
When a logic tree of causes is driven down through the mechanical, process
human, and organization or management systems causes, we can turn to solutions
that will eliminate that problem from ever happening again. What?s more
the fix can, most often, be extended to other systems in the facility.
When that is done, we increase MTBFs and hence, the reliability and availability
of our manufacturing systems. |
|
Earlier, I said that failure analysis should be practiced on two tracks.
Root Cause Failure Analysis is one of the tracks. It is so important because
we find real solutions to those l0% to20% of failures that represent 80%
of our losses. What is left out is the failures that represent the other
20% of focused losses, the failures that are not our primary focus and
those administrative blocks or failures that prevent people from doing
their best work. To fully exploit the potential inherent in failure analysis
we must address these failures also. |
|
We call our second investigative path the Failure Analysis/Problem Solving
Track. This track is reserved for mechanics, operators and supervisors,
people very close to the business of producing product. They know the impediments
to productivity, they face them daily. |
|
To employ field people in this type of work, the methods to be used need
to be designed for them. The method must provide something fulfilling for
the field analysts. Time limitations, because of other duties, precludes
the in-depth analyses utilized when Root Cause Failure Analysis work is
performed. Supporting this work with field employees also needs to be accommodated. |
|
We have found that best results are obtained when field analysts are allowed
to decide for themselves which failures or problems need to be resolved
to allow them to do their best work. Field analysts are provided with a
priority technique to select those failures or problems that have the highest
impact on their work and are the easiest to accomplish. |
|
Analysis is performed using a logic tree just as is done when Root Cause
Failure Analysis is performed. However, validation routines are less stringent.
This means that the confidence one can place in the results will be somewhat
less. However, compared to the troubleshooting or trial and error methods
commonly used, the gain in valid answers to failures is truly astonishing. |
|
When Failure Analysis/Problem Solving is utilized and supported, we have
found that even using very reasonable assumptions, manufacturing facilities
ought to be able to realize benefits of at east $2,250,000 for every hundred
people trained. Considering that both the cost of the training and the
direct support will be $287,000, the returns expected are in the area of
800%. These are gold mines that await discovery. You will want to review
the assumptions that support this extraordinary return. They are: |
Out of 100 people trained,
it is expected that only 30 will initially participate. However, this 30
will form a critical mass that will draw in others as successes begin to
accrue.
|
Only 10% to 15% of work
time will be used to work on failure analysis and problem solving. For
the critical mass of 30 people, this amounts to roughly 3 additions to
staff. In the economics, this is figured at $60,000 per person.
|
Field Analysts will prefer
to work in teams so the assumption is that there will be 15 teams of two
working on failures and problems.
|
Each team of two working
on the average 12.5% of their available time will be able to complete only
6 analyses/year. This is a conservative assumption.
|
The average analysis
will yield only $25,000 in direct cost reduction and documented opportunities
like more uptime to produce product. This is the most conservative assumption.
|
Another cost element
has been included but it needs explanation. You may or may not know that
more than $50 billion are spent each year in the United States on industrial
training.
|
Studies vary, but between
1% and 20% indicate that of this is training translated into changed behaviors
on the job. In other words, most training dollars are wasted. We have studied
existing transfer environments and have concluded that most facilities
have paradigm blocks that preclude successful transfers. Accordingly, we
have included in our economics the cost of one training/mentor for ever;
100 trainee analysts at $7O,OOO/trainer/mentor.
|
|
You can deduce that I am an advocate of trainers not merely providing class
instructions but also mentoring the students that want to participate until
students feel comfortable with their new skills. |
|
The solutions that result from in-depth Root Cause Failure Analysis and
Failure Analysis/Problem Solving will, without question, increase reliability
to heights not imagined. The manager that wants to provide consistently
good operation and product flow can have it. The documentation supporting
this claim is mounting. Don?t be left behind. |
|
Before I conclude this paper, I would be remiss if I did not express one
caveat. The claims in this paper are real but to obtain them, traditional
thinking has to change. For example, management has to set out the expectation,
but then they have to serve their employees in the attainment of those
objectives This is much easier to write than to do. The gold is available
only to those that have the will, courage and perseverance to win it. |
RCI Offers the full
range of Reliability Consulting Services and Training Programs for Industry.
We conduct facilitations, reliability assessments, FMEA & Root Cause
Failure Analysis Training - Public & On-Site.
For more information
contact:
Reliability Center, Inc.
P.O. Box 1421
Hopewell, Virginia 23860
Phone: (804) 458-0645
Fax: (804) 452-2119
Website: http://www.reliability.com
|
|
Return
to Failure Analysis Reference Library Index
|
