|About the author:
José Durán is an Electrical Engineer with a Masters
Degree in Maintenance Engineering. In the last few years he has been
working as a Consultant Engineer in Operational Reliability Improvement
process for oil companies in Venezuela and Colombia, consulting, training,
and facilitating activities like: Reliability Centered Maintenance,
Industrial Risk Management, Maintenance Optimization, Maintenance
Project Management, etc. It is covering different aspects of the oil
industry: extraction, pumping, refining. Also services: electric power
generation, gas injection, etc.
More information: please contact him:
All of us have some
ideas about what "risk" is, not to insure the car “is a risk”,
playing the lottery “is a risk” a little bit different, to risk on an
investment, etc. All these gave us some ideas about a probability that
something may happen, a probably win or loss event. In fact almost
constantly we are living in a world driven by risk, although sometimes we
are not aware that we are using this concept in the process of making
mathematic world we may explain the risk like probability (frequency) of
event times its consequences. As we can see in engineering the risk is
associated with negative consequences
R = P X C
R = F X C
Both equations are
equivalent and will have value units divided by time units ($/year, $/month,
etc.). About the differences of both equations: We must not talk about
frequencies of events that have never happened, in these cases we use
the "probability" concept.
The majority of
industrial and engineering operations are in the risk world. Almost
everything done involves probabilities and consequences. The future never
will be predicted in a 100% and there are a lot of factors that contribute
with uncertainty, for example:
influences (oil prices, political changes, etc.)
characteristics (reliability, availability)
contractors performance and suppliers
> Human errors
and bad communications
Fortunately it is
possible to transfer some calculation methods in the fields of probability
and risk to the industrial field.
What risk value is
acceptable? This is a million dollars question, it is not simple but it is
approachable. Starting with logic rules and company vision direction. What
is really true is that risk must be evaluated and controlled under accepted
levels. Tools like criticality and lost opportunity studies give us a quick
evaluation of the risk factors that are currently affecting us. For big
potential risks we must develop deeper analysis like HAZOP studies, for
Let us see how to calculate
a risk exposure:
If we have a
big event probability (fire, explosion) estimated in one every 100-50 years
and the consequences valued between 10 and 200 million dollars, then the
risk exposure would be between:
R = $100,000/year ($10 millions
and 100 years)
$4,000,000/year ($200 million and 50 years)
How to estimate the event
Information options will
allow us to see some possibilities:
a.- Computer system:
When we have it and it is well data provided. It may be a really valuable
information supply, although many times its information is too general and
does not allow us to distinguish clearly some data like event causes.
experiences: They tend to know their equipment really well but the human
capacity to estimate is not efficient enough when we start to deal with
extreme and non common quantities. Another conflictive point is the
different opinions between the workers.
c.- Available databases:
Other option is to use databases gathered by third parties for example OREDA
(Off Shore Reliability Data), IEEE (Institute of Electrical and Electronics
Engineers). They can give us approximations to some factors like failure
rates; normally those data may be referenced values or benchmarks to do
primarily evaluations. The fact is that different operational conditions
make those data really dispersed .
The event consequences
in the operational point of view could be summarized in:
failures: They are caused by reliability problems. This includes
repair costs, production losses, etc.
II- Falling Performance.
The production capacity (quality, quantitative) is affected by the event, so
the net sales and the profit margin are compromised. Typical examples are
capacity deterioration caused by fouling process in filters, compressor
III- Rise of production
cost: The event consequence that triggers a production cost rise that may be
caused by an efficiency loss (more energy consumption, materials, etc.),
maybe you require more people to do the same tasks, etc.
IV- Effects on life
cycle cost or capital investment: In these cases the consequences are
because of the equipment life cycle is affected by the events occurrence,
for example the overhaul cycles of equipment may be affected by the stopping
and starting and by the maintenance cycles. The capital investment (for
example the buy of new equipment) also may be affected by these events.
V- Legal requirements:
In these cases the event may be associated with the breaking of a safety or
environment requirements. This may unchain big impacts to the environment or
workers' health with big economic penalties.
VI- Shine factor. It
includes the other factors of a company's daily life that are difficult to
evaluate like welfare, external appearance, community relationship, etc.
Evaluating the Risk
There are several
methods to calculate the risk; in this case we will only see a really simple
way to do initial calculations. The risk may be evaluated under this scheme:
I- List the bad events.
II- Investigate the
frequencies/probabilities of the events above.
III- Estimate the event
consequences of each event.
IV- Calculate the risks
associated with each event.
V- Identify the bigger
risky events and to prepare strategies to their reduction.
VI- You can calculate
the total risk using this equation (you must use the same time base)
Total Risk = ?Fi X Ci
Changing the risk.
Starting with the
equation we can see that the risk may be modified cutting the frequencies or
consequences or both of them.
Risk Analysis in an
Operational Reliability project
In a project of
improvement of the risk analysis is always present,
when we make criticality or lost opportunities we are doing risk evaluations
of systems or situations under risk. And when we propose new operation,
maintenance or security strategies we must be capable to estimate
their future impact in the company.
So the risk concepts we use allow us to take
decisions based in more than suppositions.
Cost Risk Analysis
As we saw before, the
risk may be modified. One of the most powerful tools to decrease the
industrial risk is the Reliability Centered Maintenance. This suggests
maintenance strategies based on operational conditions of the company and
the impact of the possible failures in security, environment, production and
repair costs. Whatever the philosophy used to decrease the risk we must to
be able to evaluate the benefits of the results and a really good way to
determine what are the risk concepts.
How can we do a Cost Risk
A risk analysis of this
type (non-optimization) may be faced calculating the risk before and after
the new strategy implementation and the costs' change . Deeper analysis must
be done when we are studying critical processes or big investment (there is
cost risk software in the market). Some processes do not seem to be in this
category because they do not have a big probability, but they could produce
big consequences in any of the categories defined above.
It has some advantages
and disadvantages. Among the advantages we can see that it is a
probabilistic concept leaving a movement range. The disadvantage: this
is not a really easy concept “to sell” in the whole organization and its
When we are treating
this evaluation a weak point in any analysis may result in the difficulty to
get reliable data. The management uncertainty is a big research field.
Having absolute reliable data is extremely difficult. How can we assure
about the average “life” of some components?
Is the uncertainly bad?
In fact the only bad
thing is the incapacity to work with uncertainty, it may bring analysis
paralysis; when the will to find exact results convert people in perfect
researchers without decision making capacity.
Managing the uncertainly
A good way to face the
uncertainty is to use these cases: More likely, best case and worst case and
to do sensibility tests trying to make more reliable prognostics. Other
possible choice is to manage numbers with tolerance ranges (80 +/- 10 %).
The fact of presenting decision ranges make our studies with a good security
margin and at the same time, give them a bigger credibility than trying to
be perfect calculators generating “exact” values using uncertainly data.
Let us see some advices to research data:
1. Show how the
information will be used, what it is needed for, and what is the precision
2. Ask without
3. Test highest and
lowest values without influence of your opinion.
4. Use indirect ways to
indicate facts that may not be real.
Now there are in the
market some tools to evaluate engineering decisions based on cost/risk
concepts, the usefulness of these tools has been demonstrated with many
million dollars saved for many companies around the world using them for
these typical cases:
2. Optimize inspection intervals.
3. Project evaluation.
4. Optimize Spare part stockholding.
5. Optimize Shutdown strategies.
AT NO COST
TO YOU you can
support our website when you
visit our sponsors and advertisers
and get valuable information from them.
They make it possible for us to keep this
free service up to date for you.