142
10 Defining Probabilities of Events
societal risks faster, even if its consequences of failure remain constant…. The methodology
developed in this paper enables us to “measure” and give a sense to a complex problem, to
transparently compare alternatives, to discuss rationally and openly the survival conditions,
or to evaluate the premature failure of a structure. The only way to slow down the increase
of the probability of failure is to repair damage occurring as a result of each hazard hit,
or to entirely avoid the damage. The second is generally “not feasible” for economic and
constructional reasons. Risks, especially long term ones, can never be reduced to nil.
As this present book is focused on the service life of TSFs and the closure phase
but not the long-term, “perpetuity” phase, we refer interested readers to (Oboni et al.
2014).
10.1 Probabilities of One Event
The scope of this text does not include mathematical skill development. However
there are a few mathematical tools whose concepts are important in the deployment
of sensible risk management when statistics are poor or non-existent and projects
evolve for many years, decades.
Probabilities allow us to consider the various sources of uncertainty and evaluate
their impact on the big picture. Thus we are of the opinion that even rudimentary probabilistic analysis (https://www.riskope.com/wp-content/uploads/2015/10/
Tailings-Facility-Failures-in-2014-and-an-Update-on-Failure-Statistics.pdf) is better than working deterministically. In fact, the inclusion of uncertainties is far superior
to “artificial” parametric studies (as, for example, varying one or two parameters at
a time to see their influence on the overall results).
The key question of what constitutes the essential (understood as basic, indispensable) and ideal (understood as “perfect”) data set to use is frequently asked by
users of risk assessments. There is no “simple” answer to that question, as we often
deal with facilities that may not even have been commissioned yet and past performances may not reflect future behaviour because of system or climatic changes.
Indeed, any internal or external change to the system has the potential to invalidate
the assumption that past experiences are sufficient to understand and calibrate future
implementations.
Let’s also remark that no risk assessment ever has the ideal data set available.
Indeed available data are generally gathered for other purposes, may be censored
and biased, and, most importantly, they reflect the past, not the future. This is the
case even in extremely regulated environments. Therefore the analyst must rely on
his/her skill and specific knowledge to use available data, either specifically from the
site(s), or from specific technical literature, to define framing probabilities ranges.
Of course any factual data (for example, over-topping, erosion, drainage and
deformations) will help immensely in framing a reasonable range of probabilities.
Records of near-misses can also be considered essential. Accident records can be
considered essential, although in many studies done in the past, there were no such
records, simply because the facility was not even in service. For future facilities,
10 Defining Probabilities of Events
societal risks faster, even if its consequences of failure remain constant…. The methodology
developed in this paper enables us to “measure” and give a sense to a complex problem, to
transparently compare alternatives, to discuss rationally and openly the survival conditions,
or to evaluate the premature failure of a structure. The only way to slow down the increase
of the probability of failure is to repair damage occurring as a result of each hazard hit,
or to entirely avoid the damage. The second is generally “not feasible” for economic and
constructional reasons. Risks, especially long term ones, can never be reduced to nil.
As this present book is focused on the service life of TSFs and the closure phase
but not the long-term, “perpetuity” phase, we refer interested readers to (Oboni et al.
2014).
10.1 Probabilities of One Event
The scope of this text does not include mathematical skill development. However
there are a few mathematical tools whose concepts are important in the deployment
of sensible risk management when statistics are poor or non-existent and projects
evolve for many years, decades.
Probabilities allow us to consider the various sources of uncertainty and evaluate
their impact on the big picture. Thus we are of the opinion that even rudimentary probabilistic analysis (https://www.riskope.com/wp-content/uploads/2015/10/
Tailings-Facility-Failures-in-2014-and-an-Update-on-Failure-Statistics.pdf) is better than working deterministically. In fact, the inclusion of uncertainties is far superior
to “artificial” parametric studies (as, for example, varying one or two parameters at
a time to see their influence on the overall results).
The key question of what constitutes the essential (understood as basic, indispensable) and ideal (understood as “perfect”) data set to use is frequently asked by
users of risk assessments. There is no “simple” answer to that question, as we often
deal with facilities that may not even have been commissioned yet and past performances may not reflect future behaviour because of system or climatic changes.
Indeed, any internal or external change to the system has the potential to invalidate
the assumption that past experiences are sufficient to understand and calibrate future
implementations.
Let’s also remark that no risk assessment ever has the ideal data set available.
Indeed available data are generally gathered for other purposes, may be censored
and biased, and, most importantly, they reflect the past, not the future. This is the
case even in extremely regulated environments. Therefore the analyst must rely on
his/her skill and specific knowledge to use available data, either specifically from the
site(s), or from specific technical literature, to define framing probabilities ranges.
Of course any factual data (for example, over-topping, erosion, drainage and
deformations) will help immensely in framing a reasonable range of probabilities.
Records of near-misses can also be considered essential. Accident records can be
considered essential, although in many studies done in the past, there were no such
records, simply because the facility was not even in service. For future facilities,