4.2 A Systemic View on Tailings Dams Failure Processes
41
and lead to prosecutions and larger fines. Due to the same influencing factors negligence and Force Majeure implications will certainly drastically change in the coming
decades.
4.2.3 Tailings Dams Failure Processes
The elements described above show the need for a systemic approach of the “failure
chain process” through investigations, design and construction (IDC) of tailings dams
and then service-life management and long-term monitoring (e-IDC).
For the discussion we opted for a probabilistic causality analysis. Publicly available incident and accidents data from the last hundred years were again used. The
predictive model is geared toward filling the gap between common practice and
“path to zero failures” goal and accommodates data-mining analytic and future
“lesson learned” that could make it possible to perform Bayesian updates (Dezfuli
et al. 2009) after the first estimates (after the a priori estimate) (see Sect. 10.1.2).
The model has to include CCF (Stott et al. 2010) in operations, risk assessment,
peer reviewing and inspections of tailings dams, at least in a simplified way, for the
sake of completeness and explicit inclusion of conflict of interest and complacency
(Oboni et al. 2013).
The Reliability Model
Engineering structures (and machinery, but also processes, including e-IDC processes) are systems consisting of a number of structural/physiological elements that
can individually fail. The way elements are connected and their reliability Xj, where
X j = 1 − p fj define the reliability of the whole system (Eqs. 4.1, 4.2). For a series
system (Eq. 4.1), failure of an element results in failure of the whole system. Reliability of the system is the product of the reliability of its elements. Equivalently, the
system fails if any component fails.
Success:
¯
X =
N
1
¯
X j
(4.1)
A parallel system (Eq. 4.2) is a redundant system that is successful, if at least one
of its elements is successful.
Success:
¯
X = 1 −
N
1
1 − ¯
X j
=
N
1
¯
X j
(4.2)
41
and lead to prosecutions and larger fines. Due to the same influencing factors negligence and Force Majeure implications will certainly drastically change in the coming
decades.
4.2.3 Tailings Dams Failure Processes
The elements described above show the need for a systemic approach of the “failure
chain process” through investigations, design and construction (IDC) of tailings dams
and then service-life management and long-term monitoring (e-IDC).
For the discussion we opted for a probabilistic causality analysis. Publicly available incident and accidents data from the last hundred years were again used. The
predictive model is geared toward filling the gap between common practice and
“path to zero failures” goal and accommodates data-mining analytic and future
“lesson learned” that could make it possible to perform Bayesian updates (Dezfuli
et al. 2009) after the first estimates (after the a priori estimate) (see Sect. 10.1.2).
The model has to include CCF (Stott et al. 2010) in operations, risk assessment,
peer reviewing and inspections of tailings dams, at least in a simplified way, for the
sake of completeness and explicit inclusion of conflict of interest and complacency
(Oboni et al. 2013).
The Reliability Model
Engineering structures (and machinery, but also processes, including e-IDC processes) are systems consisting of a number of structural/physiological elements that
can individually fail. The way elements are connected and their reliability Xj, where
X j = 1 − p fj define the reliability of the whole system (Eqs. 4.1, 4.2). For a series
system (Eq. 4.1), failure of an element results in failure of the whole system. Reliability of the system is the product of the reliability of its elements. Equivalently, the
system fails if any component fails.
Success:
¯
X =
N
1
¯
X j
(4.1)
A parallel system (Eq. 4.2) is a redundant system that is successful, if at least one
of its elements is successful.
Success:
¯
X = 1 −
N
1
1 − ¯
X j
=
N
1
¯
X j
(4.2)