Titel_SS06

4.4 Assessment of risk

Within different application areas of risk assessment various rather specific methodologies
have been developed and this has had the effect that risk assessments across the boundaries of
application areas are difficult to compare and even more difficult to integrate. Numerous
procedural schemes for risk based decision making are available but these focus on the project
flow of risk assessments rather than the framework for risk assessment itself. Moreover, one
of the most significant drawbacks of existing frameworks for risk assessment is that they have
not been developed from a Bayesian perspective, i.e. do not sufficiently facilitate and enhance
the potential for utilizing evidence and/or indications of evidence in the assessment of risks.
Therefore the generic risk assessment framework illustrated in Figure 4.6 is proposed. This
framework facilitates a Bayesian approach to risk assessment and full utilization of risk
indicators.

In Figure 4.6 the system which is considered subject to a risk assessment is assumed to be
exposed to hazardous events (exposures EX ) with probabilistic characterization
p(),1,EXkEXPk n

(

, where denotes the number of exposures. Generally exposure events

should not be understood as individually occurring events such as snow loads, earthquakes
and floods but rather as the effect of relevant combination of these. The probability of direct
consequences

nEXP

cCD ij)associated with the jth state of the i constituent of the system due

to the exposure on the constituent is described by

Cij

pC( ij EXk) and the associated conditional

risk isp()CEij Xck D(Cij). The summation and integration of the conditional risk over all

system constituents and states, respectively is denoted the vulnerability of the system in
regard to the considered exposure. The risk due to direct consequences is assessed through the
expected value of the system vulnerability over all nEXP possible exposure events as:

1

()()(

nEXP

D ij k D ij k)

k

R pC EX c C pEX



 (4.1)

Finally the probability of indirect consequences associated with the system

state due to the exposure , the state of the constituents C and the associated direct

consequences is described by

cScID(, ())k DC

Sk EXk

cD()C p()SEXlk and the corresponding conditional risk is

pS EX c() (lkID(,Scl DC)). The integration of the conditional indirect risk over all possible

system states can be seen as a measure of robustness; indicating the ability of the system to
limit the total consequences to the direct consequences for given constituent state and
exposure. The risk due to indirect consequences is assessed through the expected value of the
indirect consequences in regard to all possible exposures and constituent states, as:

11

()(,())(

nEXPnSTA

ID l k ID l D k

kl

R p S EX c S c p EX )



 C (4.2)