Management summary
The Swiss Federal Office of Transport (FOT) developed
a method for evaluating the acceptance of the individual risks to passengers and railway
employees (see Slovak et al., 2012; Baeriswyl et al., 2013). Our aim was to assess the
current method applied by FOT, to validate assumptions of the current approach and to
suggest improvements of the current approach if needed. We focused on the individual
risk for passengers.
It is important for railway companies to assess individual risks of a particular technological
system or a planned structure as well as the overall individual risks induced
by the railway transport. However, investigating an occurrence of a rare event (e.g.
death of a passenger in a train collision) leads to estimating extremely small probabilities
which is usually imprecise. To solve this problem, we proposed a new approach
of individual risk assessment. Our approach is based on viewing the individual risk as
a random variable and calculating the probability, that the individual risk is acceptable.
The proposed method is demonstrated on actual data collected by FOT.
Recommended changes in the individual risk assessment methodology
Comparison with the acceptance criterion derived from the minimum endogenous
mortality is possible only in concern of fatalities. Therefore, the value of
the collective risk for passengers should be adapted.
The estimation of the average travel distance should be related to the whole railway
network of Switzerland and not to the largest Swiss railway company SBB
only.
Statistical confidence of the estimated individual risks has to be considered. We
propose the application of the Bayes’ theorem providing additional information
for more confidential decision-making. This extended approach enables more
realistic setting of several parameters that were originally set very conservatively.
We recommend a more transparent calculation of the passengers individual risk.
By the use of the proposed pre-generated tables of acceptance for evaluation
of the statistical confidence, there is no additional effort for a user of the risk
assessment methodology.
