Information Technology Reference
In-Depth Information
is questionable. However, numerical values may be used for the validation of
severity categories.
It is natural to classify railway accidents by the following guiding questions:
1. Does the accident scenario typically involve a single person (P) or multiple
persons (M)?
2. Does the accident typically lead to light (L) or serious injury (S) or fa-
tality (F) of a person involved?
Tabl e 2 . Examples of signalling functions
Code Function description
=LBB Detect track vacancy
=LBC Detect train at a particular spot
=LBD Locate train
=LCB Determine train description
=LDB Provide diagnostics
=LEB Supervise driver vigilance
=LEC Automatic train stop
=LED Supervise braking curve
=LEE Supervise maximum train speed
=LFB
Optimise train running
=LGB
Monitor points
=LGC
Lock points
=LGD
Monitor derailer
=LGE
Lock derailer
=LGF
Monitor level crossing
=LHB
Provide signal information
=LJB
Provide cab radio
=LKB
Display state to driver
=LKC
Display state to dispatcher
=LKD
Transmit commands
It should be noted that
- “Typically” does not mean worst case; in a safety meaning, it should be
interpreted as a “typical” bad outcome, i.e. worse than average.
- If statistical data were available, e.g. a 90% percentile of the outcome
distribution could be taken. This means that 90% of accidents have a less
severe outcome, but 10% have a more severe outcome.
- f no trustworthy data or expert estimates are available, then a conserva-
tive choice has to be made, e.g. if an accident may typically lead to severe
injury or fatality, but if no clear distinction is possible, then the category
fatality has to be chosen
