ABSTRACT

The issues “risk assessment” and “tolerable risk” are causing conflicting reactions not only among Health and Safety experts. Experienced designers in the machinery sector are sometimes unsettled, too. The controversies are mainly about numerical probabilistic representations. These are new in the field of general machinery safety, and the key term “probability” turned out to be ambiguous. Recently introduced probabilistic methods encounter a well-proven practical state-of-the-art, which is merely based on qualitatively defined requirements.

If objective findings are to be taken as a basis, it is obvious that the sequence of the total annual figures for reportable accidents can be considered as random independent events in a population of comparable elements. Mathematical concepts seem to make sense if the annually recorded (machine-specific) accident data are interpreted as the overall result of a huge “random experiment” in a relevant observation framework (e. g. DGUV statistics in Germany).