Title :
The prediction of human performance safety with event trees
Author :
Heslinga, G. ; Stassen, H.G.
Author_Institution :
N.V. KEMA, Arnhem, Netherlands
Abstract :
The extent to which event trees can be used to predict human errors made by operators in following written procedures and what consequences may result from these errors was studied. The problems that may occur were analyzed by means of a field study in which the startup of an actual installation was examined. The difference between analyzing single and sequences of human errors was studied as well. It appeared that discrete operator tasks can be studied efficiently because of the nature of the event tree. Continuous actions can also be analyzed, but it requires a realistic decomposition into discrete actions. Qualitatively speaking, the analysis of sequences of human errors had more value than the analysis of single human errors. Quantitatively speaking, there was not much difference since the probabilities of undesired consequences occurring were virtually equal in both types of analysis
Keywords :
behavioural sciences; human factors; risk management; safety; trees (mathematics); decomposition; discrete operator tasks; event trees; field study; human errors; human performance safety; installation; probabilities; startup; Control systems; Ergonomics; Error analysis; Human factors; Laboratories; Man machine systems; Performance analysis; Redundancy; Risk analysis; Safety;
Journal_Title :
Systems, Man and Cybernetics, IEEE Transactions on