The report should be considered a guide for the treating of human errors: for identifying their possibilities of occurrence when designing well-structured human tasks and for their improvement when they occur in reality. For these purposes a strong coupling between predictive and retrospective analysis is emphasized: In order to control human errors, post-incident analysis of cases with human errors in a given industrial plant should be performed as means of feedback from reality for the verification of results of predictive analysis and also as a general means of identifying and improving such human errors which cannot be expected covered by predictive analysis. Primarily, the guide addresses people with a knowledge of the technical plant in question and involved in the safety-oriented design and improvement of human activities and without a particular human factors background. The main report describes the procedures for post-incident analysis and for Work Analysis, which is a search strategy developed for well-defined activities, e.g. test and calibration, and constitutes a formalized procedure for the pre-identification of relevant human errors leading to a lack of task result and/or to immediate effects not covered by the lack of task result itself. Work Analysis and the post-incident analysis procedure are based on a common description system for human malfunctions. This system is explained in appendix and so are its underlying models and way of reasoning. Finally, a word index is provided for supporting the reader. 4 references. In English.
|Place of Publication||Roskilde|
|Publisher||Risø National Laboratory|
|Number of pages||70|
|Publication status||Published - 1985|