SCHERER, D.; http://lattes.cnpq.br/9876878380575511; SCHERER, Daniel.
Resumo:
The analysis of accidents and incidents is essential for the study of human error and to
propose ways of prevention. The research on the human error is focused on the classification
of the error and on identifying the relationship between the error and the human activity that
caused it. The literature on this subject proposes methods and taxonomies for categorizing
errors, however, these methods are not adequate to support the analysis and understanding of
the circumstances surrounding the error. For this it is necessary to combine methods,
ontologies and taxonomies to the decision making sequence that resulted in the error, added to
the mapping of the interrelationship between the elements associated with the error.
The work presented in this document deals with the human error study specifically in
the context of electric systems operation. It discusses the error categorizations and taxonomies
found in the literature, in an attempt to relate those with the steps of the human decision
sequence that might lead into error.
The study is based on a corpus of reports that describe accidents and incidents which
happened during the operation of electric systems in a company belonging to the electric
systems market in Brazil. In order to understand the error causes it is also needed a user
profile that contains both objective and cognitive aspects associated to the work context
description. Both the work context and the user profile are described with the proposed
taxonomy, which was validated by its use during the analysis of a second corpus of error
reports, originated in the same company. The text also presents a tool designed to support the
application of the proposed taxonomy when reporting and analyzing accidents and incidents
in the operation of electrical systems. The investigation of the relationship between work
context, user profile and the error, was performed by verifying a set of hypothesis using
conceptual maps.
The achieved results consist on the proposal of a process to be employed in the
analysis of error reports in the domain of electric system operation, and the proving of
hypothesis relating work context, operator behaviour and error incidence, in automated
electric substations, based on the corpora of study.