Error Classification and Human Performance Level Analysis of “Hiyari-Hatto” Incidents Caused by Maintenance Engineers

Abstract

Modern information systems are becoming increasingly standardized and are adopting multi-vendor system design and products. Fault tolerant designs have also been implemented and the information systems and the products themselves are becoming redundant. Consequently, in recent years, whilst incidents affecting the entire information system are suppressed, the maintenance operations are getting complicated and human errors, that have not been assumed conventionally, have occurred.

In this paper, we focus on “Hiyari-Hatto” incident data. “Hiyari-Hatto” is an incident which has the possibility to cause an accident but fortunately does not in actuality. We collected “Hiyari-Hatto” incidents experienced by maintenance engineers for information systems and classified them into similar product groups. We have applied error classification by James Reason’s theory and the human performance level by Jens Rasmussen’s theory. The analysis of the product groups could reveal the differences in the incidents that occurred for each product group. As a result, in order to prevent incidents due to human error, it was concluded that it is effective to devise measures based on the error type and the human performance level for each product group.