Community Research and Development Information Service - CORDIS

Abstract

The paper describes recent developments achieved in modelling human operators interacting with time evolving systems. The case study is referred to safety analysis of nuclear power reactors, but the methodology is quite general. This is based on the DYLAM (Dynamic Logical Analytical Methodology) approach for modelling systems in failed and transient conditions. DYLAM, extensively described elsewhere, differs from normally used reliability analysis techniques (such as fault trees, event trees) since it associates the physics of the process with the failure mode analysis. In this way, while simulating incident occurrences by introducing the possible component failure events, all relevant physical quantities (such as pressure, temperature, flow) are calculated during the transient time. As the operator surveys instruments which monitor such physical quantities, his diagnosis and subsequent interventions depend on the evolution of the plant process, and can be modelled only by dynamic techniques such as that proposed in the paper. The results show that the DYLAM modelling and simulation technique is able to incorporate any sophisticated model for approaching the complex diagnosis intervention - recovery loop characteristic of the man - machine interaction. As an example, together with a decision logic based on alarms and assigned procedures, a continuous decisional process by using fuzzy-set theory has been introduced.

Additional information

Authors: AMENDOLA A, JRC ISPRA ESTAB. (ITALY);STUDIO MESA, MILANO (ITALY);REINA G, JRC ISPRA ESTAB. (ITALY);STUDIO MESA, MILANO (ITALY);CICERI F JRC ISPRA ESTAB. (ITALY), JRC ISPRA ESTAB. (ITALY);STUDIO MESA, MILANO (ITALY);STUDIO MESA, MILANO (ITALY), JRC ISPRA ESTAB. (ITALY);STUDIO MESA, MILANO (ITALY)
Bibliographic Reference: ANALYSIS, DESIGN, AND EVALUATION OF MAN - MACHINE SYSTEMS, VARESE (ITALY), SEPT. 10-12, 1985 WRITE TO CEC LUXEMBOURG, DG XIII/A2, POB 1907 MENTIONING PAPER E 31835 ORA
Availability: Can be ordered online
Record Number: 1989124023800 / Last updated on: 1987-01-01
Category: PUBLICATION
Available languages: en