Combining system architecture modelling with dynamic process simulation for early stage fault and effect analysis

Designing of complex process plants, such as a nuclear power plant, requires the development of the physical process and the automation system controlling it. When dynamics of the physical processes are included, it becomes challenging to identify all possible consequences arising from different component failures, including common cause failures or degraded modes of operation. In this paper, we present a concept for supporting fault and effect analysis using architecture
description language capable of modelling software and hardware components and their faults of the system with a dynamic process modelling simulator. We firstly use the error modelling and analysis of the architecture to find potential critical combinations of component faults within a complex system. Secondly, we simulate the effects of combined faults on the controlled process in order to analyse system effects. We test the method on an early design of a safety system, called the Halden Safety Fan, while using Architecture Analysis and Design Language for architecture modelling and Advance PROcess Simulator for the dynamic simulation. The Halden Safety Fan system is an early conceptual design, offering a high-level description of a proposed modernisation of the existing emergency ventilation system of the Halden BWR reactor. Results indicate that
proposed early-stage failure assessment can easily be performed using a model as input, gaining confidence on design choices.