Modeling Distributed Control Systems response time: from theory to measures

Abstract

As producer and designer of electricity production plants, Electricite de France must demonstrate that the critical systems have a limited response time. Particularly in the nuclear industry, a reliable upper bound for the response time must be calculated. In this paper, we present an implementation of an analytical evaluation of the response time of a Distributed Control System (DCS). This approach is based on a complete modeling of the system in the form of a timed event graph (TEG) which can be transformed into a system of equations in algebra (max,+). To model accurately the behavior of the system, we locally introduced weights into the TEG model and adapted the equations accordingly. By solving the equations of the occurrence dates of events in (max,+) Algebra, we obtain an analytical formula allowing the direct calculation of the worst-case response time of DCS. We then compared our theoretical result with a measurement campaign on the real system.