Dynamical reliability and availability evaluation and validation of distributed control systems
Pavol Barger, Jean-Marc Thiriet, Michel Robert · 2003
The evaluation of reliability and availability of control distributed systems is the main purpose of this paper. These control systems are composed of intelligent sensors and actuators (with calculation capabilities and communication interfaces), processing units, and a communication network, e.g. fieldbus or real-time network. The modeling of the system is ensured thanks to Petri Nets. Color Petri Nets allow the evaluation of both the dependability aspects and the performances with a control point of view. In order to get some results, it is possible to achieve a simulation or to make an analysis of the occurrence graph. The work presented is a model of a tank, in which the level of the liquid is controlled. An analogue sensor is used for the measurement of the level and is used for the control of the pump which fill the tank. The sensor is extended by some possibilities not to supply the controller with a measurement, as a result of a failure, and a degraded strategy can so be set. The evaluation of the reachability of this state after failure is achieved thanks to the occurrence graph tool. The global aim is to propose a methodology for the evaluation and comparison of distributed architectures.