IEC61499 as an Architectural Framework for Integration of Formal Models and Methods in Practical Control Engineering

摘要

Modeling is extensively used in research work on intelligent manufacturing to solve such problems as optimal dynamic reconfiguration of production facilities, or optimization of material flow and throughput. Also, models are used for avoidance of possible deadlocks, or for prevention of malfunctions caused by incorrectness in control algorithms. As the components of automation systems are getting more autonomous, the more embedded and reusable must be the models. Models may give opportunities to simulate the system's behavior, to analyze system's structure (such as connectivity of certain elements), or to analyze properties of system's dynamic behavior even without conducting the simulation. A production plan may impose a branching structure of the control with multiple scenarios of automation system's behavior. However, even a purely sequential control may lead to such behavior due to malfunctions in some hardware elements. In both cases the formal analysis might be the only way to reveal potentially dangerous situations. This sort of analysis requires modeling of closed-loop plant/controller systems, where the model encapsulates properties of the plant, the controller as well as the structure of the system. Modeling can be especially beneficial for flexible, re-configurable automation systems. In traditional hierarchical control systems, re-configuration requires to redesign controllers, sometimes in a very sophisticated way. Then the new functionality must be tested to ensure the desired behavior in the closed loop, and eventually the new functionality must be encoded in a language used for controller programming. In the same manner process monitoring, diagnosis, and supervision by the operator have to be re-designed with similar levels of difficulty. All these tasks have to be accomplished quickly to minimize lost production. A faster system integration, however, could be achieved if the testing routines were substituted by formal validation of newly arisen system configurations.