Robust control filter for manufacturing systems : application to PLC training

In this paper, we show how research dealing with the design of a robust filter for manufacturing controllers can be used in the field of PLC training. Maintenance tasks and improvements in the manufacturing process involve changes in the PLC (programmable logic controller) program that can have bad safety consequences. Hence, a PLC program which is at the startup of the production lifecycle safety validated can become unsafe. Today, this problem has increased because remote maintenance monitoring is a reality in manufacturing systems. We have proposed to design a robust filter placed inside the PLC which authorizes or forbids outputs from the PLC. The filter is composed of several logical constraints which have to be respected at each PLC cycle. In order to guarantee the filter quality, it is necessary to check that all constraints have been well defined. For that, in order to validate the filter, an original formal method has been proposed. This one is based on a modular modelling approach of the manufacturing system and by considering the most permissive PLC program. UPPAAL checker is used to verify that the filter is sufficient to avoid the manufacturing system to reach dangerous forbidden states. If it is the case, the filter guarantees the safety whatever the PLC program. Hence, the filter still remains active even if there are changes in the PLC program during the lifecycle of the production system. This approach is applied to safe PLC training in order to test on line the students’ PLC program. The explanatory capacities of the filter are used to supply explanations. The paper is illustrated with an application based on a virtual “sorting system”.