Model-based training of manual procedures in automated production systems

Abstract: Maintenance engineers deal with increasingly complex automated production systems (aPSs). Such systems are characterized by an increasing computerization or the addition of robots that collaborate with human workers. The effects of changing or replacing components of such systems are difficult to assess since there are complex interdependencies between process parameters and the state of the components. This paper proposes a model-based training system that visualizes these interdependencies using domain-independent SysML models. The training system consists of a virtual training system for initial training and an online support system for assistance during maintenance or changeover procedures. Both systems use structural SysML models to visualize the state of the machine at a certain step of a procedure. An evaluation of the system in a changeover procedure against a paper-based manual showed promising results regarding effectiveness, usability and attractiveness.