Multi-level approach to virtual commissioning: a reconfigurable assembly system case

Abstract

Software plays a key role in modern automated systems. Testing the control software is one of the final steps in the development process and when done on the physical set up, it creates risks (e.g. project delay, system damage and operator injury). In addition, with flexible systems it is almost impossible to test every possible scenario on the real system. Using a digital model allows testing different configurations before the physical hardware is available and thus without risk for the operator and the system hardware. A digital model of the reconfigurable system can also help validate new configurations virtually while the real system still operates in its current configuration, allowing for a quick risk-less changeover. Depending on the level in the control hierarchy, the simulations can run faster than real time. This paper demonstrates the use of a digital model during the commissioning at different levels. On each level, other requirements need to be fulfilled and a different level of detail is needed in the model. Every level asks for a different set of input data, that needs to be consistent to result in a reliable emulation/simulation. The methodology is evaluated on a reconfigurable assembly system, consisting of multiple base units with modular, relocatable add-on modules. The most detailed level is the process and machine level, considering the low-level control logic, movements of the machine and robots, collisions, etc. Cell and plant level commissioning focuses on the interaction of multiple machines, transport, production schedule and assignment of tasks to the individual machines. This level typically covers a longer time span and requires less detail of the actions on the machine level. It interacts with higher level control software and a simplified version of the low-level control logic is implemented in the model itself. This paper describes the methodology and lessons learned to reuse the digital model at different levels of the control hierarchy while remaining the consistency by exchange of information within the digital model and with the real system. Further research is described towards automated virtual test case scenarios with a digital model for a reconfigurable assembly system.