Managing Cyber-Physical Production Systems Variability using V4rdiac: Industrial Experiences

Abstract

Cyber-Physical Production Systems (CPPSs) are highly robust and versatile production systems that utilize diverse hardware components through control software. Employing a systematic variability management approach for developing variants of control software can reduce cost and time-to-market to build such complex systems. However, employing this approach in the CPPS domain is challenging. Engineering CPPSs require multidisciplinary engineering knowledge (e.g., process, signal, mechanical). Knowledge about CPPS variability is thus typically scattered across diverse engineering artifacts. Also, variability knowledge is usually not documented explicitly but rather tacit knowledge of mostly senior engineers. Furthermore, control software is commonly implemented using a graphical Domain-Specific Modeling Language (DSML) which only provides minimal support to express variability. This paper describes our experiences dealing with these challenges in an industrial context using a multidisciplinary variability management approach called Variability for 4diac (V4rdiac). V4rdiac is an integrated approach that allows CPPS engineers to conduct stepwise product configuration based on heterogeneous variability models from multiple engineering disciplines. V4rdiac also provides a mechanism to automatically generate control software based on a set of selected configuration options. We evaluate how V4rdiac implements and manages CPPS control software variants in the metallurgical production plant domain. We describe the benefits and lessons learned from using V4rdiac in this domain based on feedback from industrial practitioners.