GSaME Kolloquium: Future Production Systems: Risk and Safety

March 25, 2022, 2:00 p.m. (CET)

Time: 3/25/22, 2:00 p.m. – 3:30 p.m.
Lecturer: Jun.-Prof. Dr.-Ing. Andrey Morozov; Institute of Industrial Automation and Software Engineering (University of Stuttgart)
Download as iCal:

Industrial Cyber-Physical Systems, Cyber-Physical Production Systems, Software-Defined Manufacturing, Smart Manufacturing, Advanced Manufacturing, Industry 4.0, Industry 5.0, System of Systems, Internet of Things, Human-in-the-Loop, Digital Twins and many other related concepts describe several distinctive aspects of modern and future production systems such as (1) intelligence, autonomy,  heterogeneity, (2) high structural and behavioral complexity (3) reconfigurability and re-purpose-ability, and (4) sophisticated software that include AI components.

All the concepts mentioned above emphasize that accurate and up-to-date risk assessment and safety measures are essential because of the high cost of downtime and strict safety requirements. However, the analytical capabilities of risk methods applied in the industry are far behind the technical level of the systems in question. These methods cannot adequately describe sophisticated failure scenarios of highly dynamic and intelligent production systems. Besides that, the reliability and safety analysis of AI is an entirely open question. An inevitable revolution in safety methods is on the way that will provide us with the next generation of risk analysis and mitigation methods.

We will talk about the challenges that safety engineers of production systems will face in the nearest future. We will review the capabilities of available risk analysis methods their advantages and drawbacks. We will talk about how different methods can be combined and where to get input data to feed them. We will cover safety issues of software-defined systems, where a software update can drastically change the entire production line and require automated risk re-evaluation. Finally, we will discuss risk assessment of the systems that contain AI components, deep-learning-based anomaly detection, and mitigation.

To the top of the page