Willkommen Gast

  • Login
Full load

Linked Forever Young Production Automation with Active Components
DFG Priority Programme 1593/2 (In cooperation with HSU, Hamburg)
In order to address software degeneration problems in long-living production automation systems, the FYPA²C project of the first phase of the priority programme introduced an anti-aging cycle that perpetually reinforces the consistency of software specifications and their actual behavior in operation. Project results of this phase included support for operators during undocumented software evolution based on methods and processes for gathering knowledge about production processes and their properties. In this approach, a knowledge carrying software checks desired pre-evolution knowledge constantly against the actual observable system behavior in order to detect evolutionary changes as a prerequisite to prevent software degeneration.

However, due to the complexity of realistic plant dynamics, evolution support in the operational phase is still rather limited w.r.t, e.g., automatic detection of evolution potential or predictions of evolutionary effects. To overcome this limitation, future production automation systems should be situated in an evolution environment which is not just aware of its properties, but also aware of its evolution potentials. As an opportunity to turn into evolution-aware platforms, nowadays production systems are already increasingly becoming parts of industrial networks. Consequently, new potential for cooperation support arise, because evolution steps have likely already been carried out on a similar system within a similar context.

Therefore, the new LinkedFYPA²C project proposes to exploit such experiences inherent contained in a network by envisioning an associated community of coevolving systems with the following main contributions to software evolution. (1) Development of an interpretable knowledge base to describe evolution processes of a Knowledge Carrying Software. (2) Methods and processes to exchange evolution experiences in a Machine-to-Machine communication in order to assess already performed evolution steps of coevolved systems and predict their impacts with context-related reasoning techniques. (3) A robust and flexible runtime platform and middleware for evolution which allows continuous anticipation and reflection of evolutionary changes by cooperatively detecting evolution potential to proactively trigger knowledge exchange. Together, these contributions provide an autonomous and decentralized management of collective experiences about evolution to establish the networked neighborhood of production systems as an active evolution environment. This will allow for a human assisting evolution support by presenting context-related evolution steps and their predicted property changes to human operators in order to achieve a better guided, foreseeable and less risky evolution for networked (production) systems.

Publikationen im Projekt
In: IEEE 15th International Conference of Industrial Informatics INDIN 2017