Salzburg University of Applied Sciences and ABB’s Machine Automation Division – B&R – are working collectively to use synthetic intelligence to enhance vitality effectivity in industrial automation.
The collaboration, anchored within the Josef Ressel Heart for Clever and Safe Industrial Automation (JRZ ISIA), focuses on translating superior analysis into sensible options for industrial drive techniques.
A current milestone of this collaboration is the submitting of a joint patent software within the discipline of energy-optimized movement management for drive techniques utilized in industrial automation – akin to robots, machine instruments and automatic manufacturing traces – the place extremely dynamic movement sequences together with positioning, acceleration, deceleration and cyclic motion should be managed with excessive precision.
The event displays ongoing efforts to bridge educational analysis and industrial software.
The collaboration addresses a identified problem in industrial automation. Whereas typical management strategies depend on more and more correct mathematical fashions, they don’t at all times totally seize real-world vitality losses that may be measured however can’t be exactly modeled or described intimately.
To handle this limitation, the collaboration explores making use of AI – specifically reinforcement studying (RL) strategies that may study immediately from actual system habits.
A studying agent, deployed on the bodily system, interacts with the machine and autonomously learns how completely different movement profiles contribute to vitality losses, adapting the management technique accordingly, with out requiring a whole system mannequin.
A key innovation of the work lies in a brand new mathematical formulation of the educational technique, which permits quicker studying with decreased information necessities.
This enables reinforcement studying strategies – historically thought-about too sluggish and data-intensive for industrial use – to be utilized extra successfully in industrial environments whereas delivering improved outcomes.
Consequently, deployment in cyber-physical techniques turns into economically and technically viable, with the purpose of creating movement sequences considerably extra energy-efficient whereas totally reflecting actual working situations.
Stefan Huber, head of analysis at Salzburg College of Utilized Sciences, says: “The collaboration and the ensuing patent submitting clearly reveal how scientific excellence and industrial observe come collectively on the Josef Ressel Heart.
“Our intention is to make sure that analysis doesn’t cease on the laboratory, however ends in tangible technological improvements for business.
“Within the discipline of synthetic intelligence specifically, Austria and Europe want analysis on the technological forefront that delivers direct impression on industrial worth creation.”
Martin Haidacher, innovation supervisor at B&R, says: “Shut cooperation with Salzburg College of Utilized Sciences permits us to deliver revolutionary analysis strategies into sensible industrial purposes at an early stage.
“It highlights the worth of mixing educational analysis with industrial experience to advance the event of options that ship worth in real-world purposes.”
The underlying analysis builds on a number of years of improvement. Preliminary work dates again to 2020, when the subject was launched throughout the EU Interreg undertaking KI-Internet.
Since 2022, the analysis has been additional developed throughout the Josef Ressel Heart in collaboration with business companions from ABB’s Machine Automation Division (B&R), COPA-DATA and others.
Essential picture: Stefan Huber (left – head of analysis, FH Salzburg) and Martin Haidacher (innovation supervisor, B&R)
