- Register


Editor's Pick


Real time edge AI control

17 July 2018

At this year’s PPMA show Omron will showcase a new solution that adds artificial intelligence (AI) functionality to packaging and processing machinery for real-time control that provides an instant reduction in waste and machine downtime.

The AI controller is based on a machine learning engine (MLE), which is incorporated into Omron’s Sysmac control platform (on both PLC and IPC) and monitors multiple signals simultaneously to detect potential anomalies (faults) and determine the appropriate response. The whole calculation can be accomplished in under a single machine cycle, minimising any disruption to the operation of the line. 

AI offers functionality which would not be achievable on traditional control and programming solutions. To demonstrate the AI controller and the value it can add to the production process, a technical cell at the PPMA exhibition will show a machine filling containers with liquid, until a fault is deliberately applied. The AI controller will detect the fault and calculate in real-time how to avoid a faulty operation and recommence the filling process. In a similar application without AI, the machine would have to be stopped, with the need for manual analysis of the fault and physical adjustments to be made. This will inevitably cause downtime to production and reduce the efficiency of the machine.

As robotics and industry 4.0 becomes adopted more widely in manufacturing, it is apparent that current cloud or networked AI solutions are ideal for collating and processing huge amounts of data for long-term performance trends and applications such as predictive maintenance. The issue with these solutions is that they cannot react in real-time, to control machine functions, even with today’s high-speed communications technology. Adopting EDGE-AI technology in your machine on the factory floor, allows tighter control of the machinery in the line, which can bring an immediate increase in productivity and Overall Equipment Effectiveness (OEE) without the complexity of modelling every aspect of the complete system.

Omron marketing manager Dan Rossek comments: “There is a huge buzz in the industry about how EDGE-AI processing can change the way we work. Today, while the majority of companies are still talking about the potential of AI applications, Omron has gone a step further and brought machine learning into the production line. Moreover, the solution that will be shown at PPMA is less complex to implement than proposed AI solutions and will allow manufacturers to access some of the benefits of Industry 4.0 immediately on a platform that is easily scalable.

By employing this solution at machine level, the AI controller can be taught to recognise anomalies and abnormal patterns much easier than traditional techniques, which would require potential faults to be determined at the production line level and require many more input signals and additional computing power to analyse the results. The AI algorithms are pre-integrated into the control system to create the framework for real-time optimisation without the need for further programming. The result is a solution that is powerful but remains easy to implement, with immediate benefits for manufacturers. Stand B32