Predictive Analytics

June 09, 2020

Predictive Analytics

Early in 2020 we announced the implementation of Adapdix’ EdgeOps™ platform technology within ficonTEC’s advanced photonics production systems. If you read the press release, then I can spare you the repetition here.

Essentially, machine learning is utilized to monitor all production process parameters and the algorithms ‘learn’ to flag anomalies – so-called cyber-physical systems are empowered to make decisions on the own, and a whole new level of visibility to intelligent data is provided, so one learns more about the systems, how they perform and, more critically, what actually affects performance, downtime and yield (and also what does not). This all falls under the Industry 4.0 tag, and is already implemented in some areas of mainstream industrial manufacturing.

The benefits should be immediately obvious: greater machine autonomy, better maintenance scheduling means reduced unscheduled downtime (20 to 40% !!) means higher overall throughput and better yield at lower TCO. To borrow a phrase used in advertising for a particular brand of credit card – ‘priceless’.

So, naturally we have the first pilot schemes running with significant industrial operators of ficonTEC systems. Unsurprisingly, they are not the only ones that have been learning – there were some (thankfully, only) minor surprises for us too. Realistically, this is of course not entirely unexpected when one considers the customization that goes into most systems, and that can only help us further improve performance for all systems.

If you are a multi-system user and think your production might benefit from intelligent autonomy, please get in touch …

Related Webinar: AI/ML-based Process Control

AI/ML-enabled applications in manufacturing are in the news, and rightly so. However, while already a cornerstone of Industry 4.0, there is still some way to go for their implementation in the photonics sector. We overview the application of this technology at ficonTEC and what this means for the customer.