Author:
Angelo Hulshout is an experienced independent software craftsman and a member of the Brainport High Tech Software Cluster.
Reading time: 3 minutes
Angelo Hulshout has the ambition to bring the benefits of production agility to the market and set up a new business around that. Next to technical challenges, this also raises business-level questions.
There’s a lot of hidden data in factories and getting that data out can help improve the manufacturing process and the logistics around it. But getting that data out of the production equipment is a real challenge, as I pointed out last time. Not every machine has the possibility to report its activity or progress. Sometimes because it’s not built into the controller (PLC), sometimes because there simply is no controller. I’ve had a few interesting discussions about how to technically approach this, but also on how to deal with it on a business level.
Technically, getting data out of a system that has no ‘data interface’ can be a relatively easy exercise. If you have a digital controller like a PLC with a network connection, almost everything is possible. Adding a little bit of software logic will make any data item available that can be derived from data inside the controller and I/O signals coming from the actual equipment. Without such a controller, we’ll have to use a slightly different approach. We can put an intelligent sensor on the power leads of a motor or set up a camera that detects machine movement, eg to see when a machine is operational. This will help us see when the machine is ‘doing something,’ although it may not tell us whether that ‘something’ is useful production work.
Automated system
Before making modifications, the question is: what data do we need from a machine? If we have a sawing machine in a factory that produces wood panels, it might be interesting to know how many production jobs pass through that machine every day, so we can match that with the number of planned jobs. It could also be that we’re interested in the comparison between the machine’s production capacity versus its actual production.
These are business-level questions. What do we want to achieve? What information is required to achieve it? What data is needed to generate the information? Are we interested in monitoring our production plan or in maximizing the production time of an individual machine? Or both? The latter might lead to conflicting solutions. After all, Eli Goldratt’s Theory of Constraints already taught us in the 1980s that a bottleneck in production often occurs exactly because of the desire to maximize the production hours of the most expensive machine in the factory.
To prove his theory, Goldratt collected the data by hand, using monitoring equipment in production lines. The data analysis would bring the issues to light and lead to possible solutions. My goal is to find a way to collect the data, analyze it and then have an automated system (eg through machine learning and/or digital twinning) define what needs to be changed to improve the production process – also across plants that are involved in production of the same materials.
Shinchoku
A lot of this may look obvious at first sight, but the more people I talk to, the more I realize it’s a big step. Smart industry is based on good ideas and supporting technology. At the same time, the target users include a lot of smaller factories that still have to take initial steps in automation before smart industry solutions become useful for them. Next time, I’ll go over a full, although partly made up, production line, to show where this has an impact and what can or should be done in different areas to get closer to smarter manufacturing.
Since all of this is about improving manufacturing processes and making progress, we’ve put a name to my idea: Shinchoku (進捗, Japanese for “progress” and “innovation”). It’s now officially part of my company. If things continue as they stand now, it will spin off into an independent startup within the next 18 months.
