Consider this. You touch something hot and you immediately drop it or snap your hand from the heat source. It’s automatic — you do this so instantly that you don't even think about it. Ladies and gentlemen, this is our wonderful nervous system at work — which consists of your brain, spinal cord and nerves — which work together so well that it feels instant or real-time. It sensed the hot object in your hand and signaled your muscles to let it go subconsciously. Basic studies have tracked this process of “feeling pain” to see that some nerves transmit at more than 200 MPH.
Quite amazing when you think about it.
One day after a dentist visit my mother (bless her) bought home some fish and chips for lunch while my mouth was still numb from the anesthetic. Two hours later I was in horrible pain because I had burned the top of mouth bad and chewed a hole in my tongue because I couldn’t feel the pain. My nervous system was blocked by the pain killers and I ended up doing some serious damage to my mouth. It occurred to me that this is what happens with manufacturing intelligence with it capturing events (both good and bad), transferring them via integration to a central place for viewing and actions. So, as a machine starts producing bad parts, the quicker these events are captured, moved and analyzed the less impact of rework or scrap occurs. This is called the “feedback loop.”
Nerves let us know when something is wrong. The more important something is the more nerves we have there such as our hands and mouth. This is the same for manufacturing. If it is important, you must be tracking these critical events and “listening” to problems. Automation solutions like historians and SCADA systems have traditionally done this — however, MES systems are becoming advanced and encroaching on this space. Either way, if it is critical to your operations such as a bottleneck machine, then it needs to be measured (it needs nerves).
Nerves filter out normal and announces the bad stuff. It is great that you have nerves in your plants but it needs logic to tell you something is wrong, just like I need something to scream that the delicious fried goodness I was eating was too hot. One customer (who shall remain nameless) boasted to me that his system was capturing events at 450 records a second per machine. Literally 10 minutes later the same machine broke, stopping production. After digging in, it was apparent there was no alerting, tolerances or feedback on these captured events. “Why are you capturing events so fast?” I asked. “Because I can,” was his reply. Just because you are capturing events does not mean you are measuring them.
Feedback must reach the “brain”. If your nerves are screaming “stop” and these same messages are blocked or not occurring, then it is pointless because your brain can’t do anything about it. It’s the same with manufacturing. Information needs to be filtered out and presented to the end user so that actions can be taken. Engineers need visibility and understanding of the process so costly mistakes can be avoided.
The faster the feedback the lower the “pain”. My brain eventually got the message that there was a problem — it was just two hours too late. If all was good, it would have taken a split second to realize the food was too hot and I wouldn’t have put it in my mouth. Same with manufacturing. Events need to be moved, transformed, analyzed and presented in near real time so that actions can be done quickly. These actions can be done automatically by the system (such as slowing down the machine) or by users such as operators, engineers, supervisors and so on. The faster this happens, the smaller the impact.
Feedback must be understood. Pain messages sent from the nerves must be understood. “Hey this food is hot you should let it cool,” would have been a lovely one to get. This is what manufacturing intelligence does — it gets the flood of data and transforms the information so that it can be consumed and understood by key people. From here, corrective and hopefully preventative actions can be undertaken.
The brain remembers so we don’t make the same mistake again. That is the beauty of our brain. We learn from our mistakes so that (hopefully) pain only gets experienced once. Manufacturing intelligence highlights the pain, but it is up to the people designing and managing production to understand the root cause and stop it from happening again. There are many ways to do this, from memos to complete process changes. This is where business process modelling becomes an essential part of manufacturing solutions today.
So there you have some examples. Hopefully many of you already have a magnificent manufacturing nervous system up and running to avoid smaller issues becoming large. If not, then please learn from the pain of others and look into ways of listening to your production process using manufacturing intelligence. I hope your days are painless, your meals are at a good temperature and your feedback loops small!
James Montgomery is a Senior Product Manager at Apriso that is focused on the company’s manufacturing and business intelligence capabilities, and how this information can be used for quality and continuous process improvement.