Factories of the Future (IIoT + AIoT)
I have a friend who has an uncle who runs a small magnet factory in Sheffield, England. This is an old family enterprise. I’m not sure exactly how long they’ve been in business, but I think it must have been running since at least the 1850s — around the time that Charles John Huffam Dickens (1812–1870) was busy creating some of the world’s best-known fictional characters.
A few years ago, my friend took me to visit the factory one weekend when no one was working. The ground floor looked a bit like the image above — lots of individual freestanding machines, each of which would have its own operator. These machines were used to cut, shape, grind, polish, and paint blank magnets in a tremendous variety of shapes and sizes.
Meanwhile, the upper floor was a different world. First, there was a room when the magnetizing process took place. In the days of yore when this process was a carefully guarded secret, this room would be kept locked and bolted — only the Master Magnetizer (I don’t know his real title) and possibly his apprentice (the Minor Magnetizer?) were allowed to cross the threshold.
The remainder of the upper floor was occupied by the owner’s opulent office. This office could have featured in a Dickens novel. The entire room (floors, walls, and ceiling) was lined with beautifully polished wood. The owner sat behind a monster wooden desk, which was mounted on a raised stage that occupied the back third of the room. A couple of chairs were located on the main floor in front of the desk. The stage was only about one foot (25 cm) high, but the effect was to place visitors at a major psychological disadvantage since they were at a lower level than the manager and had to crick their necks to look up.
My friend told me that the company constantly receives letters from around the world from companies that have been customers for over 100 years. As an example, he showed me a letter they had just received from a railway company in India requesting a quantity of magnets of a certain size and shape used to perform some unknown function on the trains or perhaps in the signaling system. Who knows?
The company in India were requesting a part number from the dim and distant past — their previous order had been more than 70 years earlier — and, amazingly enough, this family factory actually had the required parts in the storeroom (if not, they would have pulled out the original specifications and made a new batch).
Factories of Yesteryear, Today, and Tomorrow
The point of the above (yes, of course there’s a point) is that for every large manufacturing plant there are hundreds of small factories. Many of these facilities — both large and small — still have equipment like that shown in the image above; that is, legacy machines without any form of modern sensor-based control systems. In fact, it is estimated that there are trillions of dollars of such equipment in the USA alone.
What we are talking about here are machines that rely on their human operators to ensure they have sufficient oil and to keep a watchful eye to make sure the machines don’t overheat or lock-up or do something untoward.
The next level up is to augment these legacy machines — or purchase new ones — with traditional sensor solutions to measure things like oil levels, temperatures, pressures, and flows, along with things like limit switches. The data from these sensors can be monitored locally using a microcontroller (MCU) or a programmable logic controller (PLC), which can control the machine and do things like turning it off if something untoward occurs, like a temperature rising above a specified value.
Did you see my column What the FAQ are the IoT, IIoT, IoHT, and AIoT? In that column I noted that one definition of the Internet of Things (IoT) is: “A system of interrelated computing devices, mechanical and digital machines, objects, animals, or people that are provided with unique identifiers and the ability to transfer data over a network without necessarily requiring human-to-human or human-to-computer interaction.”
In the same column, we noted that: “[The Industrial IoT (IIoT)] refers to interconnected sensors, instruments, and other devices networked together with computers and industrial applications, including manufacturing and energy management. This connectivity allows for data collection, exchange, and analysis, potentially facilitating improvements in productivity and efficiency as well as providing other economic benefits.”
According to the IoT Agenda, “The Artificial Intelligence of Things (AIoT) is the combination of artificial intelligence (AI) technologies with the Internet of Things (IoT) infrastructure to achieve more efficient IoT operations, improve human-machine interactions, and enhance data management and analytics […] the AIoT is transformational and mutually beneficial for both types of technology as AI adds value to IoT through machine learning (ML) capabilities and IoT adds value to AI through connectivity, signaling, and data exchange.”
We are already starting to see machines being equipped with AI and ML capabilities that can perform sensor fusion (combining sensory data from disparate sources), detect patterns in the sensor data, and use these patterns to predict possible futures. In addition to detecting and mitigating short-term issues (“This machine is running too hot, let’s shut it down”), it’s also possible to analyze the data over time and detect longer term trends (“This machine has been getting gradually warmer for the past couple of days — if the current trend continues it will go out of spec in a week’s time — so let’s schedule a maintenance inspection ASAP”).
We are also starting to see the deployment of more sophisticate sensor solutions, including audio, vibration, and video, where the latter may include areas outside the visible spectrum, such as infrared (IR).
Consider how a human operator might gauge the “state-of-health” of a machine. The operator might place a hand on the machine to determine how smoothly it’s running from its vibrations. The operator might also be alerted by small changes in the way the machine sounds — an unusual squeak or a slight change of pitch. Based on years of experience, the operator may be able to associate these phenomena with certain conditions (“Ah, I bet the main drive shaft needs a little attention”).
It isn’t difficult to imagine a not-so-distant future in which machines in factories are equipped with AI/ML systems and appropriate sensors, and these systems are trained by their human companions to have similar levels of expertise. Even when no humans are present, a slight change in vibration, a small alteration in sound, a tiny waft of smoke may cause the automated systems to leap into action and take whatever steps are necessary to keep the factory running or to shut things down gracefully before serious problems occur.
This is only a small taste of things to come. In my next Factories of the Future column, we will consider the combination of the IIoT with augmented reality (AR) and mediated reality (MR) systems. Until then, as always, I welcome your comments, questions, and suggestions.
