When the Internet of Things arrived on the scene a few years ago, it seemed like déjà vu for me, especially in application of discrete manufacturing. Back in the late 1980s and early 1990s, connecting discrete manufacturing machinery and systems to various versions of Ethernet was all the rage. Back then, the goal was to gather tons of data from Ethernet and send them to executive suites for “dashboard” applications on computers that gave executives a look at every click, every switch, every On/Off point generated by all the equipment on the manufacturing floor.
It wasn’t long before executives were saying it was too much data, and manufacturing moved to “management by exception.” All that data proved equipment was running well, and the only time anyone needed to get involved was when a switch didn’t flip, or a click didn’t happen. These developments led to preventive maintenance services and more efficient planning of equipment operation to take advantage of time periods when the cost of energy was low.
So how is the Internet of Things different from the activities of those days? Some users installing IoT systems report some savings and productivity improvements. But were these savings because they hadn’t taken advantage of the developments back in the ‘90s and only now have done so? In several cases, that was the scenario — making the IoT approach a bit less impressive.
The Internet of Things arose out of a speech given by former GE CEO Jeff Immelt. Under Immelt, GE made a large commitment to bring Internet connectivity to all departments and branches and reap the benefits of data collection. In a recent conversation with Rich Carpenter of GE Automation & Controls, Carpenter brought up some things GE has learned over the last several years.
Most IoT systems are heavy users of the Cloud. But, as Carpenter notes, the cloud does not solve all problems. Huge amounts of unanalyzed data are stored on the cloud, usually at a high cost. To Carpenter, and a number of other users of the cloud, cloud-only solutions no longer make sense. Carpenter believes the next iteration of the IoT will involve edge devices and the cloud. Edge devices will bring analytics closer to the production situation, reducing the amount of cloud storage needed.
Increasingly, customers will see a field agent, a gateway, or a platform for running analytics near equipment. However, IoT is still at the stage of collecting lots of data then discovering what data is necessary for analysis.
The low-hanging fruit of the IoT is to use the data in preventive maintenance, something that discrete manufacturing has been doing for a couple of decades now. One notable development is that increasingly, machine maintenance is being pushed back to the OEM by the customer. Depending on the industry, IoT equipment can be useful to handle this situation.
Carpenter sees changes coming to software analytics. Customers are interested in software platforms they can customize. Analysis programs will get to the point where they behave like PC software, upgrading on their own as analysis programs determine. According to Carpenter, that’s the end goal.
The IoT industry is still very much a solution looking for a problem. But who knows? With strong software analytics, maybe IoT will solve more manufacturing efficiency problems.