New edge computing opportunities available for manufacturers
Placing devices near the edge of networks, close to machines, can bring in more data. That data can be analyzed at the edge, which can be faster than doing it at a centralized location. Collecting and analyzing data close to its source can allow users to know their production systems better.
Many industrial facilities have legacy equipment and slower-than-desired polling protocols. Edge computing can modernize and simplify things while obtaining more data with less latency. It also can allow more divisions within a company to be data consumers. Additionally, it can make the edge data the one source of truth, which improves data reliability. Edge computing also is ideal for enhancing the Industrial Internet of Things (IIoT).
Transferring data from legacy equipment
Nearly every company has legacy equipment. Getting data from there into supervisory control and data acquisition (SCADA) and other systems can be labor-intensive, especially when everyone has to connect. Companies have to poll that information, create mappings, and know exactly what each device is. There’s a lot of work involved.
Edge computing can simplify that infrastructure by doing mapping once—at the edge—and delivering data in a much more efficient way, using the publish/subscribe protocol message queuing telemetry transport (MQTT). With MQTT, devices report by exception.
If data is only sent when the values change, this means there’s less traffic on the network. That means users can get more data and they can get it faster. For example, if users need everything at a 50-ms rate, chances are they wouldn’t be able to do that with a polling strategy from a central system. However, when devices are next to the controller and publishing that data up, you can achieve faster rates and get access to more information.
Edge computing benefits
Edge computing offers simplicity with plug-and-play devices that require less maintenance. And once the data is published to a broker, SCADA can access it, as well as other systems, such as enterprise resource planning (ERP), information technology (IT), or business intelligence. With MQTT and the Sparkplug payload specification, numerous systems can auto-discover new data or information, without needing to know what the end device is. The edge computer can send raw data, a data subset, pre-processed information, already-acted upon information, or some combination, depending on the architecture and application needs.
As data collection goes more smoothly, so does control. Users are still writing back to the programmable logic controller (PLC), but it’s happening in a more efficient manner. Edge computing also can enable advanced process control (APC). There are software platforms that are just doing APC. They’re using more sophisticated algorithms than a PLC, and it’s usually expensive, specialized software.
Today, edge computing is making that world more accessible. It’s easier to build models because there are more algorithms that can be leveraged. It’s becoming easier to do real-time tuning of processes. That’s a pretty exciting area of edge computing. And the closer to the controller that you do it, the faster the whole thing will be.
Benefits: less scrap, downtime
Numerous real-world examples show how edge computing can make a big difference. Consider a semiconductor factory with many tools that put silicon on discs. The equipment is very finely tuned, and there are low tolerances for error. While it’s running, edge computing with these very complex models can be determining if the project is going off course or will be deviating from the course. The edge computer could contribute to making an adjustment (or make the adjustment, depending on the architecture) before there are real problems. Lost material could mean thousands of dollars in scrap, so this is a great example of how edge computing can save a lot of money by having a faster response than traditional computing.
Another example is prediction of equipment failure. Equipment can be extremely expensive, millions of dollars, or hundreds of thousands. It’s not advantageous to organizations to have the backup equipment sitting there on a shelf, having to spend that capital for something that’s not being used. However, if equipment breaks and they don’t have product for a day or two because they had to order a replacement, that could be millions of dollars in lost revenue.
Do users take that risk and not have a backup and hope the operators can see it’s failing? Even seasoned operators may miss such predictions. If the company has edge devices collecting data and comparing it against these models, they can detect patterns humans can’t. Having lead time in equipment failure will help the company. And it absolutely works better on the edge, because it’s focused just on that equipment.
Because data can be published and users can securely get data to the cloud, vendors that made the equipment can be a critical part of troubleshooting. The companies that built the machines can provide their expertise on potential failures. Equipment manufacturers can continuously monitor the equipment as a service and let the user know of any issues. Users don’t need to worry about that, because it’s being done for them. This is possible when equipment has an edge gateway that can publish data.
Edge computing enables other cool capabilities that never would have been considered with completely isolated, closed systems. There was a mentality of, “The system’s closed, and you can’t touch it.”
We’ve gone from that to thinking, “Hey, edge computing is letting us leverage IT, and leverage expertise from partners, and cloud platforms, and new algorithms.” Companies can do a lot more now. Customers are realizing new potential, saying, “The more I open it up, the more I can win.” And that comes with risk, of course, and greater attention to cybersecurity training, processes, and technologies. It’s very important to have platforms that are open, interoperable, and secure.
Questions to ask vendors
With the expansion of machine learning and analytics, people have more options, and they want to know how to get things done with these new techniques. There is a lot of confusion as to how the pieces fit together. Specifically, “What edge-computing platform do I get?” There’s a lot of smoke and mirrors out there. Users have to look for practical solutions they can add today.
The knowledge gaps are not in the functions, but in understanding what’s out there. It’s easy for a company not to know what to look for. Vendors have shiny new hardware, which might look open and interoperable. However, it may not be. The vendor should go beyond an IT approach and have an operations technology (OT) perspective, as well.
Legacy equipment wasn’t built to be replaced easily. If the vendor makes its product obsolete, the company better be able to replace the technology. Today, customers get to be choosy. With many edge computing options out there, customers can select all the best-in-class hardware and software they need with an interoperable approach. They can win with that without being locked into one vendor.
The IIoT has so much potential, but no one system can realize all of it. Vendors are seeing this as the future, and they’re saying, “As long as I’m open, interoperable, and secure, a customer will select me, because I have a value-add.”
Keywords: Edge computing, Industrial Internet of Things (IIoT)
Edge computing allows companies to analyze data faster and understand production systems better.
Edge computing technology vendors can help ensure connected machines work properly.
Machine learning and advanced analytics are enhancing how users can gather more information with edge computing.
What benefits can edge computing provide?