Edge computing can save IIoT projects from data inundation.
The Industrial Internet of Things (IIoT)is a disruptive technology that is gaining ground in manufacturing applications across all industries. The promise of IIoT is that better collection of streams of data from equipment will improve information and analysis, leading to more efficient production. Ironically, for such an information-centric technology, plenty of misconception and myth still clouds the topic.
One of those myths is that there will be an overwhelming flood of data that threatens to consume resources while delivering no measurable benefits. That doesn’t have to be the case.
Dealing with Data: The Power of Edge Analytics
We recently spoke with Jagganath Rao who leads Siemens Cloud Applications Services. He reported that less than 5% of all data generated in manufacturing plants today is analyzed for insights.
Where does all that extra data come from? Consider the following example:
A factory starts an IIoT project to track the performance of several conveyors and ovens. There is a continuous flow of data from each of those assets.
The electric motors continually advise their operating temperatures, vibration, rpms, and power usage. The oven sensors report temperatures across numerous locations. The conveyors report speeds, energy usage, vibration, product counts, spaces between products, and whether there is misalignment of products. This is a lot of data, particularly when it is monitored and reported several times per minute.
In cases like this, it’s no wonder that all of the data isn’t analyzed. Sending this data to the cloud would be a waste of resources.
What’s the missing piece that will allow manufacturers to take full advantage of the benefits of IIoT, while avoiding creating an ocean of data? The answer is at the edge.
If you aren’t familiar with the term “edge” in the context of cloud computing, it’s simple. While the cloud refers to the central node of a network where data can be aggregated, the edge refers to the extremes of the network.
For example, a factory may have ten PC machine controllers connected to a central intranet. Rather than sending machine data to a central server and then processing and analyzing the data centrally, processing and analytics can be performed locally at each PC, with the resulting information selectively sent to the central hub, highlighting trends, exceptions, or summary data. This reduces the need for IT infrastructure such as bandwidth and computing power.
Manufacturers can get all the benefits of better analytics that are accessible in a central location without sifting through the reams of data that are generated at each location, most of which is simply indicating that the processes are running as expected.
Which Data Should be Processed at the Edge?
Consider the following example of a production process in a consumer packaged goods environment.
The bullets on the left side of the graphic show data streams from the production floor. The collected parameters include vibration, temperature, inventory, and energy usage. From an analytics perspective, some of this data is useful without processing, and some data is only useful after processing. As discussed above, much of it simply reports normal data within expected parameters, which doesn’t require analysis.
For example, inventory level can be passed directly to the user, because the raw inventory data is relevant information by itself. Vibration or temperature data, on the other hand, is not very useful on its own. Rather, vibration data can be an indicator of machine health, for example, but only after processing to give it context.
So, edge computing can be used to handle the load of must-process data locally, while easily intelligible high-level data such as production counts and uptime can be passed directly to the cloud. In the cloud, the user can work with this data as well as the resulting edge-analyzed data, creating a cleaner, more efficient system. This prevents the issue of data inundation because the cloud never becomes crowded with the high-frequency, low-utility data.
So, how can a manufacturer begin leveraging the power of the edge?
Connecting Machines to Servers: Getting Configured for the IIoT
IIoT projects start by connecting a manufacturer’s machines to an IIoT data collection platform. With many of the options in the market, you might need to hire one company to put the data collection devices on the machine, and another company for the IIoT software. Some ecosystems, such as Siemens MindSphere, offer a more comprehensive package.
For manufacturers with legacy equipment, even without digital controls, retrofitting can be done to connect even outdated, unsupported machinery to the IIoT. However, most modern machinery is already outfitted with the sensors and I/O to connect.
Once your machine data is connected to the IIoT platform, there are a wide range of applications that are available, from predictive maintenance to asset tracking to performance optimization. In fact, many manufacturers choose to test several applications once their pilot environments are up and running, according to new research on IIoT adoption from Engineering.com.
With edge analytics in place, you can take advantage of high-volume asset data without inundating your cloud platform by configuring your assets to send information such as key performance indicators (KPI), uptime, downtime, and even to build performance or machine health indices.
How Edge Analytics Benefits IIoT Security
Another common IIoT belief is that storing factory data on a third-party server in another building, or on another continent, is a security risk. According to IIoT experts, this is not a myth. If attackers gain access to your network or to your asset I/O, it can lead to data breaches and unplanned downtime costs.
However, the myth of IIoT security is that nothing can be done to improve it. One benefit of edge analytics is that by processing the data locally, before the data crosses your firewall on its way to the cloud, you can more easily protect that data from attackers.
If you still don’t trust the cloud with your proprietary data, edge analytics can be configured to sanitize your machine data before it is sent to the cloud. In other words, if you process the data locally, you gain the power to control what goes in and out of your machine controller or other asset.
Process Locally, Aggregate Globally
The IIoT is one of the key pillars of the snowball of disruptive technologies known as industry 4.0. As that snowball gains momentum, manufacturers are feeling the pressure to adapt.
In this context, full-spectrum, turnkey IIoT platforms are worth a second look, whether you’re just getting started in IIoT or looking to take your implementation to the next level.
This article and the research it references has been sponsored by Siemens AG. They have had no editorial input to this article. All opinions are mine. – Isaac Maw