How edge devices are poised to change maintenance

Robert Skeirik, Director of Product Management at Emerson's machinery health solutions, has written, published and presented numerous papers in multiple languages across six continents. Skeirik has work with Emerson for nearly 30 years, and has served in product management for the last 25. During the live Q&A portion of the webinar, "Case Study: Real World Digital Transformation In A Manufacturing Facility," Skeirik shares how he has helped manufacturing teams bring about real-world digital transformation in their facilities.

PS: How many sensor inputs are required to monitor an asset?

RS: That's an interesting question. It depends on how critical the asset is, and it depends on what type of asset it is. A typical motor pump pair could be very effectively monitored with probably four inputs. There's a gearbox in between, and it probably needs a minimum of six. Typically, if you were looking at monitoring pressure before and after the strainer on a pump, and temperature, and possibly flow, you're up around 10 to 12 inputs.

PS: Is it common for analytics to run on an edge device?

RS: No, actually, this is a new concept and it's really come into its own through the industrial internet of things and Industry 4.0. It is really a new thing to have a single edge device that is actually taking a sensor signal, processing it, and running analytics on it all in the same box, in the field. I don't know of that being done before now.

PS: Where is the collected data stored?

RS: The data can be stored in multiple places. In the case of Emerson AMS Asset Monitor, it has its own historian. It actually can store up to 10 years' worth of data internal to the device, but you can use OPC UA. With standard industry connectors like OPC UA, you can pretty much store the data anywhere you want.

PS: Have you seen any one of the traditional MRO functions – maintenance, reliability, operations – take the lead on digital transformation projects? One trend I’ve observed is, since maintenance teams are burdened with so much of the actual fix-it work, that oftentimes reliability teams are taking on the strategic planning role to help digitize asset management.

RS: I don't think that's new. I think, especially among your top quartile, you will find they probably have a Vice President of Reliability. I've heard the Vice President of Chevron say that reliability is safety. With top quartile companies, reliability is a core competency in their facility, and it is frequently differentiated from maintenance, which is more repair. But it's also not uncommon for both topics to be addressed by the same team. It's a management philosophy question. But you get the biggest bang for your buck if you have well-trained people focused on reliability.

In that same presentation where the VP of Chevron was talking about reliability equaling safety, he expounded on how reliability impacts their financial goals, their operational goals, their HR goals, and safety as well as health and environmental. So reliability is just huge. Knowing that assets will function the way they are intended to function allows the plant to perform the way it was meant to perform and then correspondingly, to generate the revenue it was intended to generate.

PS: What are some of the elements that someone who wanted an internet device, would build into the case? Would it include things like a cloud subscription or other sort of integration services in addition to the technologies?

RS: It certainly could, and I would want to look at those costs. In the example the Emerson AMS Asset Monitor, there are no subscriptions. You could push the data to the cloud, but you don't have to because it will do the analytics in the device and then broadcast the analysis back to whomever needs to see that information along with a recommendation. When you want to look at an ROI, certainly you want to look at costs for cloud storage, you want to look at accessibility to the data in the cloud, you want to look at acquisition cost of the device, installation costs, and any types of subscriptions, which there are none with this in our example.

If it's hazardous rated, self-contained, and it can be mounted next to the machine with very short cable runs to the sensor, your total installation cost is significantly lower. Then if it can support a wireless backhaul, like we do with our solution, you avoid cabling for communication. You want to look at the level of training that's required for the employee to actually use the data. If it's in human language in the examples that I've shown, that's going to lower your training cost. It also will help to alleviate potential disruptions as there's turnover in employees.

On the return, you want to quantify the importance of that asset to your overall production, does it create a slowdown or a shut down if that device is lost? Is there a difference between controlled maintenance on the machine versus catastrophic failure?

PS: Why is it important to have the measurement of electronics in the gateway?

RS: A traditional system would measure the data in, say, a rackspace solution and then push it over into a database, but there's a lag that comes about through this multi-step process. By having a single device that brings in the sensor signal, does the measurement electronics, does the embedded prescriptive analytics, and functions as the gateway, you have real-time responsiveness to any type of changes that are occurring in the assets that can then be sent out as alerts in near real-time. Without that, you've got this delay and if people are on vacation, even with an expert system, the results might sit to be reviewed. I believe part of the value proposition of digital transformation is this automated, near real-time access to asset health information.