Making IIoT real with Predix: EDGE connectivity to GE’s Predix platform

By Mike Bacidore, editor in chief, Control Design

Jun 29, 2017

ThinkBIG 2017 Banner

Data historians, HMI/SCADA systems and edge devices generate large volumes of valuable data. Ensuring that this data is accessible in the cloud is the first step in the Industrial Internet journey, and GE Digital’s Predix provides cloud-based advanced data visualization and analytics.

At AutomaTech’s ThinkBIG 2017 User Conference in Cape Cod, Massachusetts, one of the more popular training workshops demonstrated basic system configuration and presented the concepts needed to connect GE’s Historian and HMI/SCADA data to Predix. In one of the hands-on demonstrations, participants connected a Raspberry PI to move sensor information.

The in-depth training session focused on showing users how to begin utilizing the Predix platform, to enable their existing systems to leverage the cloud, and presented a GE Digital and Predix platform overview. Included with the workshop was a hands-on lab where attendees could operate real-time connectivity between GE Historian and GE HMI/SCADA to Predix.

“We connected a Raspberry PI to get sensor information to Predix Time Series,” said Domenic Verte, manufacturing applications manger at Toray Plastics in North Kingstown, Rhode Island, and one of the participants in the session.

“The workshop was designed to demonstrate how easy or difficult it would be and to show the level of complexity to get information from an asset on the production floor to the cloud,” said Verte.

“At Toray, we would be connecting a slitter, grinder or some piece of equipment that would produce operational data and then analyzing that data in Predix,” explained Verte, whose company produces polypropylene films, polyester films and polyolefin foams for packaging.

“We currently use GE’s on-premise products—iFIX and Historian,” said Verte. “In the workshop, we connected a historian to the Predix Time Series. We set up five tags and connected the Raspberry PI. It was very easy.”

Verte said he’d previously watched the training videos online but hadn’t tried a hands-on exercise until the workshop.

“It would be great to blend the operational data with the IT data and gain insight into a problem more quickly than we have in the past,” said Verte. “Currently we deal with problems in a reactionary way. We’d prefer to use a system to alert us so we can make the correction before the problem occurs.”

Verte said he comes to AutomaTech’s ThinkBIG 2017 User Conference to see the products and gain a better understanding of their operation and updates. “The conference provides us with the information we need to build use cases that we can potentially use at our plants,” he said.

Kevin Davis, process control engineer at Middlesex County Utilities Authority ( in Sayreville, New Jersey, also found a great deal of value in the training sessions. “You know they’re good workshops when the three hours goes by so quickly,” he explained. “We were connecting data through an industrial-gateway-server (IGS) driver and used the SIM driver.

“During the workshop, it was very easy to connect historian to the cloud,” said Davis. “We are setting up Predix cloud for ourselves, and I like that the base is free, but you can choose to pay for analytics. Companies like to look at the trends, and it’s important for us to know if we’re trending up.”

Davis continued, “We collect data back to two different historians, one is real-time and the other is an FTP collection into historian. It allows us to still get the data collection if the cell goes down.

“With Predix, the trending capability and features and analysis of data are a lot more powerful,” explained Davis. “You can change the scale and do analysis in the cloud. We have 75 meter chambers and are trying to make sure our flow numbers make sense. Our five pump stations average a flow of 100 million gallons per day (MGD), and our current analysis is human analysis. They have a server-to-server client that you can put in the DMZ and then reuse that as a relay to Predix, so you’re not exposing your historian.”

One of the features Davis appreciates with Predix is the way everything auto-scales. “We have a measured flow and a calculated flow for the plant, and they should match,” he explained. “You can manipulate and program the scale with Predix. Connecting the data is very easy. For security, you could use Historian, put it on the Internet and connect to the cloud.”