Aggregating Manufacturing Data

GE’s Predix cloud platform, which was built in a new facility the company set up in the heart of Silicon Valley, is an effort to “work with customers on their own digital industrial transformation,” Magee said. Predix is a platform-as-a-service (platform) based upon Pivotal Software Inc.’s Cloud Foundry that is tuned to the unique requirements to industrial applications. Among those are the need to process many different types of data on different schedules and in different locations with goals ranging from improving efficiency to predicting failure to optimizing logistics. And that processing may have to be done under the harshest of environments, such as inside a jet engine at 35,000 feet or on a remote oil rig in the North Sea. “Industrial companies have a set of unique requirements in terms of the types of data they work with, how they work with that data and the distributed nature of their plants and field assets,” Magee said. For example, industrial machines often don’t have the luxury of high-speed Internet connections, or even any connection at all. “Industrial companies need to put computing right on a wind turbine or a jet engine,” Magee said: “They want to aggregate data in a manufacturing plant without having to send everything to the cloud.”

Optimized for the Edge

Predix achieves that through what GE calls an edge-to-cloud continuum. It’s essentially “an end-to-end cloud-based operating system that also supports edge nodes,” Magee said. “Wherever the Predix system is, whether it’s on the tiniest medical device or the largest locomotive, you’ve got the ability to run analytics at the edge and to benefit from the cloud.” A core feature of the Predix platform is the “digital twin,” which is essentially a virtual representation of the physical assets the user is trying to optimize. This construct enables diagnoses and solutions to be matched to the unique situation and avoid the scenario in which “You and I drive our cars differently, but our manuals tell us to change the oil at the same time,” Magee said. “The digital twin gives us the ability to collect all that data, not by the manufactured spec, but by the performance requirements of a real-world environment,” he said. “You can optimize just for that asset.” It also has the flexibility to support highly customized applications. For example, data streaming from a locomotive engine may be needed for on-the-spot analytics to diagnose potential equipment failures, but also stored in the cloud and for analysis in combination with data from other equipment. Predix is also intended to offer flexibility at the back end for integration with legacy systems and equipment. “You’ve got all kinds of systems that have been there for many years, and companies now want to get data out of them to make more efficient operations,” Magee said. Predix uses the cloud as an interoperability layer to combine data from many sources and normalize it for analysis. Standards will be crucial to the development of IoT. GE was one of the founding members of the Industrial Internet Consortium, which is racing to set standards based upon real world scenarios. One of the standards efforts is around interoperability.

Where to begin?

How does an industrial company get started with Predix? There are myriad use cases, but most companies start with problems that have measurable payback, Magee said. “In medical equipment and transportation, for example, uptime is critical. Those kinds of investments are often a starting point,” he said. “Sometimes it’s in operational efficiencies like getting the right information to decision-makers who are doing maintenance.” Whatever the scenario, IoT analytics promises to make computing power useful in whole new areas of the economy. Speed and variety will be characteristic of many of these new applications. “Analytics could be critical to an executive who’s making decisions about whether to expand their facilities or sell electricity on the spot market,” Magee said. “Or it could be as simple as helping an operator use a piece of equipment more effectively because of best practices discovered from data.” Predix generated an estimated $6 billion in revenue for GE last year. With 50 million connected devices expected to come online over the next four years, the company has only scratched the surface of potential new value. “I think the sky is the limit,” Magee said.

Watch the full interview (15:40) here

http://siliconangle.com/blog/2016/08/15/how-ge-predix-tackles-the-unique-challenges-of-the-industrial-iot/