Torbz - Fotolia

Three promising IIoT platform options aim to improve manufacturing

Given the exploding industrial internet of things market, surveying the choices can be overwhelming. To get you started, here's a look at three platforms and their focus areas.

Global internet of things spending is expected to reach $1.29 trillion in 2020, according to IDC. The research firm also believes a substantial portion of that spending will come from the manufacturing industry, which is investing heavily in the industrial internet of things. It's an assessment also held by P.K. Agarwal, regional dean and CEO of Northeastern University Silicon Valley in San Jose, Calif.

"Given [this information], every vendor is trying to position itself to be the platform provider, the cloud provider, the data provider," said Agarwal, who was also the chief technology officer for California under former Gov. Arnold Schwarzenegger. "But I don't see a clear dominant player emerging yet. It's still kind of a Wild West out there."

To get a sense of some of the technology being created as vendors try to stake their claim in this Wild West of the industrial internet of things (IIoT), here's a look at three emerging platforms.

IIoT platform focuses on asset performance

Predix is what GE calls an "operating system for the industrial internet of things," said Marc-Thomas Schmidt, chief architect of Predix at GE Digital in San Ramon, Calif.

Predix is the industrial, cloud-based platform from GE Digital that's designed specifically to handle industrial data and analytics. It's a tool that helps GE customer companies build software that powers internet-connected industrial machinery, such as jet turbines, trains or factory robots, said Schmidt.

Predix is similar to other cloud platforms, such as what Amazon, Microsoft and IBM have, in the sense that it enables customers to build big data analytics applications, which is what most industrial customers do, he said.

"But it's different, in some respects, than those more generic platforms, and the difference comes from the fact that we cater to one audience: the industrial space -- companies that make use of industrial machinery and extract value from that machinery," Schmidt said. He believes that "if they have assets and they want to improve the performance they get out of those assets, Predix is the platform for them."

Another differentiator, according to Schmidt, is that Predix is a hybrid platform. That means that part of the applications that customers build with Predix runs in the cloud, but a good chunk of the application runs where the machines are, which Schmidt calls "the edge."

"So we have an edge software stack that our customers use, and we have a cloud software stack," he said. According to Schmidt, "one of the services they find in our cloud that is different from anywhere else is an asset service that they use to monitor their assets by modeling digital twins of those assets."

A digital twin is a virtual model of a physical asset, whether it's a jet engine or a compressor or any other specific machine, that's built on Predix, according to GE Digital. The digital twin enables a manufacturer to unravel data from the machine and predict how that machine will perform.

"Then we have an analytics framework that makes it very easy for data scientists to create analytics or tap into a catalog of predefined analytics, then use the analytics to make the most of the information they get from their machines," Schmidt said. "And we have a set of specialized data services that take good care of the industrial data that we find."

IIoT platform seeks to open up interoperability

Siemens also offers manufacturers a cloud platform for developing, extending and operating cloud-based apps, dubbed MindSphere, which is built on the SAP HANA cloud.

"We call it the open, cloud-based internet of things operating system; not open source, but in terms of interoperability," said Jagannath Rao, senior vice president of data services at Siemens AG, in Buffalo Grove, Ill.

Manufacturers have always produced a lot of data through their control systems and their drive systems. However, because they have disparate data sources, they have never been able to harness the richness of the data by connecting and integrating all these data sources, Rao said.

"To get valuable insight, you really want to find correlations in the data, but that wasn't possible because of the disparate sources," he said.

Rao believes "this platform provides the ability to integrate all these data sources and upload large amounts of data into the repository. And the computing power that we have in the applications enables manufacturers to get insights they could never get before."

MindSphere consists of a connectivity layer called MindConnect that connects machines, plants and systems and on boards them to the cloud infrastructure, Rao said. Not only can companies connect Siemens controllers and devices, but they can connect any industry-standard device that's on the shop floor.

"On top of that is the software as a service [SaaS] layer where the value is created -- you have the applications culminating in a SaaS layer to enable Siemens and third-party applications to deliver value to Siemens' customer base," Rao said. "All these layers, infrastructure and connectivity and SaaS, were designed in a way that they are completely open."

Siemens also wants MindSphere to be cloud provider-agnostic, and is in the process of opening the platform up to Amazon and Microsoft Azure, in addition to SAP HANA, so customers aren't locked into one provider.

"And the SaaS layer is going to be an open ecosystem. We'll have an API that will be released so anybody can build applications that are of value," Rao said. "Siemens will also have its apps on there in the areas where we bring domain knowledge; for example, driving trains, analytics or machine tool analytics. That ecosystem will grow through a larger web of partners and, hopefully, in a few years, we'll offer an industrial app store."

This summer, Siemens plans to release development kits for the MindConnectivity layer so, in the future, vendors can MindSphere-enable their equipment, Rao said.

Sweetening the IIoT platform mix with new answers

Houston-based Honeywell International Inc. is also targeting the industrial internet of things platform market. The company said its approach focuses on helping customers solve long-standing business problems using the technology.

IIoT provides new methods for solving some problems that have been around for a long time, said Paul Bonner, vice president of consulting and data analytics for Honeywell Connected Plant, which helps manufacturers improve the safety, efficiency and reliability of operations so they can become more profitable.

Honeywell's answer to helping manufacturers improve operations is Sentience, a cloud industrial IoT platform that delivers secure big data capabilities for all of its connected products. One of the applications the company has developed using IIoT and its Sentience platform is CPS (Connected Performance Services), Bonner said. CPS is designed to give refineries and petrochemical and gas processing plants greater visibility into their operations.

"We've taken the reactor models and the rigorous models of UOP processes and we've turned them into digital twins," he said. "So we've created a rigorous mathematical model of UOP licensed processes and we connect the data, we have a secure data transfer from the customer's process plant, and we bring the data into our Sentience cloud platform and run the data through the models. And from that data, we can do asset management, looking at things like the energy performance, the yields, the efficiency and the reliability of the customer's process."

This application then compares the process data to a digital twin of the process and reports back to the customer, offering recommendations or suggestions on how to actually improve the yield, the efficiency and the reliability of that process, according to Bonner.

"We've built that out and deployed it, and it's one of the most advanced uses of Sentience," he said.

Next Steps

IIoT needs edge architectures -- here's why

Security is critical to industrial internet of things

Are digital manufacturing and IoT the new PB&J?

Dig Deeper on IIoT in manufacturing