The internet of things is headed for a prime-time test as Invenergy LLC, North America’s largest independent, privately held renewable energy provider. It has signed on to deploy General Electric Co.’s Asset Performance Management reliability management software on 13 turbines at six gas-operated thermal plants in the U.S.

The software is based upon GE Predix, a platform as a service that is intended for companies that are instrumenting industrial machines, controllers and sensors.

Invenergy is adopting the software to perform predictive analytics on turbines that produce an operating capacity of 3,159 megawatts, or roughly enough to power a half-million homes. GE’s APM will help plant technicians to more accurately predict and diagnose equipment failures before they occur in order to avoid unplanned outages. The software also is also expected to help Invenergy better manage its assets to achieve the delicate balance among reducing cost, managing risk and improving availability and reliability.

The Predix-based platform has already paid some dividends during a six-month testing phase by identifying a a turbine journal bearing that was experiencing early-stage vibration. A failure could have damaged the turbine and resulted in an unplanned outage, but Invenergy technicians were able to detect the problem three months early and repair it during planned downtime.

GE says it has enlisted more than 26,000 developers to use Predix. As one of the world’s top producers of industrial engines, the company has been a champion of the industrial IoT, which it believes can unlock value in data that companies already have but can’t use. GE has said it expects its revenues from digital products and services to top $15 billion annually by 2020.

Locked-up value

In most industrial companies, “the instrumentation already exists. What doesn’t exist is the ability to get value out of the data,” said Niloy Sanyal, chief marketing officer of GE’s digital business focused on power and utilities. “Typically the customers doesn’t need a lot of new sensors. They need the analytics.” Only about 3 percent of the sensor data that industrial companies collect delivers any value, he said.

GE estimates that the average power generation system is unavailable 6 percent of the time from unplanned outages. “We know our application stack can cover 70 percent of most common failure modes in industrial power and we aim to get that to 90 percent,” Sanyal said.

Invenergy’s digital cooperation with GE began in 2015 with a pilot project to test early detection of equipment or system failures on selected units using both GE and non-GE equipment. The company is licensing APM as software as a service and not on a shared-savings basis. Sanyal said APM customers typically see a tenfold return on investment by applying analytics to existing data.

He compared GE’s platform approach to that of customer relationship management giant Salesforce.com Inc. “Salesforce doesn’t have very sophisticated analytics, but they’ve created an application layer where others are able to overlay their requirements,” he said. “The big trend is the convergence of operations tech and IT in the simple but elegant notion of a digital twin.”

The digital twin is a concept GE has been promoting that involves creating virtual digital replicas of products, processes and services that can be used for testing and experimentation before applying them to physical assets. They can also be customized to the unique characteristics of each asset. “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,” said John Magee, chief marketing officer at GE Digital in an interview on SiliconANGLE’s theCUBE last August. “You can optimize just for that asset.”

Image: GE