FREE: The Industrial Internet

Published on: 

GE LAUNCHES THE DIGITAL POWER PLANT AT ITS MINDS + MACHINES EVENT

In September, GE hosted its fourth annual Minds + Machines event in San Francisco, California. More than 1,000 industry leaders, GE partners, software developers and users attended a three-day event focused on the Internet of Things (IoT) and how it can be used in power generation, manufacturing, aviation, healthcare and transportation.

The focus at the event was GE’s Predix software platform for the Industrial Internet introduced earlier this year. This cloud-based platform enables industrialscale analytics for asset performance management by providing a standard way to connect machines, data and people. When a power plant is fully connected with sensors, Predix is said to model its present and future state to increase efficiency across the entire system. GE calls this the Digital Power Plant (Figure 1).

The company claims that benefits can amount to $230 million over the lifespan of a new combined-cycle power plant (CCPP) and $50 million when applied to an existing CCPP. GE attributes these savings to lower fuel costs, increased performance, less unplanned downtime and lower emissions. The basic concept is that all equipment, systems and auxiliaries in a power plant can be connected to Predix. It acts as the operating system for all control systems, condition monitoring, maintenance systems and industrial software applications.

Alerts are created when a turbine is producing less than optimal power, for example. Similarly, anomalies are reported to the relevant employee; maintenance schedules are based on current conditions; and the plant is operated more efficiently.

Ganesh Bell[/caption]

Technicians and engineers will continue to operate gas and steam turbines with the added benefit of far more sensor information that is fed into an analytics engine to provide detailed insight on plant operations. “A typical gas power plant has more than 10,000 sensors,” said Ganesh Bell, Chief Digital Officer at GE Power & Water. “Only about two percent of that data gets analyzed today.”

All data transmitted to the analytics engine can be used in various kinds of modeling as well as simulation of a variety of what-if scenarios. Operators can better evaluate the risk of running components beyond their maintenance interval and discover how to eke out more efficiency from gas turbines, for example.

Digital twin

In particular, GE has put together what it calls the “Digital Twin.” This is an exact duplicate of a power plant including a digital replica of the gas turbines that make use of what Bell terms “digital technology and physics-based methods” to represent the actual assets in a plant.

Advertisement

The Digital Twin simplifies the gathering of data as a means of increasing performance and improving results gained through predictive maintenance. It can also be used to predict future performance or to investigate past failures.

Alternately, the Digital Twin can be used to determine how the plant would respond to a variety of simulated plant conditions. Being able to model temperature extremes would help increase the accuracy of the CCPP output.

And with wind and solar resources now being introduced onto the grid with greater frequency, the Digital Twin can be used to model plant operation in response to sudden stops and starts related to renewable resources going on or offline. This would enable plant executives to decide whether the gains from rapid cycling balance out the heavy toll exerted by the resulting wear and tear on turbomachinery and Heat Recovery Steam Generator (HRSG) components.

Two utilities are onboard with Predix, so far:

  • Exelon Generation is expected to implement Predix-powered software across its nuclear, gas and wind facilities. The utility hopes to increase power plant capacity
  • Public Service Enterprise Group is using Predix to upgrade multiple gas-powered plants. Predix is expected to be used in 7HA.02 gas turbines being installed at a new CCPP in New Jersey.

Predix technology is already being used in digital wind farms. This has resulted in the generation of 20% more power from an existing wind farm, said Bell (Figure 2).

Figure 2: The Digital wind farm already exists[/caption]

This is all part of the larger concept called the Industrial Internet, which integrates complex physical machinery with networked sensors and software. One goal is to assemble the necessary digital and physical components to create a smart grid that spans an entire region.

“Fully operational pockets of smart grids are possible within the next five years,” said Jim Connaughton, Executive Vice President of smart grid vendor C3 Energy at a parallel event in the same venue sponsored by The Economist magazine. “The world will spend two trillion dollars on grid upgrades in the next ten years, and all will come with sensors.”

In effect, utilities and grid operators are shifting from internal-facing systems to operating their assets in the cloud. That opens the door to networking far more systems and achieving higher efficiencies as an essential part of the Industrial Internet. “Twenty utilities once said they would never be in the cloud,” said Connaughton. “Now, nineteen of them are.”

Jeff Immelt, Chairman and CEO of GE, emphasized that the Industrial Internet is about “no unplanned downtime.” But it will also bring benefits such as faster product development. “Our Hclass turbine was designed in half the time it used to take,” said Immelt.

Industrial apps

To achieve such gains, however, gas turbine operators must deal with far more data than before. Further, they will have to become accustomed to dealing with a lot more software applications (apps). Immelt is hoping to have 5,000 apps available by the end of 2015 and expects GE’s Industrial Internet App Store to eventually grow to as many as half a million apps. GE has doubled its software revenues over the last four years — now about $6 billion. And the company is using Predix as the operating system for its own businesses. “We’ve generated about $400 million of internal productivity by using analytics internally,” said Immelt. “And predict that this will grow to about a billion in the near future.”