Select Country
Follow Us

Full Steam Ahead: This Software Brain For Coal-Fired Power Plants Could Help Eliminate 500 Million Tons Of CO2

GE just picked up a head of steam and put it in the cloud.

It was just in April that GE Power, the GE business that makes power generation equipment, acquired the Boston-based machine learning and data analytics startup NeuCo Inc., which uses software and artificial intelligence to improve the efficiency of coal-fired power plants. These power plants burn coal to boil water and then use the steam to spin turbines and generate electricity. They are the most common source of electricity, providing around 40 percent of the world’s electric power.

NeuCo’s code and neural networks are now part of the world’s first “Digital Power Plant for Steam,” a set of digital technologies that can dramatically reduce greenhouse gas emissions by improving the operations and efficiency of coal-fired power plants. GE made the announcement at the Minds + Machines Europe conference, which is taking place today in Paris.

GE says it could help eliminate 500 million metric tons of greenhouse gas emissions if deployed at all existing coal-fired power plants globally. That’s the same as removing 120 million cars from the road – or every tenth car in the world – or covering Earth with carbon-trapping forest the size of Europe.

The technology is “equipment-agnostic,” which means it can optimize machines made by any manufacturer, says Ganesh Bell, GE Power’s chief digital officer. “Utilities will need to generate 50 percent more electricity than they do today by 2040 and meet ambitious goals to cut emissions set by COP21,” says Bell, referring to the United Nations climate change conference that took place in Paris last year. “Using data science, software and intelligent automation to optimize every aspect of the electricity value chain – from generation to delivery and consumption – will do both.”


Above: GE’s Ganesh Bell says data and software will help utilities “generate half as much electricity as they do today by 2040 and meet ambitious goals to cut emissions set by COP21.” Image credit: GE Reports. Top image: GE’s “ultra supercritical” steam turbine. The water pressure inside reaches 4,000 pounds per square inch, more than what’s exerted when a bullet strikes a solid object. Image credit: GE Power

Bell says the new Digital Power Plant for Steam (or DPP for Steam) software interprets data from more than 10,000 sensors located across the steam power plant and can increase efficiency by as much as 2 percent. It can make the boiler perform better – that’s where NeuCo’s neural nets come in – but also analyze coal quality to reduce fuel use and create what Bell calls a “digital twin” of the entire plant inside the cloud that can identify any gaps between actual and ideal performance.

GE also has a deep knowledge of steam turbines, another critical part of coal-fired plants. It acquired Alstom’s large global portfolio of steam turbines with its energy business last fall. “The combination of the physical strengths of our legacy Alstom steam technologies and GE’s digital systems can help our customers optimize their power plants and make them more efficient, lower emissions and reduce cost,” says Andreas Lusch, president of GE’s Steam Power Systems.

GE Reports editor Tomas Kellner sat down with Bell before the Minds + Machines Europe conference to discuss the technology. Here’s an edited version of their conversation.

RDK8 3Dview

Germany’s RDK8 is the world’s most efficient coal-fired power plant. Image credit: GE Power

GE Reports: Why did you develop the Digital Power Plant for Steam?

Ganesh Bell: When we digitized the wind farm last year, we allowed operators to build the most efficient wind farm and then run it in the most optimal fashion. The DPP for Steam is similar.

Our customers are ultimately in the business of generating and transmitting electrons. They are interested in doing it in the most reliable, effective and productive way from a mix of different fuel sources. But they are also facing a lot of challenges ranging from reducing emissions to plant longevity and fuel price. Our software is helping them to find the best solution and the best mix.

GER: How so?

GB: The average utility can generate power in many different ways, using wind, natural gas, coal or nuclear fuel. But the mix is different in every market, based on demand, regulations, fuel availability and other conditions. Our software will help our customers optimize all of their power generation options, select the best combination and make sure they hit their marks. In Europe, for example, they have to comply with very strict emissions.

GER: Why is coal still so important? Aren’t we trying to get rid of it?

GB: Coal is still a big part of the picture, whether you like it or not. People don’t like to talk about it, but the majority of power generation in many parts of the world is still from coal and it’s not going away. Here’s the challenge: our customers have sunk a lot of money in coal-fired plants and they expect to run them for decades. But in the post-COP21 world, they may not be able to because they won’t comply. We’re now giving them more options to run their assets by rejuvenating their operations with software.

GER: Give me an example.

GB: Take something simple like coal. Its quality varies in different parts of the world but also batch from batch. You have to optimize the power plant for the particular type you have. Today, you set the parameters once and you forget it. But our system is continuously learning and optimizing. This kind of technology is not available today. We figured that just this part of the digital power plant for steam could reduce fuel consumption by 4,400 tons of coal per year with the same megawatt of output in a single steam power plant.

GER: Does it matter who built the power plant?

GB: No. It doesn’t. It doesn’t matter whether the power plant has machines built by GE, Siemens or Skoda. The software is fully technology-agnostic. In fact, we can extend it beyond coal to plants burning other kinds of fuel like biomass and oil.

GER: What is the big picture here?

GB: Like I said, smart utilities are moving from power generation to optimizing their entire energy mix based on the fuel source. Already the power in your home can include electrons from coal, nuclear power as well as wind, depending on demand, weather, regulations and a number of other factors. We started by optimizing turbines at the asset level, then we moved to power plants across the fleet, and now the next logical step is optimizing all of power generation.

GER: How will you do that?

GB: That’s why GE developed Predix, the cloud-based operating system for the Industrial Internet. The Digital Power Plant for Steam, for gas, the Digital Wind Farm; all of these solutions are all built on the same Predix platform. With Predix, you now have an enterprise-wide view of all generation and all of data in a single secure cloud. That gives you new, interesting opportunities to solve hard problems.

GER: Like what?

GB: In Europe, for example, Predix allows you to manage how you integrate renewables better into the mix. It gives you new demand flexibility and helps you compensate quickly with traditional fuel sources when the wind stops blowing. We can now make all kinds of power coexist and work together.

Subscribe to our GE Brief