IN DEPTH: GE's vision for wind

The wind farms of the future may look much like today’s — but only from a distance.

To reach electricity production that is on a par with — or indeed cheaper than — fossil-fuel power stations, they will need to be different in most every respect from today’s, believes US industrial giant GE.

The company is in the vanguard of those that believe the key to cutting costs lies in combining innovative emerging turbine technologies with the digitally driven advanced automation made possible by the “industrial internet”.

GE envisages the wind farm as a “virtual power plant”.

Its ranks of giant-rotored, intelligent turbines will communicate with each other in real time about changing wind patterns. Predictive analytics and remote monitoring software will catch mechanical and electronic failures before they happen. And hundreds of thousands of data points will be pored over every second to smooth electricity as it courses onto the grid.

“Bigger is where we need to go in getting to optimised LCoE [levelised cost of energy],” says GE Global Research wind technology platform leader Mark Jonkhof. “But bigger is harder. And existing technology — the onshore, geared, double-fed induction generator turbine — as it is scaled up, pushes LCoE prohibitively high.

“The whole of our research and development programme is centred around creating larger machines that will be used to build highly integrated wind farms that can produce power at a lower overall cost.

“Our mission is subsidy-free wind. Innovation is how we are going to get there.”

This philosophy has led GE to pursue potentially disruptive technology projects, targeting everything from hyper-long blades made of architectural fabric, to 10MW-plus direct-drive generators powered by superconducting electromagnets and 100-metre-tall space-frame turbine towers equipped with built-in hoists.

Earlier this year, GE unveiled its new 2.5MW 2.5-120 “brilliant” turbine kitted out with “intelligent decisioning” software to crunch huge volumes of sensor-fed data and advanced for

ecasting algorithms — and “talk” with other turbines, service technicians and operators to boost energy output.

The design was trialled over nine months in a 1.6MW unit in Tehachapi, California, with an integrated 50kWh GE sodium-nickel-chloride battery. Another unit is also undergoing tests at a site in Wieringermeer, the Netherlands.

Groundbreaking ultra-large-scale turbine technology is the lifeblood of GE’s wind-power R&D unit. However, for long-term commerciality, the company is looking to “farm-level” LCoE to transform the industry’s economics.

Viewing a wind farm made up of hundreds of turbines as a single plant means cutting against the grain of turbine makers “selling power curves” to developers, says Jonkhof. This sales pitch ignores the fact that more than 80% of turbines sit behind others and do not get “pure” wind streams.

“From this perspective, a wind farm becomes more about park aerodynamics and acoustics and analytics of fleet data,” he states. “These turbines are seeing much different winds — more complex wake winds. The physics is fundamentally different.

“So we are exploring ways of controlling turbines differently — playing with tip-speed ratios, RPMs [revolutions per minute], torque level, optimising the blades for a turbulent wind regime — so that they make the most of the wind they get.”

Short-term energy storage, such as has been experimented with at Tehachapi, is seen as one of the final pieces in the jigsaw.

Small industrial-scale batteries make variable power sources like wind “predictable”, insofar as they can smooth the natural variability in production levels before the electricity is channelled on to the grid.

“The grid can handle variable demand, but add variable generation [and] it causes problems,” says Keith Longtin, GE’s wind-power product line leader. “But if your upstream turbines are communicating to your downstream turbines about how the wind is behaving, and you have a small amount of storage to shore things up, then you can provide a forecast [of output].

“This minimises balancing requirements on the grid — and gets rid of the problem of wasted energy due to curtailment, because you can always charge up your batteries.”

GE wants to take its holistic approach further, with wind-farm-to-wind-farm communication paving the way for automated control of a development’s voltages to the grid. This would create greater stability across regional networks by managing production from a number of projects.

“Multiple wind farms coming into the same grid will compete with each other on voltage,” says Longtin. “If these wind farms are talking to each other, we can determine what the optimal voltage levels are, which will strengthen a grid.”

The GE virtual power-plant vision is fast becoming reality. GE already has turbines intercommunicating on wind patterns at a site in Eastern Europe and, within a few months, will kick off a project that intelligently ties together several wind farms to manage output onto the grid.

Inside 12 months, the company expects to start shipping battery-fitted 1.7-100 and 2.5-120s machines into the US market.

“We've been in this business for 11 years [since acquiring Enron’s wind unit in 2002],” states Longtin. “We’ve taken our [fleet] availability from less than 85% to close to 98% during that time, we’ve doubled our capacity factor and we have scaled our manufacture from making ten turbines a week to 13 turbines a day. We’ve come a long way in a relatively short time.

“Wind farms as power plants: this is our next chapter.”