
The 12-m diameter generator shows its segmented design. For scale, there are two people in the top left of the photo.
Early last year, this magazine reported on a prototype generator that would use production methods developed for printed circuit boards to take cost and weight out of a permanent-magnet generator for direct-drive wind turbines (WPE&D, Feb 2012). The developer, Boulder Wind Power , recently reports that a 3 MW proof-of-concept generator has completed testing on a stand in Montana. “The 12-m diameter ring generator and power-conversion system were electrically segmented so we could run a variety of critical tests at full-load while minimizing the investment required for multiple design iterations of the electrical system,” says Boulder Wind Power’s Co-Founder and Chief Technology Officer Sandy Butterfield. “This electrical-segmentation strategy should also reduce downtime in the field. In the event of an isolated failure anywhere in the electrical system, a fraction of the system can be isolated and taken offline while most of the generator continues producing.”

The close-up of the steel backiron in the BWP proof-of-concept generator shows stiffening ribs that carry loading from the permanent magnets. Under normal operating conditions, analysis shows that the magnetic loads in the generator can approach 1,000,000 lb.
The proof-of-concept generator and test stand have been moved from Montana and reassembled at the company headquarters in Louisville, Colo. The company says the design may take up to 20% off the cost of wind-generated power.
Butterfield says part of the generator’s success so far comes from the accuracy of BWP’s electromagnetic analysis capability. “While many machine designs settle for the accuracy provided by 2D analyses, our team has leveraged 3D analysis throughout the design to better understand the various phenomena that are important in our axial-flux machine. Through a collaborative effort with our EM-FEA software providers, we are continually pushing the boundaries of 3D analytical methods, leading to what we believe are some of the largest and most detailed dynamic models in the industry.”
Butterfield says the air core, a key feature of the BWP technology, allows larger diameters. “By

A section of printed circuit-board stator conductors (red) sits in a magnetic field (blue) from the permanent magnets (light grey) placed on a section of back iron from the BWP generator rotor (dark grey).
eliminating the attractive forces between the generator rotor and stator, our engineering and design team does not have to rely purely on stiffness in the support structure to manage a small air gap between these two components. Free of the stiffness-driven design approach that most other drivetrain designs rely on, BWP can increase diameter to achieve higher power ratings more cost-effectively. A cost-effective 10 MW design may be readily achievable with this technology,” he says.
Butterfield won’t provide a weight because these figures will vary widely within a given platform depending on the rotor speed, which defines the torque requirements for the drivetrain. “The rotor diameter and the noise related tip-speed limitations that a wind turbine manufacturer selects to optimize their products for a particular wind regime and market will have a significant impact on this torque requirement. For example, the mass for any geared or direct drive design may vary as much as 20 metric tons when comparing a design optimized for a 3 MW turbine with a 101-meter rotor diameter to the same design optimized for a 3 MW turbine using a 120-m rotor. The mass of our 3-MW generator designed for a high-torque wind turbine is comparable to that of conventional low-torque 3-MW machine with gearbox. And I can say our generator will be much lighter than state-of-the-art, direct-drive generators.” At a recent Chinese wind energy conference, Butterfield says low-torque 3 MW drivetrains weighed about 40 metric tons.

Open circuit voltage measurements collected on the generator shows good correlation with the prediction formed by detailed 3D electromagnetic, finite-element analysis.
He acknowledges that 12 meters is a large diameter and perhaps a bit out of the ordinary. Most think wind turbines should look like they do now – small, compact drivetrains in nacelles that look like trailers. Enercon’s direct-drive designs, like BWP’s machine, represent a big change in appearance. “If the cost of energy is the issue then cost should be the primary metric, and the rest of the issues should be secondary considerations. Besides, there is no satisfactory 10-MW drivetrain design sufficiently lightweight and cost-effective. The industry will have to do something different,” he says.
Butterfield adds that superconducting generators, often considered an alternative for turbines in excess of 10 MW, are quite a ways off. “They’ll have to be proven on a small scale beforehand, and recent efforts have not been encouraging. Superconducting devices work best in steady state. They don’t like thermal cycling, and wind turbines are all about thermal cycling, shut downs, and start-ups. So it will be difficult to make a superconducting generator work on a tower bobbing around with constantly changing power levels,” he says.
Regarding the next step for the generator’s development, Butterfield will only say that the company is designing prototype generators for multiple clients’ turbines, but he won’t name the companies or when they might launch. WPE
Filed Under: Generators, News