Building the “World’s Largest Wind Turbine Generator”

Introduction

This story is about the design, construction, and operation of the Mod 1 Wind Turbine Generator, along with some personal anecdotes. This was way back in 1979. It was not the first of its kind, and much larger WTGs have been built since then. But the WTG installed in Boone, North Carolina had the distinction of earning a citation in the “1980 Guinness Book of Records” as the “World’s Largest”, which it was at that time.

The Mod-1 WTG was designed, built, and operated by General Electric, Advanced Energy Products Department, located in King of Prussia, Pennsylvania, under contract with the Department of Energy, NASA Lewis Research Center in Cleveland, Ohio. NASA Lewis, (later renamed to honor John Glenn) had previously built a Mod-0 prototype WTG at its facility in Plumbrook, Ohio, and the Mod-1 was intended to be the next step toward more powerful designs.

History of Wind Turbine Generators

Windmills for grinding grain and pumping water go back to ancient times: examples were sail-cloth mills operating in the Greek Islands; more recently the iconic Dutch Windmill, and the familiar symbol of American farmland throughout the United States. The first serious attempt at constructing a large wind turbine for the generation of electrical power was the Smith-Putman project*. Smith-Putman was the first project to tackle one of the fundamental logistical challenges of such large installations: transport and erection of large blades and heavy equipment to the top of a mountain, and installation on the top of a tower.

*Wikipedia: The Smith–Putnam wind turbine was the world’s first megawatt-size wind turbine. In 1941 it was connected to the local electrical distribution system on Grandpa’s Knob in Castleton, Vermont.

It was not until the energy shortages of the 1970s prompted interest in renewable energy sources to meet the rising cost and consumption of electrical power, that WTGs were given renewed attention. The lead government agency was NASA’s Lewis Research Center. NASA-Lewis undertook to build a prototype Wind Turbine Generator at their nearby Plumbrook facility. This Mod-0 WTG was the model for subsequent megawatt-class installations, such as the Mod-1, located in Boone, North Carolina.

Design of the Mod-1 WTG

Participate in the Mod-1 Program was an engineer’s dream assignment. On most large projects, an individual engineer will have a singular, specialist role, among a large team of engineers from different organizations and disciplines. But here was a project where an engineer could participate in all phases, from beginning to end: from conceptual design; performance analyses; structural analyses; electrical engineering; software development; instrumentation; hydraulics; pneumatics; computer control;; manufacturing liaison; to learn about tribology and lubricants; to detailed design; logistics; on-site assembly; and finally, operations. There is nothing more rewarding for an engineer than to see his ideas and efforts in operation. Sometimes, you have to be patient. The New Horizons spacecraft, for example, took 9½ year after launch to reach its first destination.

Turbine Blade

The blade of a WTG has characteristics in common with an aircraft propeller or helicopter rotor. Like a propeller, it has twist, so that the regions of the span nearest the hub, where the relative wind velocity is least, presents the greatest angle of attack, with a reduced angle of attack toward the tip, where the relative wind velocity is highest. This means that all sections of the blade can operate close to their angle of attack for maximum lift. Whereas the propeller or rotor provides thrust when driven by an engine, the wind turbine blade extracts energy from the wind to power an electrical generator.

Turbine Blade Geometry

The wind energy that can be captured by a WTG is proportional to the area swept by the blade; therefore a 200-ft diameter blade will capture four times the energy as a 100-ft blade. However, the 200-foot blade will cost four times as much to build and will be only ¼ as stiff. These scaling factors favor building WTG blades of larger and larger diameter, providing the disadvantages of cost and reduced stiffness can be overcome.

Aerodynamically and structurally, a WTG blade is more like an aircraft wing than a propeller. The first step in the design of a WTG blade, as for an aircraft wing, is to choose an airfoil section. A great contribution to aircraft design by the National Advisory Committee on Aeronautics (NACA) in the early days of aviation was the codification and testing of a broad range of airfoil sections. A good airfoil choice for slow speed applications has its greatest thickness (camber) at a distance from the leading edge that is about 25% of the chord. That is where the center of lift will be located and where the greatest depth is available for location of the main spar. By contrast, the so-called “laminar flow” airfoil, which allowed the P-51 Mustang to attain higher speeds, has its maximum thickness at 50% of the chord.

The wing or blade, in generating lift, will have a non-productive component known as induced drag. To minimize induced drag, a high aspect ratio (span divided by average chord) is favored. An aspect ratio of 15 or more suits the flying regime of the best soaring birds, as it does sail planes and wind turbines. A low aspect ratio wing or blade, like the stubby wings of a crop-duster, will have greater induced drag, and for a saucer shape, with an aspect ratio of 1.0, the induced drag will be prohibitively large.

Choice of Blade Materials

For the Mod-1 WTG, there was a wide choice of materials of blade construction available: steel, aluminum, “fiberglass” (glass-fiber reinforced polymer), even wood. Studies even considered concrete, because of it low cost. Don’t laugh; concrete has been successfully used for the construction of ship hulls. The final choice is often made on the cost of construction. Remember, also that this was in the 1970s, and since then, there have been substantial improvements in the design and materials available, so that choices may differ today.

Aircraft wings are seldom made of steel. (The B-70, a tri-sonic, delta wing, is an extreme exception.) But the Boeing Company has some unique experience in the design and fabrication of hydrofoil patrol boats and ferries. The Boeing subcontractor for the hydrofoil blades was located in Tacoma, Washington, and used a common naval-grade steel, HY-80. The all-welded hydrofoil blade was close to simple monocoque construction, without the proliferation of ribs and stringers usually found in riveted, aluminum construction of an aircraft wing of that size. The sealed, welded construction kept water from entering and being trapped, and offered the unique opportunity of weld inspection by pressurizing the interior with inert gas and using a leak detector.

As a former aircraft structural design engineer, aluminum construction of the WTG blades would have been my sentimental favorite. One economic proposal was to copy the structural design of the wing of a B-24 Liberator bomber. Now that would have rung a bell with me, since I once lived in Chula Vista, California, near San Diego, where Consolidated-Vultee built many B-24s. The logo for the city of Chula Vista featured a depiction of the nearby memorial to glider pioneer, John Montgomery. On the top of Otay Mesa, where Montgomery practiced the art of hang-gliding in 1880, a B-24 wing has been erected as a pylon, its polished skin reflecting the sun for many miles.

A glass-epoxy composite construction has been used with success for large aircraft propellers. Hamilton Standard, a leading propeller contractor, was awarded a contract by NASA Lewis Research Center to attempt construction of a 60-foot, filament-wound blade. Filament winding was an attractive solution because the process can be automated to control precisely the spacing, direction, and number of layers of glass fibers. The difficulty is to ensure complete penetration of the epoxy in areas where there is an abrupt change in section. If successful, these blades might have been retrofitted to the Mod-0 WTG in Plumbrook, and would have been the leading contender for Mod-1 blades. However, Ham Standard was never able to prevent resin voids in the difficult area near the hub, where structural loads are the greatest. NASA also funded Kaman to develop a filament-wound glass-epoxy blade, 100 feet in length, that might have been retrofitted to the Mod-1 WTG.

NASA Lewis also contracted with Gougeon Brothers, Inc., for blades made of Douglas fir and birch. These blades operated successfully for 8,000 hours on the Mod-0A WTG in Kahuka, Hawaii. The blades were taken out of service because of failure of the metal shank, attributed to stress concentrations at a corrosion pit. Among the concerns with wooden blades are the effects of humidity, and the need for lightning protection.

Eventually, welded steel construction for the Mod-1 WTG blades was chosen, an economical choice because of the previous Boeing experience with hydrofoil blades. Modern WTG blades tend to be glass-reinforced polymer.

Hub Assembly

The hub assembly attaches the blades to the drive shaft, with bearings at the root of each blade to enable the blades to rotate in pitch. The hub assembly also incorporates a large bearing that supports the hub and drive shaft.

Associated with the hub, are the actuators and mechanisms for blade pitch control. For Mod-1, the force was provided by two of the biggest hydraulic actuators you can imagine. The actuators were capable of forcing the blades into a feathered position in a fault condition or when wind speed exceeded maximum allowable. Storage tanks holding high-pressure air were a back-up in the event of loss of hydraulic pressure. When in the feathered position, the actuators latched into a lock mechanism that was intended to prevent any further change in blade pitch.

Drive Shaft and Gearbox

The drive shaft, turning at the same rotational speed as the blades, sees the highest torque in the drive train. It is therefore robust and heavy. In the Mod-1 WTG, the drive shaft consisted of a heavy-walled steel pipe with industrial bolted flanges on each end. Attaching the flanges to the drive shaft is routinely done with a shrink fit, by machining slight tapers on both parts and slipping the heated flange on the cold shaft. How to remove the flange to make repairs? A clever solution is to machine a small groove in the taper of the coupling, and to apply hydraulic pressure through an external port. You can readily apply 10,000 psi hydraulic pressure with a hand pump, and the taper will cause the flange to pop right off. There are other trade secrets in designing large drive shafts that we learned from the Large Turbine Department of General Electric in Schenectady.

The gearbox increases the slow rotational speed of the blades to the high speed required by the electrical generator. Here, we had the services of Philadelphia Gear, tops in the large gearbox business, located right at our doorstep in King of Prussia. (The facility is no longer there.) They provided a custom design that utilized twin torque paths, the load shared equally by a quill shaft feature. The gearbox requires an active lubricating system, and a radiator to keep the oil cool. Between the gearbox and the generator, we installed a large industrial clutch, intended to prevent an overload. However, when the time came, the clutch let us down. When electrical power is applied at its terminals, the generator will function as a motor, applying torque, multiplied by the gearbox, to the low-speed shaft, hub and blades. More about the clutch later.

Nacelle

Nacelle is a word carried over from aircraft design, and usually refers to the appendage attached to the wing that houses the engine and its accessories. For a WTG, it includes the structure supporting the blade hub assembly, gearbox generator, as well as the enclosure, or fairing that protects the machinery from the elements.

The Mod-0 WTG had a nice aerodynamically-shaped nacelle. The senior designer that was given the task of designing a nacelle fairing for Mod-1 came up with a concept that somewhat resembled the nose of a Spitfire. Not surprising, since he had served his apprenticeship at Supermarine. However, a WTG, unlike a Spitfire, is not expected to experience airspeeds of 400 mph, and the aerodynamic drag forces are not significant when compared to the induced drag forces on the blade. A rectangular box would do, and was cheaper to manufacture. We went to the fabricator in Oklahoma who built enclosures for General Electric auxiliary power gas turbines. They custom-designed and built an enclosure that fitted our needs exactly. It turned out to be a serendipitous find, because the same company helped us later recover from a fabrication problem on the tower. The flat-roofed enclosure also proved to be a safe platform from which we could service the anemometer installations.

Pintle Assembly

“Pintle” is one of those old Anglo-Saxon words that have been adapted to modern usage. Its original meaning was, literally, “penis.” It has come to us through nautical usage for a pinned connection that allows rotation, such as the pin and gudgeon connection of a ship’s rudder. Early windmills had a short vertical shaft, called a pintle, allowing the assembly holding the blades to yaw with the wind, and the word has been adapted in modern wind turbine jargon, for the assembly that allows (or forces) the nacelle to yaw.

For the Mod-1 WTG, the pintle structure was a massive steel weldment, capable of withstanding severe operating loads, including the forces of hurricane winds. To minimize friction resistance required a large-diameter yaw bearing. Here, we turned to the experts in large bearings, those who had designed and fabricated the yaw bearing for the 250-ton crane installed in the Philadelphia Naval Shipyard. For the Mod-1 application, roller bearings were called for, with four-inch diameter rollers, angled so they could withstand both vertical and horizontal loading.

The nacelle was driven in yaw to face the prevailing wind by two very large, Brtitish hydraulic motors, which engaged the teeth of a large internal gear, below the yaw bearing. These four-inch teeth required a special lubricant with extreme “tenacity”, black and thick as molasses, and very expensive.

For the pintle structure, we turned to another General Electric Department with expertise in large structures: Dominion Bridge in Montreal. Dominion has experience in design and fabrication of massive weldments for hydroelectric and steel mill installations. A large scroll assembly for a hydroelectric generator can be as large as fifty feet in diameter. In their shops, Dominion has lathes that can turn an eight-foot diameter shaft, a milling machines with a 100-foot long bed, and heat treatment furnaces that can accommodate giant welded assemblies.

Another valuable consultant was the Civil Engineering Department of Lehigh University. Members of the Leigh staff have expertise in design of large steel weldments such as the bridges for the Intersate Highway system. Their laboratory has the capability of performing full-scale destructive tests on modern bridge structures, with welded joints between very thick, high-strength steel plates. We followed Lehigh guidance on the design of all critical weld joints.

Tower

If you have large blades, you need a tall tower. The wind speed is reduced by the ground effect, so that its speed increases with distance above the local surface. Therefore, it is advantageous for a wind turbine to have a tall tower. However, the cost of the tower increases by the square of the height, so that a 400-ft tower costs four times that of a 200-ft tower.

Design of the tower is primarily governed by the structural dynamics of the system. Due to the natural gradient of the wind near the ground, the wind speed is greatest when a blade is at the top of the rotation, and is least when a blade passes the bottom point, nearest the ground. In addition, a blade passes through interrupted flow and greater wind turbulence each time it passes the tower. The tower effect will be greatest if the blades are positioned downwind of the tower. The resulting dynamic load will be roughly sinusoidal, with a frequency that is equal to the rotational speed, multiplied by the number of blades. The tower structure must be designed so that its fundamental frequency is not in resonance with the driving frequency. There are two approaches.: design the tower to be stiff, with a resonant frequency above the driving frequency, or design a soft tower with a fundamental bending frequency below that of the driving frequency. In some cases, the second vibration mode of the tower may also need to be examined. The stiff tower approach is heavier and more costly, while the soft tower will be in resonance for a brief time each time the driving frequency passes through the fundamental bending frequency of the tower as it starts up or shuts down. Soft towers can be a simple monopole, the approach favored by most modern WTGs, but for the Mod-0 and Mod-1 WTGs, the more conservative stiff tower approach was chosen,

When the REA brought electrification to American farms, the transmission towers borrowed their design of the truss tower construction from the earlier farm windmills, and even the tallest, long distance transmission lines with 200,000 VDC, have retained the four-legged truss as the most economical tower construction. So it is not surprising the first modern wind turbine generators should borrow their tower design, from, (guess what?), high-voltage electrical transmission lines.

The truss tower will naturally be cranked, with the upper section restricted in width to provide clearance with the blades at their maximum deflection. Maximum bending of the blades toward the tower occurs at maximum wind speed for an upwind rotor. Below the point where the blade tip is closest to the ground, the legs of the tower can be splayed out for maximum stiffness. This gives a resemblance to an electrical transmission tower, where the upper section is restricted in width to provide sufficient clearance with suspended wires and their insulators.

The struts and cross members will have greater stiffness if they are tubular, which lends itself to all-welded construction, rather than hundreds of bolted connections. We subcontracted the detail design and construction to a steel company in Albany, New York, who had built similar large tubular structures. Here’s where our troubles began. We ordered the tubing from US Steel, the only domestic source for such large diameter tubing. US Steel completed our order, just before closing the door on their mill forever. Our order was loaded on a flatcar that was immediately covered in snow and lost in the rail yards. Meanwhile the owner of our steel fabricator in Albany, fired the president and chief engineer for incompetence and declared bankruptcy. We were fortunate enough to locate our steel tubing under the snow, buy it from the bankruptcy court, and ship it to another fabricator. Our fabrication subcontractor in Oklahoma came to the rescue, checked all the geometry, and spread out the entire tower assembly on the ground before welding.

The use of very large bolts in the assembly took some getting used to by our crew of aerospace structural engineers. Instead of small bolts to MS (Military Standard) or NAS (National Aerospace Standard) fasteners, we were guided by the Blue Book of the American Institute for Steel Construction. If you could not manage an eight-foot long torque wrench for these large fasteners, you resorted to a special wrench and sledgehammer. The objective was the same, to preload the fastener to its maximum capability, indicated by the onset of yielding. When you observed a sudden rise in torque, or no further rotation from your hammer blows, you knew that the maximum preload had been achieved. Bolt preloading extends fatigue life.

System Dynamics

I have mentioned how important it is to have a good understanding of the structural dynamics of a WTG. Operating almost continuously, the installation is like one giant fatigue-testing machine. The operator expects to have the WTG operate unattended, with minimum maintenance for twenty-five years or more. The number of fatigue cycles completed in that times may be orders of magnitude greater than the tens of millions of cycles for which fatigue design data may be available. For that reason, it is essential to know or to be able to predict accurately, the structural dynamics of the system. Designing the tower to separate its fundamental modal frequencies from the dynamic force input of the rotor, to avoid resonance, is essential. The two most important modes are tower bending, and torsional stiffness. The torsional mode was determined partially by the stiffness of the load path through the yaw motors and their geared connection with the pintle assembly, a connection that did not lend itself to finite element analysis in a way that the strictly structural elements did. It was important to obtain an empirical verification.

Fortunately, there was an opportunity to make a measurement during a trial mating of the nacelle and pintle assemblies in one of our facilities in Philadelphia. This type of test is not uncommon for large structures, but is very rare in the aerospace industry. If you were to tell a spacecraft Project Engineer that you wanted to take the biggest sledge hammer you could find, and strike his baby several times with as much force as you could muster, he would probably ban you from the spacecraft assembly laboratory. For our hammer test, we found a length of timber, 1 ½ feet square, and 12 feet long, and suspended it horizontally from the overhead shop crane. We positioned the timber so that when swung like a pendulum, it would strike a glancing blow to a protruding feature on the nacelle structure, aimed to promote a response in torsion. The attached pintle structure was firmly anchored to the shop floor. The timber was controlled by a tag line, held by an engineer located on the balcony overlooking the shop area. The torsional frequency was recorded by accelerometers. It took some skill to maneuver the timber to get it swing with a good amplitude, while steering it to impact the nacelle end-on. Several of us took a hand at it, to see who was best at “ringing the bell”.

The assembly of the Mod-1 WTG provided an appreciation for the tribulations of the crew who erected the Smith-Putnam WTG many years ago. The various assemblies of the WTG were transported from around the country to Howard’s Knob, a municipal park, near Boone, North Carolina. The winding road up the mountain was lined with trees, so that delivery of the blades, the longest items, was accompanied by a crew with chainsaws, to clear the trees that interfered. Delivery and installation of the blades was the most dramatic event of assembly operations. For transport from Tacoma, the two blades were supported on an auxiliary trailer that was steered by a prone driver, who coordinated his steering with the tractor driver up front. When imminent arrival of the blades was announced, I drove out on the highway near Boone to witness their passage. I almost missed them, as they whizzed by at high speed.

Transport of Mod-1 WTG Blades From Tacoma to Boone

Our GE Crew Admires the Blade Delivery Transporter

Installation of the blades called for the skills of experienced “high-iron” workers, those who specialize in skyscraper girder erection. Few know that this line of work is the purview of members of the Mohawk tribe. Mohawk were given the right to participate in construction of the Quebec Bridge that touched on their reservation. They were so comfortable walking the girders, that it has become their tradition ever since. Most New York City skyscrapers owe their success to the skills of Mohawk iron workers. The crane that lifted the WTG blades into position in Boone was directed by a Mohawk, who stood on the horizontal blade to bolt the blade flange to the hub assembly, and then torqued the nuts. When he had finished, he calmly walked the length of the blade to remove the lifting sling, and stepped off the tip to ride the crane hook to the ground

Site Hazards

Assembly and operating experiences provided some physical hazards that were uncommon to design engineers sitting at their desks in an office environment. Access to equipment mounted in the nacelle required operation of a cable-supported lift cage, a common safety provision used by construction workers and window-washers. In the event of an electrical failure, or the cable pulleys becoming jammed with ice (it happened), we provided an escape device: a sling positioned under the arms, attached by cable to an inertia brake that allowed the cable to be paid out at slow speed, similar to the inertia devices associated with aircraft seat harnesses. The protocol was to snap the cable hook to the cage railing, and to step off into space for a controlled ascent at about the speed of a parachute landing. It was my duty, as Site Manager, to demonstrate the lift and escape sling operation to every newcomer to the site who needed access to the nacelle, including the occasional VIP and their female companion.

The best views of the surrounding countryside were from the roof of the nacelle, reached through an access panel, roughly 120 feet above ground level, on top of a 450-foot hill. There were non-skid walkways painted on the flat roof that guided you to spots were you might safely hook your safety belt, such as around one of the poles supporting twin anemometers. The view of the countryside, high in the Piedmont region of the Appalachian Mountains, was spectacular. No, I never ventured on to the roof while the blade was turning, that would have been the ultimate thrill. I remember sitting on the edge of the roof, straddling one of the met poles, to tighten some loose bolts, where, looking down, a solid bank of clouds slowly drifted by provided a comforting background. But when the clouds suddenly parted, I was shocked back to the scary reality of solid ground, 120 feet below me.

There were occasion when I had to climb the truss work of the tower, to the dismay of our Quality Assurance Engineer, who acted as Safety Manager. To measure clearances of the blades with the tower, we rigged an array of wooden dowels on the tower at the point of closest approach. When the blades were suddenly feathered, the dowels clipped by the blades showed the clearance under maximum bending conditions. It was a crude, empirical verification of analysis.

Our site crew included electrical and computer engineers who often marched to a different drummer than the mechanical types. For safety reasons, all tests conducted on site required a written test plan, approved by the Program Manager, Quality Assurance Engineer, and Site Manager. But when the software engineers made changes to the operating code that “did not affect the hardware”, the programmers thought they had free reign. On one occasion, when our mechanical engineering crew was working in the vicinity of the drive shaft and the giant pitch actuators, the system suddenly took the blades out of their normal feathered and parked position, and easily broke the lock mechanism, sending the heads of ¾-inch bolts zinging around like bullets in a Western gunfight. “Not our doing,” said the software engineer, “we’re just making software changes, and the Program Manager said we don’t need a test plan.” That episode convinced me that you can be injured by faulty software as easily as by faulty hardware. It was a lesson I took to heart, and years later, I included in the Configuration Management Plan for Space Power Engineering at Lockheed Martin, the requirement that all such software be documented, and that changes to software be approved by the Configuration Management Board, with the same procedures and documentation required of mechanical drawings and specifications.

Site Operations

When all assembly operations had been completed, we began methodical steps to begin operations. The first time that the blades turned was an exciting event. A crowd of curious folks from town had gathered at the barricade near the base of the tower. When each blade passed the tower it gave a loud “woosh”, and soon we were producing a rapid series of ‘woosh-wooshes” that scattered the crowd down the hill.

The final step was to connect with the local utility. After some initial success, that step did not go well

A fault on the utility side of the electrical interconnect, caused the current to reverse direction, and the generator acted as a motor, attempting to drive the entire drive train and rotor backwards. The clutch on the high-speed shaft slipped and soon overheated, causing a fire. The first indication of a malfunction was the smoke from the burning clutch, billowing from the nacelle vents. Before the machine could be turned off, an overload caused a mechanical failure in the hub assembly. Without funding to make repairs, the Mod-1 WTG was abandoned. In accordance with the municipality, the installation was dismantled and the park returned to its original condition.

Postscript to Boone WTG Operation.

After failing to negotiate a contract with Hilo Electric, in Hawaii, for installation of Mod-2 WTGs, General Electric withdrew from the wind turbine business and closed the Advanced Energy Products Department that had produced the Mod-1. That was more than forty years ago and it is only recently that General Electric has returned as a primary player in the wind turbine business.

Modern WTG Design

In the eighty years since Smith –Putnam and the forty years since Mod-1, there have been many advances in WTG technology, while some things have remained the same. The principles of operation have not changed, nor the breakdown of assemblies to blade, hub assembly, drive train, gearbox, generator, pintle assembly and tower. Some of the advances in aerodynamics have been brought about by advances in analytical capability, such as the use of Computational Fluid Dynamics (CFD) which has confirmed the validity of data that was previously empirical. Finite Element Analysis (FEA) for the analysis of structural loading, deflections, stresses, statics and dynamics was available for Mod-1 in the 1970s, but today’s computers are faster and the codes more elaborate, allowing the analysis of more and more complex structures. As an example the FEM (Finite Element Model) used to design the B-70 wing in 1959 had 1235 nodes and taxed the capacity of existing computers to invert the stiffness matrix. Today it is not uncommon to create a mesh with tens of thousands of nodes, just to ensure that the solution has converged. One of the primary engineering challenges that remains from the days of Smith-Putnam is logistics, the method and equipment required to transport larger and larger blades and taller towers.

GE’s latest WTG model is called Cypress. The transport solution for Cypress blades that are over 500 feet in diameter is to make them in two pieces, joined together on site.

--

--

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store