COMPONENTS AND OPERATIONThis is an Acrobat version of the special issue of theBonus-Info 1998 newsletter. It was made for printing onyour printer. You can also read it from the screen. Use thethumbnails (left) to navigate, and the magnification tool

(below) to make the text fit to your screen.

Table of contents page 4

BONUSENERGY A/S

¨

THE WIND TURBINE

BONUS INFOT H E N E V E R E N D I N G S T O R Y

Autumn 1999

THE WIND TURBINE COMPONENTS AND OPERATION

Special Issue

BONUS-INFO is a newsletter forcustomers and business associates ofthe Bonus Energy A/S. This news-letter is published once or twice ayear.

The first number came out in 1998,and the newsletter has now beenpublished in four issues.

Each number has included anarticle on the components and opera-tion of the wind turbine. We havereceived many suggestions andrequests that these articles should bereprinted and published as a specialsingle issue.

Bonus is pleased to have herebyfulfilled this request with thepublication of this special issue.

Author:Henrik Stiesdal

Responsible under the press law

Lay-out/ Production:Claus Nybroe

Translation:John Furze, Hugh Piggott

Autumn 1999

BONUS ENERGY A/SFabriksvej 4, Box 1707330 BrandeTel.: 97 18 11 22Fax: 97 18 30 86E-mail: bonus@bonus.dkWeb: www.bonus.dk

4

THE WIND TURBINE COMPONENTS AND OPERATION

The Aerodynamics of the Wind Turbine 5Basic Theory ¥ The aerodynamic profile ¥ The aerodynamicsof a man on a bicycle ¥ Wind turbine blades behave in thesame way ¥ Lift ¥ The change of forces along the blade ¥What happens when the wind speed changes ¥ The stallphenomena ¥ Summary

The Transmission System 11The hub ¥ Main shaft ¥ Main Bearings ¥ The clamping unit ¥The gearbox ¥ The coupling

The Generator 15Direct current (DC) ¥ Alternating current (AC) ¥ Threephase alternating current ¥ Induction and electromagnetism ¥The wind turbine generator as a motor ¥ Generator operation¥ Cut-in ¥ Closing remarks

Control and Safety Systems 20Problem description ¥ The controller ¥ Hydraulics ¥ Tip bra-kes ¥ The mechanical brake

The three bladed rotor is the mostimportant and most visible part of thewind turbine. It is through the rotorthat the energy of the wind is transfor-med into mechanical energy that turnsthe main shaft of the wind turbine.

We will start by describing why theblades are shaped the way that they areand what really happens, when theblades rotate.

BASIC THEORYAerodynamics is the science and studyof the physical laws of the behavior ofobjects in an air flow and the forces thatare produced by air flows.

The front and rear sides of a windturbine rotor blade have a shape roughlysimilar to that of a long rectangle,with the edges bounded by the leadingedge, the trailing edge, the blade tip andthe blade root. The blade root is bolted tothe hub.

The radius of the blade is the distancefrom the rotor shaft to the outer edge ofthe blade tip. Some wind turbine bladeshave moveable blade tips as air brakes,and one can often see the distinct lineseparating the blade tip component fromthe blade itself.

If a blade were sawn in half, onewould see that the cross section has astreamlined asymmetrical shape, with theflattest side facing the oncoming air flowor wind. This shape is called the bladeÕsaerodynamic profile

THE AERODYNAMIC PROFILEThe shape of the aerodynamic profile isdecisive for blade performance. Evenminor alterations in the shape of theprofile can greatly alter the power curveand noise level. Therefore a blade desig-ner does not merely sit down and outlinethe shape when designing a new blade.The shape must be chosen with great careon the basis of past experience. For thisreason blade profiles were previouslychosen from a widely used catalogue ofairfoil profiles developed in wind tunnelresearch by NACA (The United StatesNational Advisory Committee for Aero-nautics) around the time of the SecondWorld War.

The NACA 44 series profiles were usedon older Bonus wind turbines (up to andincluding the 95 kW models).

This profile was developed during the1930Õs, and has good all-round proper-ties, giving a good power curve and agood stall. The blade is tolerant of minorsurface imperfections, such as dirt on theblade profile surface.

The LM blades used on newer Bonuswind turbines (from the 150 kW models)use the NACA 63 profiles developedduring the 1940«s. These have slightlydifferent properties than the NACA 44series. The power curve is better in thelow and medium wind speed ranges, butdrops under operation at higher windspeeds. Likewise this profile is moresensitive with regard to surface dirt.This is not so important in Denmark,but in certain climate zones with littlerain, accumulated dirt, grime and insectdeposits may impair and reduceperformance for longer periods.

The LM 19 blades, specificallydeveloped for wind turbines, used on theBonus 500 kW, have completely newaerodynamic profiles and are thereforenot found in the NACA catalogue.These blades were developed in a jointLM and Bonus research project someyears ago, and further developed andwind tunnel tested by FFA (The Aero-dynamic Research Institute of TheSwedish Ministry of Defence).

THE AERODYNAMICS OF A MAN ON A BICYCLETo fully describe the aerodynamics of awind turbine blade could appear to berather complicated and difficult to under-stand. It is not easy to fully understandhow the direction of the air flow aroundthe blade is dependent on the rotation ofthe blade. Fortunately for us, air con-stantly flows around everyday objectsfollowing these very same aerodynamiclaws. Therefore we can start with theaerodynamics of an air flow that most ofus are much more familiar with: A cycliston a windy day.

The diagrams (next page) show acyclist as seen from above. The diagramsare perhaps rather sketchy, but with agood will one can visualize what they

THE AERODYNAMICS OF THE WIND TURBINE

Blade profiles

Blade tip

Lead

ing

edge

Tra

iling

edg

e

Blade root

NACA 44

NACA 63

Hub

The different components of a wind turbine blade

5

6

represent. The diagram (A) on the left, illustrates a situation, during which acyclist is stationary and can feel a sidewind ÒvÓ of 10 meters per second (m/s)or roughly 22 mph (this is known as afresh breeze). The wind pressure willattempt to overturn the cyclist. We cancalculate the pressure of the wind on thewindward side of the cyclist as roughly80 Newton per square meter of the totalside area presented by the cyclist againstthe wind. Newton, or N for short, is theunit for force used in technical calculati-on. 10 N is about 1kg/force (Multiply by0.2248 to obtain lbf.). The direction ofthe force of the wind pressure is in linewith the wind flow. If we consider that anormal sized cyclist has a side areafacing the wind of about 0.6 squaremeters, then the force F from thepressure of the wind will be 0.6 x 80 N =app. 50 N/m2.

In the center drawing (B) our cyclisthas started out and is traveling at a speedÒuÓ of 20 km/hour, equivalent to about 6meters/second, still with a side wind ÒvÓof 10 m/s. We can therefore calculate thespeed of the resulting wind ÒwÓ strikingthe cyclist, either mathematically or bymeasurement on the diagram as 12 m/s.This gives a total wind pressure of100 N/m2. The direction of the wind pres-sure is now in line with the resultingwind, and this will give a force ÒFÓ on thecyclist of about 60 N/m2.

In the right hand drawing (C) theforce of the wind pressure ÒFÓ is nowseparated into a component along thedirection of the cyclistÕs travel and intoanother component at a right angle to thedirection of travel. The right angledforce ÒFvÓ will attempt to overturn thecyclist, and the force ÒFmÓ along the axisof travel gives a resistance that slows

down the cyclistÕs forward motion. The size of ÒFmÓ is about 30 N/m2. This isthe resistance force that the cyclist mustovercome. A beginner, unused to cycling,may wonder why the wind has changeddirection and a head wind is felt onreaching speed. This beginner might wellask Ò How can it be that I felt a side windwhen I was at rest and standing still,could the wind have possibly changed itsdirection? Ò But no, as any experiencedcyclist unfortunately knows, head wind isan integral component of movementitself. The wind itself has not turned. Thehead wind is a result of speed, the faster

one travels the more wind resistance oneexperiences. Perhaps, as a famous Danishpolitician once promised his voters, thatif elected he would insure favorable tail-winds on the cycle-paths, things maychange in the future. However we othershave learnt to live with the head windsresulting from our own forwardmovement, whether we run, cycle or goskiing.

WIND TURBINE BLADES BEHAVE IN THE SAME WAY Returning to the wind turbine blade, justas in the situation for the cyclist, we canobserve the aerodynamic and force

diagrams in two different situations,when the wind turbine is stationary andwhen it is running at a normal operationalspeed. We will use as an example thecross section near the blade tip of aBonus 450 kW Mk III operating in awind speed ÒvÒ of 10 m/s.

When the rotor is stationary, asshown in drawing (A) below, the windhas a direction towards the blade, at aright angle to the plane of rotation, whichis the area swept by the rotor during therotation of the blades. The wind speed of10 m/s will produce a wind pressure of80 N/m2 of blade surface, just like theeffect on our cyclist. The wind pressure isroughly in the same direction as the windand is also roughly perpendicular to theflat side of the blade profile. The partof the wind pressure blowing in thedirection of the rotor shaft attempts tobend the blades and tower, while thesmaller part of the wind pressureblowing in the direction of the rotationof the blades produces a torque thatattempts to start the wind turbine.

Once the turbine is in operation andthe rotor is turning, as is shown in the

center diagram (B), the blade encountersa head wind from its own forward movement in exactly the same way as thecyclist does. The strength of head windÒuÓ at any specific place on the bladedepends partly on just how fast the windturbine blade is rotating, and partly howfar out on the blade one is from the shaft.In our example, at the normal operatingspeed of 30 rpm, the head wind ÒuÓ nearthe tip of the 450 kW wind turbine isabout 50 m/s. The ÒmeteorologicalÓ windÒvÓ of 10 m/s will thus give a resultingwind over the profile of about 51 m/s.

This resulting wind will have aneffect on the blade surface with a force

Air flow around a man on a bicycle

Airflow around a blade profile, near the wing tip

A CB

v v

F F

w

u

Fm

Fv

A B C

v vw

u

F F Fa

FdPlane of rotation Plane of rotation

of 1500 N/m2. The force ÒFÓ will not bein the direction of the resulting wind, butalmost at a right angle to the resultingwind.

In the drawing on the right (C) theforce of the wind pressure ÒFÓ is againsplit up into a component in the directionof rotation and another component at aright angle to this direction. The forceÒFaÓ at a right angle to the plane of rota-tion attempts to bend the blade backagainst the tower, while the force ÒFdÓpoints in the direction of rotation andprovides the driving torque. We maynotice two very important differencesbetween the forces on the blade inthese two different situations and forceson the cyclist in the two correspondingsituations. One difference is that theforces on the blade become very largeduring rotation. If vector arrows illu-strating the forces in the diagrams weredrawn in a scale that was indicative of thesizes of the different forces, then thesevector arrows of a wind turbine in opera-tion would have been 20 times the size ofthe vector arrows of the same windturbine at rest. This large difference isdue to the resulting wind speed of 51 m/sstriking a blade during operation, manytimes the wind speed of 10 m/s when thewind turbine is at rest. Just like thecyclist, the blade encounters head windresulting from its own movement,however head wind is of far greaterimportance on a wind turbine blade thanfor a cyclist in motion.

The other important differencebetween a wind turbine blade and acyclist is that the force on the blade isalmost at a right angle to the resultingwind striking the profile. This force isknown as the lift and also produces asmall resistance or drag. The direction ofthis lift force is of great importance. Acyclist only feels the wind resistance as aburden, requiring him to push down extrahard on the pedals. However with a windturbine blade this extra wind resistancewill act as a kind of power booster, atleast in the normal blade rotational speedrange. The reason for this difference isdue to the blades streamlined profile,which behaves aerodynamically com-pletely differently as compared to theirregular shaped profile of a man on abicycle. The wind turbine blade experi-

ences both lift and drag, while a cyclistonly experiences drag.

LIFTLift is primary due to the physical pheno-mena known as BernoulliÕs Law. Thisphysical law states that when the speed ofan air flow over a surface is increased thepressure will then drop. This law iscounter to what most people experiencefrom walking or cycling in a head wind,where normally one feels that thepressure increases when the wind alsoincreases. This is also true when one sees

an air flow blowing directly against asurface, but it is not the case when air isflowing over a surface.One can easily convince oneself that thisis so by making a small experiment. Taketwo small pieces of paper and bend themslightly in the middle. Then hold them asshown in the diagram and blow inbetween them. The speed of the air ishigher in between these two pieces ofpaper than outside (where of course theair speed is about zero), so therefore thepressure inside is lower and according toBernoulliÕs Law the papers will besucked in towards each other. One wouldexpect that they would be blown awayfrom each other, but in reality theopposite occurs. This is an interestinglittle experiment, that clearly demonstra-tes a physical phenomenon that has acompletely different result than what onewould expect. Just try for yourself andsee.

The aerodynamic profile is formedwith a rear side, that is much more curvedthan the front side facing the wind.Two portions of air molecules side byside in the air flow moving towards theprofile at point A will separate and passaround the profile and will once again beside by side at point B after passing the

profileÕs trailing edge. As the rear side ismore curved than the front side on a windturbine blade, this means that the airflowing over the rear side has to travel alonger distance from point A to B thanthe air flowing over the front side.Therefore this air flow over the rear sidemust have a higher velocity if these twodifferent portions of air shall be reunitedat point B. Greater velocity produces apressure drop on the rear side of theblade, and it is this pressure drop thatproduces the lift. The highest speed isobtained at the rounded front edge of the

blade. The blade is almost suckedforward by the pressure drop resultingfrom this greater front edge speed. There is also a contribution resultingfrom a small over-pressure on the frontside of the blade.

Compared to an idling blade theaerodynamic forces on the blade underoperational conditions are very large.Most wind turbine owners have surelynoticed these forces during a start-up ingood wind conditions. The wind turbinewill start to rotate very slowly at first,but as it gathers speed it begins toaccelerate faster and faster. The changefrom slow to fast acceleration is a signthat the bladeÕs aerodynamic shapecomes into play, and that the lift greatlyincreases when the blade meets the headwind of its own movement. The fastacceleration, near the wind turbineÕsoperational rotational speed places greatdemands on the electrical cut-in systemthat must Òcapture and engage Ò the windturbine without releasing excessive peakelectrical loads to the grid.

THE CHANGE OF FORCES ALONG THE BLADEThe drawings previously studied, mainlyillustrate the air flow situation near the

An experiment with BernoulliÕs Law

Blow!

A

B

Air flow around an aerodynamic profile

7

8

blade tip. In principle these sameconditions apply all over the blade,however the size of the forces and theirdirection change according to theirdistance to the tip. If we once again lookat a 450 kW blade in a wind speed of10 m/s, but this time study the situationnear the blade root, we will obtainslightly different results as shown in thedrawing above.

In the stationary situation (A) in theleft hand drawing, wind pressure is still80 N/m2 . The force ÒFÓ becomes slightlylarger than the force at the tip, as theblade is wider at the root. The pressure isonce again roughly at a right angle to theflat side of the blade profile, and as theblade is more twisted at the root, moreof the force will be directed in the direc-tion of rotation, than was the case at thetip.

On the other hand the force at the roothas not so great a torque-arm effect inrelation to the rotor axis and therefore itwill contribute about the same force tothe starting torque as the force at the tip.

During the operational situation as shown in the center drawing (B),the wind approaching the profile is onceagain the sum of the free wind ÒvÓ of10 m/s and the head wind ÒuÓ from theblade rotational movement through theair. The head wind near the blade root ofa 450 kW wind turbine is about 15 m/sand this produces a resulting wind ÒwÓover the profile of 19 m/s. This resultingwind will act on the blade section with aforce of about 500 N/m2.

In the drawing on the right (C) forceis broken down into wind pressureagainst the tower ÒFaÓ, and the bladedriving force ÒFdÓ in the direction ofrotation.

In comparison with the blade tip theroot section produces less aerodynamic

forces during operation, however more ofthese forces are aligned in the correctdirection, that is, in the direction of rotation. The change of the size anddirection of these forces from the tip intowards the root, determine the form andshape of the blade.

Head wind is not so strong at theblade root, so therefore the pressure islikewise not so high and the blade mustbe made wider in order that the forcesshould be large enough. The resultingwind has a greater angle in relation to theplane of rotation at the root, so the blademust likewise have a greater angle oftwist at the root.

It is important that the sections ofthe blade near the hub are able to resistforces and stresses from the rest ofthe blade. Therefore the root profile isboth thick and wide, partly because thethick broad profile gives a strong andrigid blade and partly because greaterwidth, as previously mentioned, isnecessary on account of the resultinglower wind speed across the blade. Onthe other hand, the aerodynamic behaviorof a thick profile is not so effective.

Further out along the blade, the profilemust be made thinner in order to produceacceptable aerodynamic properties, andtherefore the shape of the profile at anygiven place on the blade is a compromisebetween the desire for strength (the thickwide profile) and the desire for goodaerodynamic properties (the thin profile)with the need to avoid high aerodynamicstresses (the narrow profile).

As previously mentioned, the bladeis twisted so that it may follow thechange in direction of the resulting wind.The angle between the plane ofrotation and the profile chord, animaginary line drawn between theleading edge and the trailing edge,is called the setting angle, sometimesreferred to as ÒPitchÓ.

WHAT HAPPENS WHEN THE WIND SPEED CHANGES?The description so far was made withreference to a couple of examples wherewind speed was at a constant 10 m/s.We will now examine what happensduring alterations in the wind speed.

In order to understand blade behaviorat different wind speeds, it is necessaryto understand a little about how lift anddrag change with a different angle ofattack. This is the angle between theresulting wind ÒwÓ and the profile chord.In the drawing below the angle ofattack is called ÒaÓ and the settingangle is called ÒbÓ.

The setting angle has a fixed value atany one given place on the blade,but the angle of attack will grow as thewind speed increases.

The angles of the profile

A B C

v v w

u

FF

Fa

Fd

Plane of rotationPlane of rotation

Chord Plane of rotation

a

bw

Air flow around a blade profile near the blade root

9

The aerodynamic properties of theprofile will change when the angle ofattack ÒaÓ changes. These changes of liftand drag with increasing angles of attack,are illustrated in the diagram above usedto calculate the strength of these two for-ces, the lift coefficient ÒCLÓ and the dragcoefficient ÒCDÓ. Lift will always be at aright angle to the resulting wind, whiledrag will always follow in the directionof the resulting wind.

We will not enter into the formulasnecessary to calculate these forces, it isenough to know that there is a direct con-nection between the size of ÒCLÓ and theamount of lift.

Both lift and drag abruptly changewhen the angle of attack exceeds 15-20degrees. One can say that the profilestalls. After this stalling point is reached,lift falls and drag increases. The angle ofattack changes when the wind speedchanges.

To further study these changes, wecan draw diagrams, shown to the right,illustrating three different wind speedsÒvÓ (5, 15 and 25 m/s) from our previouscross section, this time near the blade tipof a 450 kW wind turbine. This situationis rather convenient as the setting angleÒbÓ near the wing tip is normally 0 degrees.

The head wind from the movementÒuÓ is always the same, as the windturbine has a constant rotational speed

controlled by the grid connected generator (in these situations we do notconsider the small generator used oncertain small wind turbines). The free airflow ÒvÓ has three different values andthis gives three different values of theresulting wind ÒwÓ across the profile.The size of ÒwÓ does not change verymuch, from 50 m/s at a wind speed of5 m/s to 52 m/s in a 25 m/s wind. Thereason for this relatively minor change isdue to the dominating effect of the headwind.

However, the angle of attack ÒaÓbetween the resulting wind and the chordof the blade changes from 6 degrees ata wind speed of 5 m/s to 16 degrees at15 m/s to 27 degrees at 25 m/s. Thesechanges are of great importance fordetermining the strength of the aerody-namic forces.

Studying the diagram showing the liftcoefficient ÒCLÓ and the drag coefficientÒCDÓ we may note the following:¥ At a wind speed of 5 m/s (A), theangle of attack is 6 degrees. The liftcoefficient is 0.9 and the coefficient ofdrag is 0.01. Lift is therefore 90 timesgreater than drag, and the resultant forceÒFÓ points almost vertically at a rightangle to the mean relative wind ÒwÓ.¥ At a wind speed of 15 m/s (B), theprofile is almost about to stall. The angleof attack is 16 degrees. The liftcoefficient is 1.4 and the coefficient of

drag is 0.07. Lift is now 20 times drag.¥ At a wind speed of 25 m/s (C), theprofile is now deeply stalled, the angle ofattack is 27 degrees, the lift component is1.0 and the component of lift is 0.35. Liftis now 3 times greater than drag. We cantherefore note the following:¥ During the change of wind speed from5 to 15 m/s there is a significant increasein lift, and this increase is directed in thedirection of rotation. Therefore poweroutput of the wind turbine is greatlyincreased from 15 kW to 475 kW.¥ During the change of wind speed from15 to 25 m/s, there is a drop in liftaccompanied by an increase in drag.This lift is even more directed in thedirection of rotation, but it is opposed bydrag and therefore output will fall slightlyto 425 kW.

Coe

ffici

ents

of

Lift

and

Dra

g (C

L &

CD

)

Angle of attack ÓaÓ

Relationship between lift and drag coefficients and the angle of attack

F

Plane of rotationA

u

v (5 m/s)w

B

F

Plane of rotation

u

wv (15 m/s)

F

Plane of rotation

u

wv (25 m/s)

C

Situations at three different wind speeds

LiftDrag

10

THE STALL PHENOMENA The diagrams showing the components oflift and drag illustrate the result of stall.Lift diminishes and drag increases at angles of attack over 15 degrees. Thediagrams however do not illustrate thereasons for this stall phenomena.

A stall is understood as a situationduring which an angle of attack becomesso large that the air flow no can longerflow smoothly, or laminar, across theprofile. Air looses contact with the rearside of the blade, and strong turbulenceoccurs. This separation of air massesnormally commences progressively fromthe trailing edge, so the profile graduallybecomes semi-stalled at a certain angle ofattack, but a full stall is first achieved ata somewhat higher angle. From thediagram showing the lift and dragcomponents, one can estimate that theseparation at the trailing edge starts atabout 12 degrees, where the curveillustrating lift starts to fall. The profileis fully stalled, and the air flow isseparated all over the rear side of theblade at about 20 degrees. These figurescan greatly vary from profile to profileand also between different thicknesses ofthe same profile.

When the stall phenomena is used torestrict power output, as in all Bonuswind turbines, it is important that bladesare trimmed correctly. With the steep liftcurve, the angle of attack cannot bealtered very much, before maximumoutput also changes, therefore it isessential that the angle of the blade isset at the correct value.

One cannot alter the different angleson the blade itself, once the form, shapeand blade molding has been decided uponand fabricated. So we normally talkabout calibrating the tip angle. Notbecause the blade tip has any special

magical properties, but we can place a template at the tip, which allows us tomake measurements using a theodolite.Adjusting of the tip angle can thereforebe understood as an example of how theangle of the total blade is adjusted.

Of importance for power outputlimitation is also the fact that in practicelift and drag normally behave exactly aswould be expected from the theoreticalcalculations. However this is not alwaysthe case. Separation can often occurbefore expected, for instance due to dirton the leading edges, or it can be delayedif the air flow over the profile for somereason or other, is smoother than usual.When separation occurs before expected,the maximum obtainable lift is not ashigh as otherwise expected and thereforemaximum output is lower. On the otherhand, delayed separation can cause con-tinuous excessive power productionoutput.

Accordingly profile types chosen forour blades have stable stall charac-teristics with little tendency to unforeseenchanges. From time to time, however,it is sometimes necessary to actively alterthe stall process. This is normally doneby alteration to the leading edge, so thata small well-defined extra turbulenceacross the profile is induced. This extraturbulence gives a smoother stall process.

Turbulence can be created by an areaof rougher blade surface, or a triangularstrip, fixed on the leading edge. This stallstrip acts as a trigger for the stall so thatseparation occurs simultaneously all overthe rear side.

On a wind turbine blade, different airflows over the different profile shapes,interact with each other out along theblade and therefore, as a rule, it is onlynecessary to alter the leading edge on

a small section of the blade. This alteredsection will then produce a stall over thegreater part of the blade. For example,the Bonus 450 kW Mk III turbine, isusually equipped with a 0.5 meter stallstrib, which controls the stall process allover the 17 meter long blade.

SUMMARYThe main points as described in thisarticle can be shortly stated in thefollowing:

¥ The air flow around a wind turbineblade is completely dominated by thehead wind from the rotational movementof the blade through the air.

¥ The blade aerodynamic profileproduces lift because of its streamlinedshape. The rear side is more curved thanthe front side.

¥ The lift effect on the blade aerodyna-mic profile causes the forces of the air topoint in the correct direction.

¥ The blade width, thickness, and twist isa compromise between the need for stre-amlining and the need for strength.

¥ At constant shaft speed, in step with thegrid, the angle of attack increases withincreasing wind speed. The blade stallswhen the angle of attack exceeds15 degrees. In a stall condition the aircan no longer flow smoothly or laminarover the rear side of the blade, lifttherefore falls and drag increases.

Seperation of the air flow at the profile trailing edge Interference in the stall process (stall strip)

Stall strip

11

Just how much of a wind turbine thatbelongs to the transmission system is amatter of definition. In this chapter wewill include the components thatconnect the wind turbine rotor to thegenerator.

THE HUBThe blades on all Bonus wind turbinesare bolted to the hub. Older Bonus windturbines (up to and including the 95 kWmodels) with Aerostar blades, have aflange joint, where the glass fiber is molded out in a ring with steel bushesfor the bolts. The newer wind turbines(from the 150 kW models) have threadedbushes glued into the blade root itself.In both cases bolts from the bladepass through a flange on the cast hub.The flange bolt-holes are elongated,enabling the blade tip angle to beadjusted.

The hub is cast in a special typeof strong iron alloy, called ÒSG castironÓ. Because of the complicated hubshape which is difficult to make in anyother way, it is convenient to use castiron. In addition the hub must be highlyresistant to metal fatigue, and thisis difficult to achieve in a weldedconstruction.

In contrast to cast iron of the SG type,normal cast iron has the disadvantage ofbeing rather fragile and often canfracture under blows. This unfortunatequality is due to the high carbon contentof cast iron. High carbon content enablesthe cast iron to melt easily and thuseasily flow out into the casting form.When cast iron solidifies, carbon exists asgraphite flakes suspended in the pureiron. These flakes form weak zones in thematerial, easily prone to zig-zag fissuresfrom flake to flake. These weak zones areonly important, if forces attempt to pull

the material apart. Graphite has greatcompressibility strength, and is thereforenot easily compressed. Normal cast ironhas the same compressibility strength assteel, but its tension resistance level isonly 10% of steel tension resistance.

For many uses these strength qualitiesare more than sufficient, however inconstructions subject to heavy usage,properties such as low tension resistanceand weakness under blows are notdesirable. For this reason special SG castiron with tension resistance equal to thatof steel has been developed during thepast 50 years.

In producing SG cast iron severalspecial materials, mainly silicium, areadded during casting. After casting hastaken place, it is further heat treated forabout 24 hours, thereby changing the freecarbon from their usual flakes into smallround balls. The name SG cast iron isalso short for Spherical Graphite cast iron(latin: Sphere = ball).

This round ball shape binds thenecessary carbon in a more compactform. The graphite is not a hindrance forthe binding structure in the metal itself,and there is likewise a better structurebetween the crystals of iron. Therebyachieving the higher strength qualities

THE TRANSMISSION SYSTEM

The link between the wind turbine blades and the generator

Wind turbine hub

Main bearing

Hub

Main shaft

Gear

Coupling

12

necessary for a wind turbine hub.On account of the extra heat treatment,SG cast iron is somewhat more expensivethan normal cast iron.

MAIN SHAFTThe main shaft of a wind turbine isusually forged from hardened andtempered steel. Hardening and temperingis a result of forging the axle after ithas been heated until it is white-hotat about 1000 degrees centigrade. Byhammering or rolling the blank is formedwith an integral flange, to which the hubis later bolted.

The shaft is reheated a final time to aglowing red, following the forgingprocess, and then plunged into a basin ofoil or water. This treatment gives a veryhard, but at the same time rather brittlesurface. Therefore the axle is once againreheated to about 500 degrees centigrade,tempering the metal and thereby enablingthe metal to regain some of its formerstrength.

MAIN BEARINGS All modern wind turbines, including theBonus models, have spherical rolllerbearings as main bearings. The termspherical means that the inside of thebearingÕs outer ring is shaped like a crosssection of a ball. This has the advantageof allowing the bearingÕs inner and outerring to be slightly slanted and out-of-track in relation to each other withoutdamaging the bearing while running.The maximum allowable oblique angle isnormally 1/2 degree, not so large, butlarge enough to ensure that any possiblesmall errors in alignment between thewind turbine shaft and the bearinghousing will not give excessive edge

loads, resulting in possible damage to thebearing.

The spherical bearing has two setsof rollers, allowing both absorptionof radial loads (across the shaft) fromthe weight of the rotor, shaft, etc. andthe large axial forces (along the shaft)resulting from the wind pressure onthe rotor.

The main bearings are mounted in thebearing housings bolted to the mainframe. The quantity of bearings andbearing seats vary among the differenttypes of wind turbines: Ò Small Ó windturbines up to and including 150 kW havetwo bearings, each with its own flangedbearing housing. The 250/300 kW windturbines have only one main bearing,with the gearbox functioning as a secondmain bearing. The 450 kW, 500 kW and600 kW wind turbine models have twomain bearings, using the hub as ahousing. Each bearing arrangement hasadvantages and disadvantages, and theevaluation of these properties haveprovides each individual type with itsown setup.

The main bearings are alwayslubricated by greasing, no matter whichbearing arrangement is selected. Specialgrease having viscose properties even inhard frost is used.

Sealing of the bearing housing isinsured by the use of a labyrinth packing.No rubber sealing is used, the labyrinthwith its long and narrow passagewayprevents grease from escaping. Water anddirt are prevented from entering from theoutside by the long passageways filledwith grease, which is constantly andslowly trying to escape from the bearing.This may appear to be a rather primitivearrangement, but labyrinth packing is amuch used method where there is greatrisk of pollution by water and dirt. It ismore expensive to use than a rubber sea-ling, because the labyrinth is complicatedto fabricate on machine tools, howeverthe seal is not subject to wear, and undernormal conditions it is a safe method tokeep out the pollutants that otherwise in ashort time could ruin roller bearings.

THE CLAMPING UNITBy the means of a clamping unit the mainshaft of the wind turbine is coupled to thegearbox. The gear has a hollow shaft that

fits over the rear end of the main shaft.Torque between the two components istransferred by friction between the two.

A clamping unit, normally composedof an inner ring and two outer rings withconical facings, is placed on the outsideof the gearÕs hollow shaft. When the mainshaft is placed inside the hollow shaftduring the assembly of the wind turbine,the conical facings of the clamping unitare loosely positioned on the hollowshaft. Following control of the correctalignment of the gear and the main shaft,the rings are tightened by the means of alarge number of bolts. The outer rings arethereby pressed together, while the innerring, positioned on the hollow shaft ispressed inwards under the tightening ofthe bolts. The inner ring now presses sohard against the hollow shaft that theinner part of the hollow shaft is in turnpressed hard against the main shaft. It isbecause of this pressure that the torque is

Clamping unit ¥ (TAS Sh�fer)

Outer rings

Inner ring

Hollow shaft

Main shaft

¥

¥

¥

Spherical roller bearing ¥ (Niemann)

13

transferred from the main shaft to thewind turbine gear hollow shaft. Onemight also say that the hollow shaft isshrink-fitted on the main shaft as a resultof pressure from the clamping unit.

Transferred torque is dependent uponfriction between the main shaft andthe hollow shaft. Therefore it is vital thatthe components are carefully cleanedand completely dry, before they areassembled. If they are at all greasy, theycould slip in relation to each other duringhigh loads, for example during the cut-inprocess in strong wind conditions.

Many know of the parallel keymethod, often used in assembling a shaftto a hub. The main shaftÕs torque istransferred by forces across the parallelkey (a parallel key is often called awedge, even though it is not wedgeshaped). This assembly method is notoften used with a large shaft, there beingtoo great a risk that in time the differentparts could loosen, unless they fit uncom-monly well together. If the parallel keyjunction assembly method is used forlarge shafts, parts must fit so welltogether, that in practice one is unable todismantle them in the field, should it benecessary during possible replacement incase of damage or repair.

THE GEARBOXOne of the most important main com-ponents in the wind turbine is thegearbox. Placed between the main shaftand the generator, its task is to increasethe slow rotational speed of the rotorblades to the generator rotation speed of1000 or 1500 revolutions per minute(rpm).Without much previous experience withwind turbines, one might think that thegearbox could be used to change speed,just like a normal car gearbox. Howeverthis is not the case with a gearbox in awind turbine.

In this case the gearbox has always aconstant and a speed increasing ratio,so that if a wind turbine has differentoperational speeds, it is because it hastwo different sized generators, each withits own different speed of rotation (or onegenerator with two different statorwindings).

As an example of a gearboxconstruction, we can study a Flender

SZAK 1380 gear for a 150 kW windturbine. This gear has two sets of toothedgear wheels, a slow speed stage and ahigh speed stage. In the slow speed stagethe large gear wheel is mounted directlyon the gearÕs hollow shaft, while thesmaller gear wheel is machined directlyon the intermediate shaft.

The difference in the size of thewheels is 1:5. The intermediate shafttherefore turns 5 times every time thehollow shaft makes one completerevolution. The large gear wheel in thehigh speed gear stage is also mounted onthe intermediate shaft, while the smallgear wheel in the high speed gear stage ismachined on the generator shaft itself.Here the difference in size is also about1:5, so that the output shaft to the generator shaft turns 5 times for everyone rotation of the intermediate shaft.

When the two ratios are combined,the output shaft will turn 25 times forevery rotation of the hollow shaft and themain shaft of the wind turbine combined

One can say that the gear has a gear ratioof 1:25.

Normally the ratio in every set of gearwheels is restricted to about less than 1:6.The 150 kW wind turbine has a rotorrotational speed of 40 rpm and with agenerator speed of about 1000 rpm, thegearbox must have a total gear ratio of40/1000 or 1:25. This is possible using atwo stage gearbox. A 300 kW windturbine has a rotor rotational speed of31 rpm and a generator with a rotationalspeed of 1500 rpm. It therefore requires agearbox with a gearbox ratio of 31/1500or 1:48. This is not possible using a gear-box with only two stages, so the 300 kWwind turbine gearbox has an extraintermediate shaft, giving in all a threestage gearbox.

Wind turbines, from 450 kW andlarger, have an integrated gearbox with aplanet gear and two normal stages. Theplanet gear is a special version of thetoothed gear. This type of gear is of great delight to gearbox technicians, as it can

1 Hollow shaft2 Intermediate shaft3 High speed shaft

for the generator

Slow set4 Large toothed wheel 5 Small toothed wheel

High speed set6 Large toothed wheel7 Small toothed wheel

Flender SZAK 1380 2-trins gear Planetgear ¥ /DIN 686/Niemann)

1 Ring wheel2 Planet wheel3 Sun wheel4 Planet carrier

14

be combined in countless different com-plicated variations, each one carefullycalculated with its own special innerlogic. The form of planet gear used onwind turbines is however always of thesame basic design: An interior toothedgear wheel (ring wheel), three smallertoothed gear wheels (planet wheels)carried on a common carrier arm (theplanet carrier ) and finally a centrallyplaced toothed gear wheel (the sun gearwheel). It is this construction, with threesmaller gear wheels orbiting a centrallyplaced common gear wheel that has giventhis type of gear its name of planet gear-box.

The ring wheel itself is stationary,while the planet carrier is mounted on thehollow shaft. When the planet carrierrotates with the same rotational speed asthe rotor blades, the three planet wheelsturn around inside the inner circum-ference of the ring wheel and thereby alsogreatly increase the rotational speed ofthe centrally placed sun gear wheel. Onecan usually obtain a gear ratio of up toabout 1:5. The sun gear wheel is fixed toan shaft driving the two normal gearstages placed at the rear end of thegearbox.

The fact that there are always threegear wheels supporting each other andthat all gear wheels are engaged at thesame time, is one of the advantages of theplanet gear. This means that it is possibleto construct rather compact planet gear-boxes, because the larger ring wheel doesnot need to be as large as a gear wheel ina traditional type of gearbox. In principleit only needs to be about a 1/3 of the size.However in reality it not quite so simple.If a gear is needed to transfer heavyloads, it is often somewhat cheaper to usea planet gear.

However it is in the very nature ofthings that trees do not grow up intoheaven, and also planet gears have theirown special disadvantages. The compactconstruction, very practical for the designand construction of the rest of themachine, can be in itself a disadvantage.The compact construction makes itdifficult to effectively dissipate excessheat to the surroundings. A gear is not100% effective, and as a rule of thumb itis estimated that roughly 1% of thepower is lost at each stage. A 600 kW

gearbox running at full capacity, musttherefore dispose of about 18 kW ofwaste heat. This is equivalent to ninenormal household hot air blower-heatersoperating at full blast. This waste heatshould preferably be radiated by surfacecooling and of course the less gearboxsurface area, the higher the temperaturemust be inside the gearbox to transfer thenecessary, unavoidable excess wasteheat.

Another disadvantage of the planetgear is that they normally cannot beconstructed with bevelled machinedteeth. Bevelled teeth are always used innormal gearboxes in order to reduce thenoise level. When the teeth are set at anangle, the next tooth will start to engageand take up the load before the previoustooth has slipped contact. This results in aquieter, more harmonious operation. Forinterior gear wheels bevelled teeth canonly be machined using special machinetools that up until now have solely beenused for the machining of very largeturbine gears for use in ships. Thereforeplanet gears have always straightmachined teeth, unfortunately however,resulting in a higher noise level. Bycombining a planet gear stage and twonormal gear stages, one obtains anacceptable compromise of the advantagesand disadvantages with the two differenttypes of gear.

No matter what type of gear is used,the shape of the teeth in the differentgear stages are adapted to the specialconditions for wind turbine operation,especially those that are related to thenoise level. Teeth as a rule are case-har-dened and polished. Case-hardening is amethod of giving surface strength to aspecific material. During this process, theinner material maintains its previousstrength, which can often be lost innormal steel hardening processes.

Hardening can only take place underconditions where there is a carbon contentin the steel. The gear wheels are made of aspecial low carbon chrome-nickel steel.The teeth are first machined, andfollowing the machining process, the gearwheels are packed into large boxes full ofbone flour or some other form of highcarbon-content powder. The boxes areplaced in an oven and heated for about 24hours to a red glowing temperature.

During this baking process some of thefree carbon will be transferred from thesurrounding carbon-rich powder in theboxes to the gear wheel teeth surfaces.This is described as the method of harde-ning the teeth in boxes or cases, andtherefore from this process comes thedescriptive name of case-hardening.

The increased carbon content of theteeth surface allows the top edges of thegear wheel teeth to become harder, sofollowing case hardening, the gear wheelis lifted out, still red hot, and lowered intoan oil bath. This completes the process ofhardening, and the gear wheel now has ahardened surface, while the innermaterial still has ductile and not hardenedproperties. The hardening processslightly deforms the material, so it isnecessary to finish the process bygrinding.

THE COUPLING

The coupling is placed between the gear-box and the generator. Once again it isnot possible to consider the coupling asthe same as a clutch in a normal car. Onecannot engage or disengage the transmis-sion between the gearbox and the genera-tor by pressing a pedal, or in some othersuch way. The transmission is apermanent union, and the expressionÒcouplingÓ should be understood as ajunction made by a separate machinecomponent.

The coupling is always a ÒflexibleÓunit, made from built-in pieces of rubber,normally allowing variations of a fewmillimeters only. This flexibility allowsfor some slight differences in alignmentbetween the generator and the gearbox.This can be of importance underassembly and also during running opera-tion, when both gearbox and generatorcan have tendencies for slight movementin relation to each other.

Coupling¥ (Flender BIPEX)

15

The generator is the unit of the windturbine that transforms mechanicalenergy into electrical energy. The bladestransfer the kinetic energy from thewind into rotational energy in the trans-mission system, and the generator is thenext step in the supply of energy fromthe wind turbine to the electrical grid.

In order to understand how a generatorworks, it is necessary to first of all under-stand the deeper principles in theelectrical system to which the generatoris connected. Therefore we will firstdiscuss the electrical systems based onDirect Current (DC) and those based onAlternating Current (AC).

DIRECT CURRENT (DC) During the first use of electricity forlighting and power in the previouscentury, systems based on direct currentwere used. In DC systems the voltage isat a constant level. This could be1.5 Volts (V) as in a modern alarm clock,12 V as in a car or 110 V as in the firstproper electrical grid.

DC has the advantage that batteriescan be connected, enabling a continualsupply of electrical power even if thegenerator at the power station ceasesoperation and shuts down. Therefore thefirst power stations had large storerooms full of long rows of batteries.Such systems were well adapted to theuse of wind turbines as a main powersource, for with such large stocks ofbatteries, power could still be suppliedeven in calm periods.

In spite of the advantages of batteryenergy storage, DC is no longer used inlarger grid electrical supply systems. Thisis due to some important disadvantagesof direct current, while on the other handthe competing electrical system alterna-ting current offers important advantages.

One of the big disadvantages of DCis the strong electrical arc produced,when the electrical current connectionfrom supply to user is cut at highervoltages. For example, in larger instal-lations with connections to electricalmotors DC switches are both large andcomplicated. Therefore in practice DCsystems can be rather inconvenient.

Another ÒdisadvantageÓ is that theadvantages of battery energy storagedo not in reality exist with theelectrical grid systems in common usetoday. This is because our present-dayenergy consumption greatly exceedsthe capacity of this technology.

A typical Danish family has an energyconsumption of about 5.000 kWh peryear, or about 13.7 kWh per day. Anormal car battery has a capacity of about60 Ah (Ampere-hours). This means that acar battery can supply an electrical cur-rent equal to 1 Ampere for about 60 hoursat a battery voltage of 12 Volts. Theenergy in a fully charged battery can becalculated by the use of a simple formula:

E = 60 Ah x 12 V = 0.72 kWhTherefore less than 1 kWh is stored in afully charged car battery. A typicalDanish family with a daily requirementof 13,7 kWh kWh per day will thus need19 fully charged batteries just to cover the

power consumption of a single daywithout a supply from the power stationgrid network.

Another example: In a good highwind period a 600 kW wind turbine cantypically produce about 10.000 kWh perday. This is enough to charge about14.000 car batteries per day, were it is notpossible to supply this energy productionfor the direct consumption or use by theowner, or for supply to other consumersconnected to the grid.

In connection with such large quanti-ties of energy, storage in batteries is notfeasible, and the storage possibilitiesoffered by the use of DC systems are notreally practically relevant.

ALTERNATING CURRENT (AC) The voltage of the current constantlyvaries around zero in an AC electricalsystem. The maximum voltage must besomewhat higher than a DC system inorder to give the same power. One canspeak of an effective medium voltage as akind of average of the voltage.

AC measuring instruments usuallyshow the effective middle voltage valueand not the maximum voltage.

A lamp connected to an alternatingelectrical current will blink, as thevoltage constantly varies. The frequencyof the voltage variation or cycles inDenmark, and most other countries is 50Hz (50 cycles per second). Such rapidcycles make the blinking of the lamp ofno real importance. The glowing wire in

THE GENERATOR The wind turbine electrical system

DC-system

AC-systemThe battery store room of a wind power plant at thebeginning of the 1900«s ¥ (H.C.Hansen: Poul la Cour)

Voltage (V)

DC-current

Voltage (V)

Max. voltage (V)

Time

AC-current

Time

Eff. medium voltage

16

a normal electric bulb does not have timeto become cold in the short periodbetween cycles, and therefore does not inpractice blink. In comparison light emit-ting from a neon tube is completely shutoff each time the voltage is at zero. Theeye however cannot distinguish variati-ons in light intensity that occur fasterthan 15 times a second, so therefore wesee light from a neon tube also as con-stant.

The main advantage of alternatingcurrent over direct current is that thevoltage can be altered using transfor-mers. This is not the place to describe indetail the functioning of a transformer,but in principal it is possible to alter fromone voltage to another voltage almostwithout loss of energy.

Most know the small transformersused as power supply to radios, mobiletelephones, etc. A small box is pluggedinto a 220 volt outlet connected to thegrid and 9 volts comes out at the otherend (normally also rectified to directcurrent, but that is another story). For thegrid as a whole, it is the transformation toa higher voltage that is of importance.

The advantage of high voltage is thatenergy losses in power transmissionlines, are greatly reduced by usingincreased voltages. In order to under-stand this, one must know a couple of thefundamental formulas in electricalengineering. As an example consider thecase of a typical 220 volt electrical tool,a 2.200 Watt (W) grinder.

The current one obtains at specificpower and voltage ratings may be calcu-lated with the formula:

I = P / UWhere ÓIÓ is the current, ÓPÓ is thepower and ÓUÓ is the voltage. In theexample of the grinder, with power P =2.200 W and voltage U = 220 V Weobtain the current of 2.200 / 220 = 10 A.

The power loss from the wires may becalculated with the formula:

T = R x I 2

Where ÓTÓ is the power loss and ÓRÓ isthe resistance of the wire. A normalhousehold electric wire with a crosssection of 1.5 mm2 has a resistance of0.02 Ohm per meter. A 10 meter longwire will have a resistance of 0.2 Ohmand the power loss in the wire willtherefore be T = 0.2 x 10 2 = 20 W.

This is not so much, only about 1% of thegrinderÕs usable power.

The power loss is however quitesignificant, when one considers thedistance from the user to the powerstation. With a typical distance of about20 km , the resistance in a 1.5 mm2 wirewill be about 400 Ohm, and the powerloss will therefore be T = 400 x 102 =40,000 W or almost 20 times the powerof the grinder! Of course small 1.5 mm2

wires are not used as power supply cab-les from the power station out to theconsumer, but even with large 50 mm2

cables, the power loss is still larger thanthe rated power of the grinder.

It is in this situation that high voltagetransmission wires have their use.If instead of 220 V the power stationsends an electrical current of 10.000 Vout in the electrical grid to theconsumer, the first formula for currentwill give I = 2.200 /10.000 = 0.22 A, and the other formula for power loss willgive T = 400 x 0.222 = 20 W still usingthe same (unrealistic) wire dimension of1.5 mm2. The use of high voltage powerlines has therefore reduced power lossfrom an unacceptable level to that whichis more acceptable.

In practice current is transmitted frompower stations with a voltage of up to400,000 V . This is then transformed to alower voltage in large centralized trans-former stations, for example down to10,000 V. Near the consumer the finaltransformation down to 220 V is made.

For safety reasons high voltage is notused near the consumer, as electricalcurrent becomes more dangerous, thehigher the voltage is increased. Likewisethe demands on the safety insulation ofelectrical material also increases.

Voltage at any one given place on thegrid is therefore a compromise betweena desire on the one side for a minorpower loss (requiring high voltage), andon the other hand the necessity of a lowor moderate risk of danger and at thesame time reasonably cheap electricalinstallations (requiring lower voltage).

THREE PHASE ALTERNATING CURRENTEven though the cycles in the alternatingcurrent are of no great importance forlamps and other such things, it is

impractical for certain other machinesthat the current is always alternatingaround zero. Therefore, years ago, it wasdiscovered that AC could be suppliedwith three phases.

The principle of 3 phase electricalpower is that the generator at the powerstation supplies 3 separate alternatingcurrents, whose only difference is thatthey peak at three different times.The knack with these three separate alter-nating currents, or phases, is that it isthereby possible to ensure that the sumof the delivered power is alwaysconstant, which is not possible with twoor four phases.

It is perhaps a little impractical withthree phase current, because it is necessa-ry to run four different wires out to theconsumer, three different phase wiresand a neutral wire (zero). However forelectric motor use, the advantages ofthree phase alternating current are many.The voltage difference between two ofthe phases is greater than that betweenany one single phase and zero. Where thevoltage difference is 220 V between onephase and zero, it is 380 V between twophases.

This is often used in high energyconsumption equipment such as kitchenovens etc., which normally always areconnected to two phase power. In ahousehold installation usually only oneof the phases plus the neutral wire is ledto an ordinary socket. Normally theinstallation has several groups, and onephase will typically cover one part of thehouse, and another phase will run to theother rooms. Three phase sockets arerather large and are often known aspower sockets, mainly because of theiruse in electrical motor operation. Forease in distinguishing between the diffe-rent phases, in Denmark the three phaseshave been named R, S, and T.

Three phase AC (three super-imposed sinus curves)

Voltage (Volt)

Time

17

On the older Danish transmission linessupported by wooden masts, phases wereplaced in a certain specific order, readingfrom the bottom up, according to theDanish words for root (R), trunk (S) andtop (T).

INDUCTION AND ELECTROMAGNETISMBefore finally describing the generatoritself, we must briefly explain a couple ofthe basic principles of electromagnetism.

Many perhaps remember our schooldays, when the physics teacher placed amagnetic bar inside a coil of copper wireconnected to a measuring instrument.

If the magnet is stuck inside the coil, anelectric current is registered in the coilcircuit. If the magnet is withdrawn, acurrent of the same strength is registered,but in the opposite direction. The fasterthe changes of the magnetic field in thecoil, the greater the current. The sameoccurs if instead of the magnet beingstuck into the open coil it is merelymoved past one of the ends of the coil.The effect is especially powerful if thecoil has a iron core.

One can say that alterations in themagnetic field, induce a current in thecoil, and the phenomena is known asinduction.

In just the same way that a magneticfield can bring about an electric current,

so can an electric current likewise cause amagnetic field to be created. Electro-magnetism was first demonstrated by theDanish scientist H.C ¯rsted in hisfamous experiment, where an electricalcurrent was able to turn a compassneedle. He had therefore demonstratedthe first electromagnet.

In practice a good electromagnet isbest made as a coil with an iron core, injust the same way as the previouslymentioned form of coil that produces anelectric current when a magnet is movedpast at a close distance. Like a permanentmagnet an electromagnet has two poles, anorth pole and a south pole. The positionof these two poles depends on the directi-on of the flow of electrical current.

THE WIND TURBINE GENERATORAS A MOTORThe asynchronous generator we willdescribe here is the most common type ofgenerator used in Danish wind turbines.It is often referred to as the inductiongenerator, too. As far as we know theasynchronous generator was first used inDenmark by Johannes Juul, known forthe 200 kW Gedser wind turbine from

1957. Already some years prior to this construction he erected a 13 kWexperimental wind turbine with anasynchronous generator at VesterEgesborg in the south of the largeDanish island of Zeeland.

The asynchronous generator is inreality a type of motor that can alsooperate as a generator, and we will firstconsider this type as a motor. This is themost common electric motor, sitting inalmost every washing machine, andwidely used as a motor unit in industry.

The motor consists of two main parts,the stator and the rotor. The statorcontains a series of coils, the number ofwhich must be divisible by three. Themotor illustrated on this page has sixcoils, placed in slots on the inside of thestator, a cylinder assembled of thin ironplates. The rotor sits on an axle placedinside this stator. The rotor is alsoassembled of thin iron plates. A row ofthick aluminum bars joined at each endwith an aluminum ring, fit in key ways onthe outer surface of the rotor. This rotorconstruction looks a bit like a squirrelcage, and accordingly the asynchronousmotor is also called a squirrel cage motor.

The principles of induction

S

S

7. Coil8. Stator plates9. Coil heads

10. Ventilator11. Connection box

1. Generator shaft2. Rolling bearings3. Rotor 4. Rotor aluminium bar5. Rotor aluminium ring6. Stator

Components of an asynchronous motor

1

2 3

4 5

6

78

9

10

11

Current(I)

Current(I)

18

The six coils in the stator are connectedtogether, two by two to the threedifferent phases of the electrical grid.This arrangement insures that there is arotating magnetic field inside the statoritself. This is best illustrated by the abo-ve diagram.

At a specific time Ò1Ó the current inphase R is at its maximum, and thisproduces a magnetic field with a strongnorth pole at both the opposite coilsconnected to the phase R. At phase S andphase T the current is somewhat underzero, and the two pairs of coils produce a

medium strength south pole, producing apowerful south pole halfway between thetwo coils.

At time Ò2Ó the current at phase S isat a maximum, and the north pole is nowat the two opposing coils connected tothis phase. The current at phases R and Tis likewise reduced to under zero, and thesouth pole is now between these twocoils.

At time Ò3Ó the current at phase Tnow is at a maximum, and the north poleis at the two coils connected to phase T.The south pole has also turned, and is

now halfway between the coils connectedto phases R and S.

At time Ò4Ó the situation has nowreturned to as it was at the start of theelectrical current rotation, with the northpoles at the end of the coils connected tophase R.

In one complete cycle, from thecurrent peak to the next following peak,the magnetic field has rotated throughhalf a circle. There are 50 cycles persecond, so the field turns at 25 times persecond, or 60 x 25 = 1.500 rpm(revolutions per minute).

To understand how a generatorworks, it is easiest to first consider twodifferent situations where a generatoroperates as a motor, at 0 rpm. and at1.500 rpm.

In the first case the rotor is stationary,while the stator turns at 1.500 rpm. Thecoils in the rotor experience rapidvariations of a powerful magnetic field.A powerful current is thereby induced inthe short circuited rotor wire windings.This induced current produces an intensemagnetic field around the rotor. Thenorth pole in this magnetic field isattracted by the south pole in the statorÕsturning magnetic field (and of course,the other way round) and this will givethe rotor a torque in the same direction asthe moving magnetic field. Therefore therotor will start turning.

In the second situation, the rotor isturning at the same speed as the statormagnetic field of 1.500 rpm. This rotati-onal figure is called the synchronousrotational speed. When the stator mag-netic field and the rotor are synchronized,the rotor coils will not experience variati-ons in the magnetic field, and thereforecurrent will not be induced in the shortcircuited rotor windings. Without indu-ced current in the rotor, there will be nomagnetic field in the rotor windings andthe torque will be zero.

On account of bearing friction themotor must produce a little torque tokeep rotating, and therefore cannot runat exactly the same speed as the rotatingmagnetic field. As soon as the speedslows down, there will be a differencebetween the speed of the rotating mag-netic field and the rotor. The rotor thusagain experiences a variation in themagnetic field that induces a current in

4 situations of the rotation magnetic field

Vo

ltag

eV

olt

age

Vo

ltag

eV

olt

age

Time

TimeTime

Time

19

the rotor windings. This current then pro-duces a magnetic field in the rotor, andthe rotor can produce a torque.

During motor operation, the statorexperiences a constantly changing mag-netic field, being dragged round by itsrotating magnetic field. During this process, electrical current is induced inthe stator, which results in a powerconsumption. In fact, the slower the rotorturns in relation to the rotating magneticfield of the stator, the stronger the indu-ction in the stator, and therefore the gre-ater the power consumption.

The fact that the rotor has no torque atthe precise synchronous rotational speedand therefore will always run slightlyslower has given this motor type itsname, the asynchronous motor.

GENERATOR OPERATIONAs we have previously mentioned, theasynchronous motor can also run as agenerator. This simply happens whenyou, instead of forcing the rotor toturn at a rotational speed lower thanthe synchronous speed, exceed thissynchronous speed by applying an out-side energy source, such as a diesel motoror a set of wind turbine rotor blades.

Once again, the greater the differencebetween the rotating magnetic field of thestator (which is always 1.500 rpm) and thespeed of the rotor, the greater the torqueproduced by the rotor. When a workingas a generator, the rotating field howeveracts as a brake in slowing the rotor. Thestator experiences a variable magneticfield from the rotor that ÒdragsÓ its rota-ting magnetic field and thereby induces anelectrical current in the stator. In compari-son to motor operation the induced cur-rents in the rotor and stator will flow inthe opposite direction, which means thatpower will be sent to the grid. The fasterthe rotor turns in relation to the rotatingmagnetic field of the stator, the greater theinduction in the stator and the greater theproduction of power.

In practice the difference between thespeed of rotational magnetic field of thestator and the rotational speed of the rotoris very little. A rotor will typically turnabout 1% faster at full power production.If the synchronous rotational speed is1.500 rpm then the rotor rotational speedat full power will be 1.515 rpm.

The interesting torque curve of theasynchronous electric motor, also operat-ing as a generator, is shown below. Atspeeds below the synchronous rotationalspeed, the motor yields a positive torque.

Typically a maximum torque of about2.5 times the torque of the nominalpower. If the rotational speed exceeds thesynchronous level, the torque becomesnegative, and the generator acts as a brake.

At the Bonus factory, we have a ratherinteresting apparatus, that demonstratesthis shift between a motor and generator.A small asynchronous motor is connectedto an electric meter. The motor has a gear-box giving a shaft speed of 60 rpm.

A small crank handle is fixed to theshaft. The motor starts when it is pluggedinto a normal mains socket coming fromthe electrical grid and consumes a smallamount of electrical energy due to frictionloss in the motor and gearbox.

If one attempts to resist the rotation ofthe shaft by holding back the crank, theconsumption of energy will increase. If thecrank however is used to increase thespeed of the motor, then the electric meterwill start to run backwards, showing thatcurrent is flowing the other way. In thisway one can, by using human musclepower, feed electrical power to the grid, injust the same way that a wind turbine feedspower to the grid. It is difficult to achievemore than 1/20 kW so a work force oftwelve thousand employees is needed tocompete with one single 600 kW windturbine operating in a good wind. Visitorsto Bonus may try their hand at ourgenerator demonstration model.

CUT- IN If a wind turbine is connected to the gridduring a period of no wind, the asyn-chronous generator will operate as a motorand drag the rotor blades round like alarge electric fan. The wind turbine

therefore is disconnected from the gridduring periods of calm.

The wind turbine is likewise discon-nected during periods of low wind speeds,allowing the blades to slowly rotate. Thecontrol system of the wind turbinehowever constantly monitors the rotatio-nal speed, and after the blades reach acertain pre-set level, the system permits agradual cut-in to the grid.

The cut-in to the grid is carried out bythe use of a kind of electronic contactscalled thyristors, allowing continuouslyvariable up and down regulation of theelectrical current. Such thyristors allowsmoother and gentler generator cut-in,thus preventing sudden surges of currentcausing possible grid damage. Likewisethis gentler switching procedure preventsstress forces in the gearbox and in othermechanical components. A direct cut-in,using a much larger electrical switchingunit result in violent shock-effects, notonly to the grid but also to the whole trans-mission system of the wind turbine itself.

Unfortunately, thyristors have thedisadvantage of an power loss of about1-2%, so after the finish of the cut-inphase, current is led past the thyristorsdirect to the grid by the means of aso-called Ò by-pass switch Ò.

CLOSING REMARKSIt has been necessary to make manysimplifications in the above description.We have considered such important terms,as self-induction, reactive current andphase compensation to be too complicatedin a more general description such as this.During the induction process, in reality itis not an electric current that is created, butan electromotive force giving rise to acertain current dependent upon theresistance.

We have used the rotational speed fora 4-pole and 6 coil generator (3 x 2).In the diagram showing the rotating field,one can observe that there are 2 northpoles and 2 south poles, 4 in all. Othergenerators may have 9 coils, which wouldmean 3 north poles and 3 south poles.Such a 6 pole generator has a synchronousrota-tional speed of 1.000 rpm.

Bonus wind turbines up to and inclu-ding the 150 kW models have 6 polegenerators, while the larger models have 4pole generators.

Torque curve

Synchronous rpm

Torque

MOTOR

GENERATOR 100%

100%

rpm

20

Control and safety systems comprisemany different components. Commonfor all of these is that combined toget-her they are part of a more com-prehensive system, insuring that thewind turbine is operated satisfactoryand preventing possible dangeroussituations from arising.

Details in control and safety systems aresomewhat different according to differenttypes of wind turbines. We have inprevious articles described componentsand their functions that roughly covermost Bonus wind turbine models,regardless of their age. However it isnecessary in this article to be much morespecific, so we choose to concentrate onthe Bonus 600 kW Mk IV.

PROBLEM DESCRIPTIONIn constructing wind turbine control andsafety systems one is soon aware of acouple of rather important problems.These problems pose special demands onthe systems, because they have tofunction in the complex environment of awind turbine.

The first problem is common to allcontrol and safety systems: A windturbine is without constant supervision,apart from the supervision of the controlsystem itself. The periods betweennormal qualified maintenance schedulesis about every 6 months, and in theintervening 4,000 hours or so the controlsystem must function trouble-free,whether the wind turbine is in anoperational condition or not.

In almost every other branch of indu-stry there is a much higher degree ofsupervision by trained and qualified staff.On factory production lines, operativesare normally always present duringproduction. For example, in powerstations the system is constantly super-vised from a central control room. Shoulda fault or breakdown occur, rapid inter-vention is possible and, as a rule, one hasalways some sort of good impression ofwhat has actually happened in any

unforeseen occurrence. However a windturbine must be able to look after itselfand in addition have the ability to registerfaults and retrieve this stored informationconcerning any special occurrence,should things possibly not go exactlyquite as expected.

The high demands on reliability requ-ire systems that are simple enough to berobust, but at the same time give the pos-sibility for necessary supervision. Thenumber of sensors and other active com-ponents need to be limited as far aspossible, however the necessary com-ponents must be of the highest possiblequality. The control system has to beconstructed so that there is a high degreeof internal control, and to a certaindegree the system must be able to carryout its own fault finding.

The other problem most of all relatesto the safety systems. A wind turbine, ifnot controlled, will spontaneously over-speed during high wind periods. Withoutprior control it can then be almostimpossible to bring to a stop.

During high wind, a wind turbine canproduce a much higher yield than its rated power. The wind turbine bladerotational speed is therefore restricted,and the wind turbine maintained at therated power, by the grid-connectedgenerator.

If the grid connection is lost, byreason of a power line failure or if thegenerator for some other reason isdisconnected, while the wind turbine is inoperation, the wind turbine wouldimmediately start to rapidly accelerate.The faster the speed, the more power itis able to produce. The wind turbine isin a run-away condition.

The following diagrams dramaticallyillustrates run-away in high wind. Thefirst graph shows the power curve for the600 kW wind turbine as a function of theblade rotational speed. The bottom curveillustrates the normal power curve con-trolled by the generator, at a blade rotati-onal speed of 27 rpm. The three othercurves show power production at 30 rpm,40 rpm and 60 rpm.

Power curves at different rotational speeds (rpm)

CONTROL AND SAFETY SYSTEMS

Po

wer

(kW

)

Wind speed (m/s)

60 rpm

40 rpm

30 rpm

27 rpm

21

At a wind speed of 20 m/s, a wind turbine will normally produce slightlyunder 600 kW. Allowed to accelerate amere 10% to a blade rotational speed of 30 rpm, it is then able to increase powerproduction to 1.000 kW. At a bladerotational speed of 40 rpm the powerincreases to 2.000 kW and 3.300 kW at60 rpm. At a wind speed of 25 m/s, ifthe blades were permitted to rotate at aspeed of 60 rpm, the power productionwould be as high as 5.400 kW.

The second graph illustrates just howrapidly the blade rotational speedaccelerates in a run-away situation. Aftera mere 0.6 seconds the rotor speedaccelerates to 30 rpm, and after 2.5seconds the blades achieve 40 rpm. Asnoted above the power output at 40 rpmis 2.000 kW, an output far above theability of the braking system to restrain.

So it is vital that the safety systemsmust possess very rapid reactive responsein order to prevent such runaway.

95% of all deliberations behinddesign of wind turbine safety systemshave to do with this one task of safelyregaining control of the wind turbine,should the generator speed controlsuddenly become non-operative duringhigh wind conditions, and thereaftersecurely bring the wind turbine to a halt.

Basically there are two main methods bywhich one prevents a run-away:

1. Either one can prevent that theblades are actually able toachieve this increased powerproduction under this con-dition of rapidly acceleratingblade rotational speed.

2. Or by some other means onecan prevent the rotational speedfrom rising to an unacceptablydangerous level.

Here we have the principles for the use ofaerodynamic braking (1) and the mecha-nical brake (2).

THE CONTROLLERIn one way or another the controller isinvolved in almost all decision-makingprocesses in the safety systems in a windturbine. At the same time it must overseethe normal operation of the wind turbineand carry out measurements for statisticaluse etc.

The controller is based on the use of amicro computer, specially designed forindustrial use and therefore not directlycomparable with a normal PC. It has acapacity roughly equivalent to that of a

80286 PC system processor. The control program itself is not stored in a harddisk, but is stored in a microchip calledan EPROM. The processor that does theactual calculations is likewise amicrochip.

Most wind turbine owners arefamiliar with the normal keyboard anddisplay unit used in wind turbine control.The computer is placed in the controlcabinet together with a lot of othertypes of electro-technical equipment,contactors, switches, fuses, etc.

The many and varied demands of thecontroller result in a complicatedconstruction with a large number ofdifferent components. Naturally, themore complicated a construction and thelarger the number of individual compo-nents that are used in making a unit, thegreater the possibilities for errors.This problem must be solved, whendeveloping a control system that shouldbe as fail-safe as possible.

To increase security measures againstthe occurrence of internal errors, one canattempt to construct a system with as fewcomponents as possible. It is alsopossible to build-in an internal automaticÒself-supervisionÒ, allowing the control-ler to check and control its own systems.Finally, an alternative parallel back-upsystem can be installed, having more orless the same functions, but assembledwith different types of components. Onthe 600 kW Mk. IV wind turbine, allthree principles are used in the controland safety systems. These will be furtherdiscussed one at a time in the following.

A series of sensors measure the con-ditions in the wind turbine. These sensorsare limited to those that are strictlynecessary. This is the first example of thetargeted approach towards fail-safesystems. One would otherwise perhapsthink, as we now have access to compu-ters and other electronic devices withalmost unlimited memory capacity, thatit would merely be a matter of measuringand registering as much as possible.However this is not the case, as everysingle recorded measurement introducesa possibility for error, no matter howhigh a quality of the installed sensors,cables and computer. The choice of thenecessary sensors is therefore to a highdegree a study in the art of limitation.

Rotational acceleration during run-away

Time after run-away (sec)

Rev

olu

tio

ns

per

min

ute

(rp

m)

22

The controller measures the followingparameters as analogue signals (wheremeasurements give readings of varyingvalues ) :

¥ Voltage on all three phases¥ Current on all three phases¥ Frequency on one phase¥ Temperature inside the nacelle¥ Generator temperature¥ Gear oil temperature¥ Gear bearing temperature¥ Wind speed¥ The direction of yawing ¥ Low-speed shaft rotational speed¥ High-speed shaft rotational speed

Other parameters that are obviouslyinteresting are not measured, electricalpower for example. The reason being thatthese parameters can be calculated fromthose that are in fact measured. Powercan thus be calculated from the measuredvoltage and current

The controller also measures thefollowing parameters as digital signals(where the measurements do not givereadings of varying values, but a mere anon/off signal) :

¥ Wind direction¥ Over-heating of the generator¥ Hydraulic pressure level¥ Correct valve function¥ Vibration level¥ Twisting of the power cable ¥ Emergency brake circuit¥ Overheating of small electric

motors for the yawing, hydraulic pumps, etc.

¥ Brake-caliper adjustment¥ Centrifugal-release activation

Even though it is necessary to limit thenumber of measurements, certain ofthese are duplicated, for example at thegearbox, the generator and the rotationalspeed. In these cases we consider that theincreased safety provided, is more impor-tant than the risk of possible sensory fai-lure.

Internal supervision is applied onseveral levels. First of all the computer isequipped with certain control functions,known as ÒwatchdogsÒ. These supervise that the computer does not make obvious calculation errors. In addition the wind

turbine software itself has extra controlfunctions. For example in the case ofwind speed parameters. A wind turbineis designed to operate at wind speeds upto 25 m/s, and the signal from theanemometer (wind speed indicator) isused in taking the decision to stop thewind turbine, as soon as the wind speedexceeds 25 m/s.

As a control function of theanemometer the controller superviseswind speed in relation to power. Thecontroller will stop the wind turbine andindicate a possible wind measurementerror, if too much power is producedduring a period of low wind, or too littlepower during a period of high wind.

A wind measurement error could becaused by a fault in the electrical wiring,or a defect bearing in the anemometer. A constant functional check of therelationship between wind speed andpower production ensures that it is almostimpossible for the wind turbineto continue operation with a wind measurement error, and the possibility of a wind turbine being subject to stronger winds than its designed windspeed rating, is therefore more or lesseliminated.

The third safety principle for thecontroller lies in duplication of systems.A good example is the mechanicalcentrifugal release units. These supervise

the blade rotational speed and activatethe braking systems, even if the speedmeasurement system of the controllershould fail.

A 600 kW Mk IV wind turbine hastwo centrifugal release units. One ofthese is hydraulic and placed on thewind turbine hub. It is normally called aCU (Centrifugal release Unit). Shouldthe wind turbine operate at too high arotational speed, a weight will be thrownout and thereby open a hydraulic valve.

Cup anemometer for wind speed indication (left) ¥ Lightning conductor (middle) ¥ Wind direction indicator (right)

Interior view of the CU

23

Once the valve is open, hydraulic oil willspill out from the hydraulic cylinders thathold the blade tips in place, therebyactivating the blade tip air brakes.No matter what actions the controller orthe hydraulic system thereafter attemptsto carry out, pressure cannot be maintai-ned in the cylinders and the air brakeswill continue to remain activated, until aserviceman resets the centrifugal releasemanually.

The advantages of the hydrauliccentrifugal release units is that it iscompletely independent the controllerand the hydraulic system. This ensuresthat a possible fatal software designerror, not discovered during designreview, will not result in a possiblerun-away of the wind turbine.

The second centrifugal release unit isan electro-mechanical unit, fixed to thehigh speed shaft of the gearbox. This isnormally called an HCU, where H isshort for Òhigh-speedÓ. Should the windturbine over-speed, two small arms arethrown out mechanically cutting off theelectrical current to the magnetic valvesof the air brakes and the mechanicalbraking system.

This is a so-called fail-safe system,where the electrical circuit must remain

switched on in order to maintain thevalves for the air brakes and for themechanical brake in a closed position.Should the electrical circuit be brokenbecause of a disconnection from the gridor as a result of a shut down from the con-troller itself, the valves will open andactivate the brakes causing the windturbine to slow down and stop.

The HCU is able to mechanically cutthe braking circuit, and thereby activateboth braking systems. The hub-mountedCU only cuts the blade hydraulic system.The HCU therefore is superior, howeverits successful operation is based in turnupon satisfactory operation of the normalvalve systems, while the CU has its ownextra valve system. Both systems thushave their own advantages and dis-advantages considered from the point ofview of safety.

Both centrifugal release units areadjusted to be activated at very nearthe normal operational rotational speed,therefore, on rare occasions, releasecan occur prematurely. This is notnormally the case in Denmark, butfollowing from unexpected power cuts atcertain foreign projects, causing theturbines shortly to operate in stand-alonemode, we have experienced release

activation. Otherwise centrifugal releasesystems are only intended to be activatedduring maintenance testing.

HYDRAULICSThe controller decides which operationsare to be carried out in the safety system,while the hydraulic system operates thebraking systems.

In a hydraulic system a liquid underpressure is used to move certain com-ponents. This liquid is called hydraulicoil, having a resemblance to lubricatingoil. The operating pressure is about1.000 Bar (one Bar is equivalent to oneatmosphere). The moving componentsare pistons in hydraulic cylinders. With apressure of 100 Bar a piston in a 50 mmhydraulic cylinder (similar to the unitsused in pulling the blade tips intoposition) produces a force of 2 tons.

The hydraulic systems of both the tip-brakes and the mechanical brake are alsofail-safe systems, i.e. hydraulic pressureis necessary for the wind turbine tooperate. The hydraulic system ensuresthat pressure is established when thewind turbine starts. It also releases thepressure when the turbine must stop.

Pressure is built up with a pumpcontrolled by a pressure sensitive switch.Following attainment of the requiredpressure level, occasional operation ofthe pump maintains the level. A reservepressure tank is also included in thesystem. This small steel tank contains arubber membrane separating the hydrau-lic oil from an enclosed body of air.When the oil is under pressure, this willpress against this body of air, which inturn will act as a kind of cushion givinga counter pressure, thereby enabling thepressure in the whole system to bemaintained.

The release of pressure from the tip-brakes and the mechanical brake iscarried out by the means of magneticvalves. These are held in a closedposition by the use of an electromagnetand will automatically open with a lackof electrical current. They are thereforeoperated by being simply switched off.

In order to avoid operational failureproblems that any one specific make ofvalve could possibly produce, twodifferent makes of valves from two diffe-rent manufacturers are placed in parallelHCU placed on the high speed shaft of the gearbox

24

in each of the two different systems forboth air brakes and the mechanical brake.Secure and safe operation is ensured evenwith only one single operational valve,and their functioning is checked at everyroutine maintenance schedule.

In addition the mechanical hydraulicCU is fixed at the hub of the rotor bladeitself. This unit is completely indepen-dent of the functioning of the magneticvalves in releasing the pressure in the airbrake hydraulic cylinders.

TIP BRAKESThe moveable blade tips on the outer2.8 meters of the blades function as airbrakes, usually called tip brakes.

The blade tip is fixed on a carbonfiber shaft, mounted on a bearing insidethe main body of the blade. On the end ofthe shaft inside the main blade, a con-struction is fixed, which rotates the bladetip if subject to an outward movement.The shaft also has a fixture for a steelwire, running the length of the bladefrom the shaft to the hub, enclosed insidea hollow tube.

During operation the tip is held fastagainst the main blade by a hydrauliccylinder inside the hub, pulling with aforce of about 1 ton on the steel wirerunning from the hub to the blade tipshaft.

When it is becomes necessary to stopthe wind turbine, the restraining power iscut-off by the release of oil from thehydraulic cylinder, thereby permittingcentrifugal force to pull the blade tipoutwards. The mechanism on the tip shaftthen rotates the blade tip through 90degrees, into the braking position. Thehydraulic oil outflow from the hydrauliccylinder escapes through a rather smallhole, thus allowing the blade tip to turnslowly for a couple of seconds before it isfully in position. This thereby avoidsexcessive shock loads during braking.

As previously described in the sectionon the hydraulic system, the constructionset-up is fail-safe requiring an activecomponent (oil pressure) in order to keepthe turbine in an operational mode, whilea missing active component (no oilpressure) activates the system.

The tip brakes effectively stop the dri-ving force of the blades. They thereforhave the function as described under

point 1 in the section dealing withproblems - to prevent the blades having agreatly increased power production withincreased rotational speed. They cannothowever normally completely stop bladerotation, and therefore for every windspeed there is a corresponding free-wheeling rotational speed. However evenfor the highest wind speeds experiencedin Denmark, the free-wheeling rotationalspeed is much lower than the normaloperational rotational speed, so the windturbine is in a secure condition, even ifthe mechanical brake should possiblyfail.

THE MECHANICAL BRAKEThe Mechanical brake is a disc brakeplaced on the gearbox high-speed shaft.The brake disc, made of steel, is fixed tothe shaft. The component that does theactual braking is called the brake caliper.Likewise this is also a fail-safe system,

hydraulic oil pressure is necessary toprevent the brake unit from braking.Should oil pressure be lacking, apowerful spring presses the brake blocksin against the brake disc.

Braking is a result of friction betweenthe brake block and the disc. Windturbine brakes experience large stressforces, therefore it is necessary to usespecial materials for brake blocks onlarge wind turbines. These are made ofa special metal alloy, able to functionunder high temperatures of up to700 degrees Centigrade. By comparison,the temperature of the brakes on a carrarely exceed 300 degrees.

The mechanical brake function isas described under point 2 of thesection dealing with the possibleproblem situations - to prevent therotational speed of the blades fromincreasing above the rated rotationalspeed.

Tip brake in function

The Mechanical Brake

A/S reg. nr. 67.911Fabriksvej 4 ¥ box 170

7330 BrandeTlf: 97 18 11 22

Fax: 97 18 30 86E-mail: bonus@bonus.dk

Web: www.bonus.dk

Energy A/S

¨

BONUS