PDF Viewing Options

Not Your Device? Search For Manuals or Datasheets below:

File Info : application/pdf, 22 Pages, 1.01MB

Document DEVICE REPORTWindmill From Nai Vol 5-1979
If the section on energy for the fourth Journal was

a bit thin, registering only our on-going and unalterable

opposition to nuclear power and offering no word of

our own work, it was because a report of our recent

windmill work was, et that time, premature. Happily,

that is not the case this year. Having found a reliable

pump made from a trailer tire that is geared to the

capacity of our water-pumping sailwing on Cape Cod,

we feel that we have developed a water-pumping or

;rr;gat;on system that is worthy of replication and

adaptation. The New Alchemy Sailwing is the subject

of the windmill article by Ear/e Barnhart and Gary

Hirshberg end was first published in Wind Power

Digest for the winter of 1977/78. lt summarizes the evolution of the mill over the four yews from its

inception to its present form. One notable windmill-

ish fact that Ear/e and Gary do not mention in their

sketch of windmill history is that England's Domes-

day Book of 7086 documents five thousand mills,

one for every f;fty households. There is something

rather nice about being pert of the renewal of a

technology that has served people well before, a

sense of commonality with the pest, perhaps.

The companion article to that by Ear/e and Gary

is by Tyrone Cashman and describes the mill that

he built for the Zen community et Green Gulch in

California, using the New Alchemy sailw;ng es a

prototype. In addition to the background and the

actual construction of the Green Gulch seihving,

he discusses the modifications he made to adapt to

conditions very different to those on the Cape. This

is interesting for us es New Alchemists because, like

Meredith Olsen's in the Aquaculture Section, it is

the first documented feedback we have had on a

second generation of the application of our ideas

and, es such. it opens wider channels for comparison

and critique.

The final article in this year's Energy trilogy is

by Joe Seele who has spent the last year working

with the HYDROWIND, the electricity.generating,

hydraulically-operated mill that Mew Alchemy has

developed on Prince Edward Island in Canada. Joe,

who has since joined us on Cape Cod, explains the

pros end cons of HYDROWIND I after a year's

study and testing. He discusses the machine in

some detail and evaluates not only this particular

mill but various theoretical approaches to aspects

of windmili design. He sees a need to make available

in writing a body of knowledge that has been until

now largely oral and, es such, scattered and lacking

in organization.





Over the l~asf scveml years WChave been investigating the applications of wind-powered water-pumping systerns. We have been particularly interested in sdilwing windmills which general&- can he consrructed of indigenous marcrinls with limited cquipmenr by people who arc nor trained specialists. In contmsf to more sophisriratcd `!esigns, sailwings can be adapted to placrs pwr in resources such as rural arens LXthird workl co~~ntrics. \Vindmills have been used for irrigaatiou for mo;c dlan ~wcI\-c centuries in areas whew cultivation would be otberwisc unfeasible.
At New Alchemy on Cape Cod we needed a waferpumping system for aquaculture projects and co irrigare the gardens. We wanted to design B windmill that could be constructed h\- a do-it-yourselfer using local andlor available materials. It was important rhat our wax-pumping system be simple and incxpensivc, require very little maintenance and bc storm-resistant. The windmill had to be adjustable for varying wind speeds and wind directions. If was essential that it be operative in areas of low wind speeds for it fo be broadly practicable.
Four years of esperimcntation and research, beginning with Marcus Sherman's bamboo/cotton sailwing windmill in Southern India, have culminated in the design and construction of tbc New Alchemy Sailwing which meets these objectives. Familiarly known as ,"Big Red" (Figure l), it was named for its first set of bright red sails. If pumps in winds as low as 6 miles `per hour and in gales of up IO approximately 40 mph. Although our water-pumping needs do nof extend into the winter, the Sailwing is operable throughout the pear. People with year-round demands need only take normal cold weather precautions against pump or water-pipe freezing damage.
On top of rbe 26.foot wooden lattice tower, a horizontal axle leads fo the junction of three steel masts (Figure 2). The multicolored Dacron sails are attached fo the masts by grommets and pegs, like tbc rigging of a sailboat. Elastic shock cords connected to the adjacent masf pull the sail roof auf fo form a smooth surface for arching tt c wind. The shock cords allow for self-feathering and easy furling in storm conditions. The sail tips are attached fo fixed triangular pieces at the ends of each mast.

The axle and sails are oriented downwind from the tower, eliminating the wed for a tail. Wind power is tranrfcrrcd along tlw rotating axle through a pair of scaled commercial bearings (Figwe 3 1. A steel disc crankshaft mounted at the base of the axle trnnsfcrs the axle rotation to the vertical motion of the pump shaft. Five distinct stroke settings are provided by holes drilled at different radii from the disc center. The assembly is centered on a rtcel plate turntable abovc the rower. The adjustable stroke disc is tenrered directly above a hole in the turntable through

1.hc windmill toww is made of tight 2 6 4 legs bolted tu buried sections of tclcphonc pole. Cxvcd wooden buttrcsscs add support at the bar< and two sets of latticcwork give additional stability above. A sccundary platform rests approximat+ halfway up the towr.
o+i,s of rhtS. "ilWi,,~ Wb,du,;ll.
A brief iook at wimlmill histoq, mdicatcs that the Ncn .Alrhemy Sailwing is a scion of considcnble hcriragc. The origins of windmills arc somewhat obscow. but primitivc horinmral mills arc thought to have bern cmploycd in serentb century Persia. Lcgcnds rccount that prisoners of the Grnghir Khan carried the idea of win<!-pmvcrcd grinding and warcr-pumping miils to coastal China. Tberc, horizontal mills with matted sails came into osr. `rhese primitive mills bccimtc obsalcrc by the end of thr twelftl: century as the application of Chinese rail-making increased the sophistication of windmill construction. The railwing had an unparalleled maintenance-free life-span due to its durability and simple, lightweight design. It gained widespread application throughout caastnl China. `The same criteria explain the ubiquitous employment of sailwings in China and Southern Asia today.
The European windmill developed independently of its Asian counterpart. `The first documented mill was used for grinding grains in Eng!and during the latter part of the twelfth centuty By the s~venteentb cmtq, due largely to the cxtensivc cxploitatinn of wind both on I& and sea, the Nethcrlimds bad bccome one of the wealthicrt nations in tbc world. Clotlr wils the commonly-used material for windmill sails during this period, reflecting its application on .sailing ships. Among the advantages of cloth were light weight, cast in handling, low cost and availability. Most importantly, when supported at tbrcc or more points, cloth forms a strong uniform surface for catching tbc wind.
These advantages hold true today, as is evidenced by the widcsprcad use of cloth for windmill sails. Currently, bzndcraftcd sailwing; arc employx~ in Crete, India, Ethiopia, China and Thailand. among others. Kccearchcrs ar Princcum recently hare dcwlopcd B two-hladcd high-speed aerodynamic Ihcroni RI sailwing for use in the United States.


,, lowspeed. eight-bladed Cretan sailsing and the higbspeed aerodvnsmicsll~-effjciettt Princetw model "A Windmill in India", the swond Jorrwd Using a bullock cartwheel rough@ one meter in diameter as
,,,~~,, the hub, he attached to it triangular wilwing frames mode of bamboo and nylon. A cloth sail adds strctch-
,,, ed owr the fratne to produce a rtablc, lightweight `Y airfoil. The rotor asrembiy was attached to a used ,,, automobile axle. Marcus msde zt turntable from hall
bearings sandwiched between two doughnut-shaped discs. The axle wzr mounted horixmwil~ on top of the turntable. .4 rudimentary "squirrel cage" asrcmbi> for bousing the drive chain and gasket pump was centered on the axle directly aborr tbc one-fourdiameter hole in the turnt&le.
The railwing war headed downwind to prc~nt th: bamboo poles from bending and striking the teak pole tower in monsoon winds. In this way the blades served i;s their own tail, trailing in the wind. In the process of the subsequent well digging, the mill was used with a pulley assembly to rairc soil and rock from the 20.foot-deep well. Because of its high starting tcxpe at low wind speeds, the mill proved well suited for yewround irrigation in Southerr. India.
Back at New A!chemy the foilwing summet. hlarcus,with Earle, gave the sailwing concept another try. With lumber and hxdware, they brilt a durable prototvpc well z!de to withstand the often blusrey Cape Cod climate. Fur a total material cost of $300. they developed iln IX-foot-diameter, clotI, sailwing capable of pumping 250 @Ions per !wur in 6 mph winds. Tbrec tapered cloth sails, supp~wted 1,) `, tubular steel masts. extended from a triangulw ply wood bob. A mowable boom war secured at the

root of each saii by a lather strap to further relffcathcring. Lung metal doorsprings conncctcd cxh of the three sail booms. In eaarly tests, the feathering mechanism withstood a force-nine gale.
A used automobile crankshaft formed the hub a~,i! crank. The assembly turned on a ball-hexing turntable which allowed the windmill to seek a downwind operating position. A recycled piston rod on the crankshaft transferred power to a reciprocating vcrtical steel pipe pomp shaft. The shaft operated a high capaci~ piston-type pump bclo~r. The entire strut-

ture was mounted on a firml!ybracrd cigbt-lrggcd wooden rower.
The windmill supplied water ro a \crics of twrnt! rmdl ponds urrd in our midge expcrimeuts. I* nil* opcrationd in high winds. nltbouyh the cloth sails were rcmwcd in sewre storm comiitions. `Tbr cotton sGls were later replaced by I)acron iRi, which is longer-lived, haldr its rhapr better, does not sbsorb water during rains 2nd is stronger and lighter than cotton. On pr&ninary testing, Marcus frmed tl~cperfwmanrc of the mill to be significnntl~ lower than its calculated pumping capscity. A double pump was used. This and subsequent models empioyxl downwind sailwing blades which minimize the chances of the sails t;mgling in the tower while feathering ami riiminatr *be costs of a Inrgc fail.



`fhc currcnr model was built in 1975 and incorporates nmay of the fcarures of its proro~pes. Scversl new ideas wrc tried. An ratensian shaft was added to position

the hubs and blades further from the tower. WC had noticed that the slip-on sock-like sails frayed wbcrc they

~, were wapped around the blade shaft. Traditional sail ,, makers advised us to attach the sails with grommets and `, pegs and to position a subilizing cable from wing tip

to hub to prevent flexing of rhe blades. In addition. we

added a simple spring-feathering device to each of the

,,,,, sailwing tips.

I ::

The double pumps used previously with the prata-

;;I,,`, type proved undersized for the strength of the new

,:i', model, soa higher capacity and more compact diaphragm

pump was oied. Tests were also carried out with a


deep wooden piston pump like a marine bilge pump, but the diaphragm was more reliable.

The auto crankshaft that made ui> the hub and

crank on the prototvpc 1~s found to yield too

`small a stroke for the mill. so Mac Sloan, an engineer

: who adviscs us on windmill problew, de&cd an in-

`,' genious disc-bearing assembly to driv., the pump

(Figure 3). The disc/bearing arrembly functL:rs as a

crank with a variable pumping woke. Sealed cam-

mercial bearings were added to the windmill shaft

at this time, as the original homemade bearings wore too quickly and demanded frequent lubrication.

Curved buttresses were attached to the tower legs to support the additional wight of the crankshaft, extension and other hardware which had been added subsequent to the original design.

Several other features have been improved since the spring of 1976. Strong elastic shock cords bwe replaced the door springs used for self-feathering, `:" resulting in increased flexibility and smoother nil i ~' motion. The shock cords are easier to maintain and

Self-fmtbering: In nwm*l winds (0-I 5 mph) the taut sails catch the wind and drive the pump. In higher winds (15-30 mph), incrcascd forces on the sail press downwind, stretching the elastic shock cord mti allowing some of the wind to spill past the sail. This automatic feathering results in continuous pumping in higher winds without destruction of the blades.
Reefirrg: During very high winds and gales (: 30 mph) we protect the windmill by reefing the sails. The blades are stopped by hand from the mid-tower platform and each elastic cord is unhooked from its metal mast Gtachment. Each sail is wrapped around its own mast pole several times and then bound by winding the cord around the sail and honking it. This arrangement leaves only a small triangle of sail exposed at the outer end of each mast, which in bigh winds is often cnaugl~ to continue pumping.
A`fjustmrwrs: Several *djustmcnts can bc made to adapt the wimimil; to diffcrcnt average winds or pumping requircmmts:

The Pur)p Stroke -`The pump stroke of the windmill hias fire settings depending on file attachment point of the pump rod to the crank disc. In a given wind, the windmill can perform a fixed amount of pumping work which ran take the form of a low lift of high ~olumc or a high lift of low wlumc. Our winds areragr 8-9 mph in summer and our need is to lump the largest volume possible to a height of four feet. Our custom-made diaphragm pump lifts ,875 gallons per two-inch stroke in average winds. In high winds we hare measured 700-800 gzdlons per hour.
Combinations of stroke-length, pump wlumc and height of lift musf bc developed for each application and site. Sailwing windmills such BSthis one have been used with piston pumps for 20.foot lifts/ IO-foot heads and with chain and trough pwnps in Southeast Asia for low-lift irrigation.
Shock Cord Tension -The responx of the windmill to varying winds delxndr on the tension of the clack shock cord holding the snil fauf. A mild tension aids operation in IOU winds by creating a steeper angle "f attack as the tail part+ feathers, bur spills most of the higher winds. A rrrcmg wnsi~m reduces starting ahiliF in low winds bv creating a flat attack angic. hut spills less cner~~ in high winds.

Ulade Tip Angler The angle of the sail fo the wind at the rip of each blade also affects the windmill in varying winds. A steep angle creates high starting torque but limits rpm once the windmill is turning rapidly. A tlar angle give less starting torque, but once started ca~scs a greater rpm in higher winds.
In our aquaculrurc circulation applicaian, where continuous pumping is idenl, a mild shack cord fer,sion and a rtecp tip angle of 40° results in low- `. wind starting and pumping as often as possible.
WC offer these simple cautions in working with rhe mill:
Stop the windmill and wear a safer? harness while on the fewer. - .Avoid allowing the windmill to free-wheel without a pumping load. -Protect water liner from frcoing for winfcr pumping. We are, ovcrz~ll, well pleased with the Ncn Alchemy Sailwing It is beautiful, functional llld durable. It performs well fhc rark wc ask of it. After four yenrs, it mccfs ihc objcctivcs wc originally postulated and. in terms of cwt. I;d~or, rfficicnq and usefulncsr, W~CII contriwcd with mow standard rcreiwch awl dcvclopnw~~f mod& it seems gcnuincly to qualify as appropriarc tcchnol~~y~.


The J0um.7~of rhe New Mcbemirrr

The Green Gulch Sailwing .-T~yr"CrarerlH,zna

The Green Gulch sailwing windpnwcrcd irrigatinn pump was concciwd of and implcmenrcd jointly by fhe Arca Fwndution, the Zen Center and the NW Alchemy Institute. For the Zen Center it rrprcscnts the first step in the integration of wind cner~y info rbeir stewardship of their valley.
In the long range, Green Gulch Fam), as a permanent agricultnral/natural cvcle based community. will reap increasing advantage in food and energy from rhc rccycling of wafer, of human, plant and animal wacfc marerial, and a p-agmatic and benign use of wind. sun and gravitypowerrd water Row.
The community is destined to hecomc a model of gentle stewardship of land - receiving from it a true abundance. while under its guiding hands the roil, ponds, sardens, hillsides and mtaf landscape become richer, more biologically divcrrc and m~tre beautiful year by yxr.
The Gwen Gulch Sailwing provides irrigating wnrr, a constant reminder of natural furccr, and basic cxpcrience in wind technology for those who helped design and huild it and for those who will operate aad maintain it. Further wind projects for grcy-water aeration, milling, winnowing, water pumping aed

electrical generation will casil\~ be incorpuratcd by the communiry through the prnctical cxperiencc and urxlerstanding this windmill is providing.
Bcrides the practical and economic wluc of this mill, there is an aspect of cqunl importance and equally npprecinrcd by the communit)c the mill is beautiful as it wrns. The evening sun glows through its sails. It graces the vnllcv 9s it responds m gentle movcmcnfs of air. And it is quirt - 8s it must bc (and few Gndmills are) for its location within yards of the meditntian hall in B mcditativc community.
As a research entity in gcntlc rechnologics, NC\\ Alchemy sought two goals with this sailwing: (a) to crcatc a wind-powered irrigation system carefull? tuned fo the bio-region and cvcn fo the microclimate of Green Gulch Farm and (h) to advance its rcscarch on the pmhlcms of incrpensiw. durahlc. simple, high volume/low wind windmills - mills that can also withstand relatively high winds withwot human inwrvenficm.

In dcsigoing for the Green Gulch vatIcy, there is the advantage that, for thrcr sca5om of the year - spring, summer and fall - when most of the irrigating must be done, the winds arr quite tame. Rarely will a wind come up that is over 20 mph. In winter, storms cm hc cxpcctcd with winds to gale force and beyond.
The weragc recorded wind in the lowest wind soawn .+ril/Xla;r? - at the windmill site was 4.44 mph in 1977. The design of a mill xvhich will do regular and significant work in a 4 mph wi.xd rcgimc and still weather Gnds of 30 tu 35 mph without lhumim attcntinn was our goal.
:\ windmill that is available commercially, such ~$5the American multi-hladr sold by Aermotor, Ihmpstcr, is the end pro;ect .)f a design that cca5cd evolving significantly in the 1930'5. It begins pumping in light winds due to a ;-to-l down-stepping gear ratio, giving it added torque to owrcomc start-up inertia, static head and friction in light winds; but it pays for that torque with a loss of Z/3 potential valumc of flow, extra friction in the gear mcchanism and considerable expcnre intrinsic to the praduction of rtrong, slowspccd gear wheels.
The rotor, which is made of rigid metal blades, must bc turned out of the wind when winds arc too high - and cxtm mcihaiii5ms arc provided to pcrform this function either automatically or manually. The construction is of metal.
The advantage of the American multi-blade design is that it is proven to bc wfc and rcliahlc, rcquiring little operator attention. These mills arc especially well-designed for remote stock-watering apcratiunr.
We priced the largest Dempster. which has a rot", diameter of 11 feei and an appropriate tmvcr for our sitc,and found that the combination turned out to be over $4,000, not irduding the pump.
The windmills fwnd on the shores of the Mcditcrrancan arc designed for the c""sm"t. relatively light winds of their region. Cl& is used for sails and wooden rpars, windshafts and cvcn wooden hearing5 are tradititmal. These mills arc inexpensive and rcsponsirc to light winds, hut their sails must hc furled or rcmorcd before a storm or if the mill is to be left unattended for a long period of time. Such constant watching and attention arc not to be expected of the American farmer or horticulturist.
For severill years, New Alchemy has bacn cngagcd in the dcvclopmcnt of simple. Imv~cost sailwing water-pumping windmills which would be adaptable to diverse wind regimes and which would combine the best qualities of the American multi-blade and Mcditw rancan mills. The Green Gulch Sailwiy may not he that model yet, but it is, on several fr *.x5, a large step

Parallel to the dcv~.Iop~~~cnt "f Nca ,\lchc",y Sailwings. il sillzll group of "Gssicrnarics i" Urn", Kthi"pin. have hccn dwzloping wd test~ing a vuicty "f sailwing mills for lo\\-level irrigatio" pu"lping. The rcwlts "t their cxpcrimcnts arc rccordcd i" a bwk by Percr Fraenkcl of the l"termcdiatc Tcchnolo:y I~c~4opmcnt Croup in London: Fwdfi-om M'bh~illr (London: ITDC, Parncll House. 23 Wilton Rd.. SWIY IJS. 1975). These practical cxpcrimcntcrs found 50°C of tbc h'tw Alchemy idea and dafa useful in their w"rk and. in turn, wmc of their results have been helpful in the dcrign of the Crccn Gulch Sailwing.
The mill site was chosen by scvcral criteria: 1. Close enough w the water source that the pun,p at the base of the "xii1 could bc no m"rc than ap proximately 10 feet above the water swface in all seasons. At this distance a pump which "pcmtcs frcqucntl)~ enough t" keep it5 leather piston rings moist will not need to be primed. This is essential for a windmill pump which of "cccsaity stops opcmting far periods of time when tl,e wind dies entirely. 2. Good xccss t" tile wind. The site is one hundred y"rd5 from the nearest obstruction of its "WI height, n run of windbreak trees, and is dircctl) hchind a gap in the trees in the direction of the prevailing winds. 3. Enough space around the tower h~sc to albw additional dcvices (air-compressor. winnower. thresher, grinder, etc.) to bc conncctcd to the present pump shaft shmdd need dictate and appropriate devices be found or built. 4. The mill is located within sight of the office, the dwelling area and the fields. It is not wise to locate a newly designed windmill met of sight. Young people arc tempted t" climb it while it is working, if no one is looking. Also, if any aspect of the mill needs "ttcntio", it can receive it hcfwc the mill damages itself. 3. The site is twt of the way of walking and gardcning traffic.
8. Tower
The tower is designed t" "se low cost, light weight materials, pint 2 x, 4'5. Tower strength is achiwctl by spreading weight and windprcssurc wcr tight legs. The tower fwadntion is tight rccylcd rcdwwd railroad ties, pai"tcd with c~cos~nc and b"ricd three to four feet in the cl"!; ~4. Strcnyth n"d conw,ic"cc in mill mai"tcnancr is provided by tw" circular platforms diGding the wwcr i,, thirds. Rigidity is oh-

4. To providr the chassis for att tower-top mcchenisms.
CONSTKVCTlON WC calculated that a free-floating tear axle from a
314 ton truck. uhen set ou end, was capable of mceting all these specificatiotls with a large margin. Such axles arc available in scrap metal yards throughout this coxnn~ and in many others around the world. The cheapest and probably highest quality turntable we cculd have wed was this recyctcd truck axle. At the tow revolutions per day of this application, there should bc many ?ears uf life left in it.
Before the axle can hc used for a turntahlc. the powrr shaft is rrmorcd ftom the ax!`, teaving B 3-square-inch passaycway in thr ccntcr for tlx pump shaft. The brake parts WC removed and a 20.inch diameter disc of steel, % inch: thick, is wctdcd as a hasc for the axte to sit on. Braces undcrncatb the wooden tower platform give added sttength against lateral pressure from the wind.
Care must he taken to protect the hearings from rust through exposure to the elements. We ptotected the beatings with an aluminum cowling which covets the entire transmission. Another protective measure used by the Fataltones Institute is to grease the bearings well with hoat trailer axle grease (designed to be immersed in water) an2 cover the opening with the appropriate size jar lid. punctured so that rhe cut edges lead rain watct Jnwn the pump shaft and away from the hearings.
Purpose: to transfer power ftom rotary (wind shaft) ,to reciprocal (pumpshaft) motion with maximum effic&y and durability and minimum cxpensc and complerity.

`This is accomplished with a I+inch diamrtcr stcct disc. !i inch in tbirkncss. wclJcJ to a 2-inch stcc\c. machincJ with it kryway xnd drittul for a hardrncd bolt. `This disc and stccvc fit over tbc end of a *-inch cold-rolled stcct bu. 32 inches tong, wbtch functions as the wind shaft, baring at1 the weight of the rotor and transferring tbc nwtion of the rotor to the disc.
Tbc disc is drilled with thrrc l-inch holes, to aa) of which a I+incb long. % inch diameter cold-rolled srcct hat. fitted with rod end bearings on either end, can be attached. This connecting rod is further cncased in a section of galvanized steel pipe for greater strength.
The connecting rod attaches to the xrticat pump shaft at a junction point comprised of a steel box welded of `A-inch plate with two industrial castors welded to each side. Tbcse pwvidc a rolling lafcral bracing which forces the side-to-side motion of tbc disc md connrrting rod to be translated into purr verticvt motion. `The castors rile in tbc pathways of srcct channcts welded to the I beam base.
The tratsmission and turntable are designed and built for great strength and durability yet simply enough for snmeonc with wvetdingabitity toconstruct.
Ii. Rotov
1. The rotor was designed to function downwind from the tower without a tail or rudder. This saves weight and expense, since a tail capable of keeping a 20.foot diameter totot facing the wind must bc very tong and large. In addition. sails which are able to stretch hack away from the wind during gusts are in danger of rubbing against the tower and catching or snagging when winds are strong. If the rotot is downwind from the towet, the stronger the wind, the further from the tnwcr the sails strctcb.
2. To overcome the disadvantage, relative to an American multi-blade, of one-to-one gear ratio in extremely tight start-up winds. the totot was made larger. Vety little is added to the expense of a sailwing totot by extending the tnzsts, for example, from sewn to ten feet. However, since the atca of a disc is quadrupled when rbe diameter is doubled, the increase of wind energy available to a 20.foot rotor is double that of a 14.foot totot. This increment of wind cncrg) is furttwr augmented by the inctcnse in ii&tanicat advantage of a IO-foot lever wm aver a 7.foot atm. The expeasc. complexity and weight of a garbox is thus ctiminatcd.
3. Tbis saitrving rotor is dcsigncd for bath II)\\' and high winds:

Four sails. cacb sail with win&catching area of 2, sqllarc feet. The choirc to spread X4 sqwrc feet elf sail, hut "`,I n,orc. was made when ttw ideal sotidit~~ factors f"r slou speed water pumpers were balanced against tbc need tu limit dangcmus Levels of drag in high winds. The erpcricncc of the research team in Omo, Ethiopia. that rhc coefficient of power (owiltt system efficicnc!;) is more a function of windspeed than of numhcr of sails deployed was also a cm~sidrration.
More experimentation weds t" bc done 011best tip speed ratios and solk!ity factors for sailwit,g wiadmitls. The senson-II)-season f"nctio"ing of this milt and testing of different pumping and air-compressing tasks will help in the further refinement nf design.
b. Duigu for bigl, zinds
The prima? design feature of this lowwind-sensitive. modcrate-solidity sailwing to withstand the esponeatiat incrcascs in energy in high winds is the flexible shock cord sheet connecting the outboard corner of the sail r"ot to the successive mast. The shock cord hrings the whole sail (from root to tip) into tension and, in

The second function "f tbc shack cord is to allow tbc sail t" strctcb back o"t of the wind when a suddm large gwt bits it, and to hc"d back spit!ing tbc majority ot thr rncrgy it is receiving in winds above 20 mph. The wind-spilling ability dors not interfere with the regular pumping action of tbc mill - since, 110matter bou far dowwind tttc sail is stretched by the wind, it always retains enough cncrgy 1" pump at an efficient rate.
The second &sign factor allowing tbc sailwing t" withstand high winds is the fact tbat ctoth sails arc flcsihle and ca" bc rccfcd (a five-minute "pcrarion) during stormy Sc~soIIs. When reefed, approximately tbrec square feet of sail remains deployed at tbc mast tips. In this condition, the milt will weather gale force winds and continue pumping the whole time.
A third factor of design for high wimls is the abitit!~ to set the tips of the sails at an angle that is aerodynamically inefficient. thus creating luffing of the sail tips while the roofs are being driven heforc the wind. The net effect "f this precautio;xtry tcchniquc is to prevent the rotor from overspecding ill high winds. crcecding totcrahte centrifugal forces as wrtt as tolerable stroke raw of the pump.

4. The rotor masts arc coned downwind by wircrope stays. -,.tle purpose O^f cantng IS to create extra rigidity in the mast. Coning is done downwind so that the effect of the pressure of the wind, mbich would he t" bend the mast further, is c""ntcracted by the tendency of centrifugal force to straighten it. 5. The rails are desigtted with 17.inch tips widening to 3%inch roots, with a catcnary curve ctit in rhc trailing edge. These proport'ons spread the energy of the wind relatively evenly "ver the length of the mast while the catenary curve prevents cncrgy loss and noise due to vibration of the roach.
6. The tip booms are a new design worked o"t to provide extra torque bt the tight wind season by aitowing a steeper than n"rnmat tip boom angle at startup. Once the rotor is moving, the tips arc dcsigned to return automatically to the corr`~ct a"glc for efficient operation in motion. The final imptemcntation of this tip design has not hccn accom plished as of this writing.

1. The masts arc standad 1% inch, 10 foot trmg TV antenna toasts. These were chosen for their length. tow cost ($3.75 when purchased in June, 1977, as opp"scd t" 537.50 for aluminum spars of adequate strength) and for the ease in obtaining replacements with uniform weight.
2. Masts arc stayed, front and rear, by l/8 inch galvanizcd wire rope, 7 I 19 strands (for essential flcsibility). At the 213 point on each mast. a wire rope connects it to the preceding and succeeding toasts. Thus, itt wery direction where stress is encountcrcd, tbc mast is hraccd. The stays provide a comhinativn of comprcssivc f"rccs ca"si"g a curved downwind cone.
3. The r"t"r is rcmavablc from tbc windshaft by four bntts. A Z-1116 inch It) steel pipe fits.xsaslccvc over thr Z-inch windstuft. To this arc w&led four Y-inch sockets l-l/S inctt ID int" which are inserted the f"ur masts. Mast ends have brarcd beads f"r snug fit.

4. `The new tip md~aniw~ i* rrutcd fnm iin intlllstriill ci,stor ,vith \VhPCI Kll",\cd. wl4`kYl to l-l ix inch 00 water pipe which inserts into the mast cd - and wrldcd as a `r to a picce of light conduit to form tip boom and tip angle control weight mount. The tip boom rotstcs 360" "11the cast"r double bearings when the sail is "c)t lashed t" it. It call be set msnwdly at a given angle- by inscrtblg a Pi", "r lcfr t" wmk aun~maticall~ by the actiun of the \w(ght. 5. Sails are made "f 5.+-"""ce &cd Dacr""!Ki p"lyester cloth with Dacron `E'cord sewa i"t" the traili"y cdgc f"r strength and rigidity. Szil is lashed w tbc mast I,>~Dacr"" ~xi cnrd witb llcxihlc tubing imcrts wbcrc the cord parses over the sharp edge of a grommet
I'urporc: to recrive cncr= in the form, intatsity and speed B 20.foot railwing dclirers and to translatr .L. energy into the m~rement of ~atcr t" a higb head.
Low speed demands a pasitiw displacement pump. Centrifugal pumps rcq"ire a" the "rdz "f 1,000 rpm for cfficicnt functioning.
Reciprocal acti"" demands either it diaphragm or pistol1 pump.
High head eliminates diaphqm Pump. It has t"" many interior square inches fur a head that produces 75 puunds per square inch.
A piston Pump is ideall>~ &ptcd t" the vertical mud"" and lengtb of st~nke that a windmill can be made t" deliver.
Experimentation in Omo. Ethinpia, cuncludcd that two ringi:le&ting. commercial windmill pistan pumps, operating "n a Irver arm such that one was voiding while tie other was tilling and vice vcrs~. increase the v":ume of flow a windmill cm produce by virtually 100 per cent. They discovered that x mill is not rignificand\~ slowed when forced t" P~m~p on the down stroke as well as the up. We were not able t" discover a commercial pump manufacturer who made a double-acting pump adapted to a windmill, 5" we decided t" build our own. In the ITDG report "n the Ethiopian research, a design is offered for a double-acting pump which, at the time of the report, bad nut been built. The pump for the Green Gulch Sailwing was built frnm that design. In initial tests the pump has proved vey well adapted t" the windmill. Further testing and shakedown must still be made.
An added advantage was seen immediately in tbc double-acting pump: the "utflow ef water is smooth. with wry little pulsing. This minintizcr the danger c"mm"n t" reciprocal pumps used far high heads, in that they are subject t" return shock bales frum their own pulses, causing rtnin and sometimes damage. This is called the "water hammer effect."

Since we have three scrtingr f"r strakc lea@ "" tbc crank disc, 5 imbcs. 8 inches and I2 inches, a burr "f 2 imhrs may turn WI f" bc right for such high prerstwcs. A head "f one hundred-fifty feet. creating 75 psi of static prcssurc 011 the fzcc "f the pist"" which is 3.14 sqm~c irwhcs. is resisting tlw piston's motion with 235 pounds "f f"rcc. !lot cwxsidcring friction and inertia.
From testing so far WChave ~1 doubts that the mill will pump I" the fifty-two-foot h"lding tank and the cightyfout resewnits as designed. In all likelihood. we shall be able t" set stroke length for higher volume in winds "f the IO mph mngc.
The pump was built from standard parts nnd pieces available in plumbing 2nd pump-and-well shops. The barrel is a I(,-inch section "f 2-inch diameter PVC water pipe. Male adapters arc glued t" this. top and bottom, and screwed int" galranizrd irun T's at each end. Tw" I-inch galvanized pipes witb in-line check valves arc conncctcd t" the T's at carh end.
The sucker r"d pures tbrougb a packing gland which prevents water, when under pressure in the upward directi"", from leaking. The piston is equipped with tw" cupped leathers - "nc facing upward and o,,c duwn.
With this design it is pursible t" discunnect "ne entry way into the pump s" that it pumps "n only ""e-half of its total stroke - and in place of the half stroke - connect another device, such as an air-compressing bellows for grey-water aerati"n.
This windmill is the m"st advaaccd of tbc New Alchemy sailwings "n se"eral c"""ts:
1. It "tiliees a far lcv, cxpensire turntable of a quality equal t" the IO-inch ID twntablc benring used on PII tcc~nt NAI mills.
2. KY the ndditic," of a wnnccting rc,d, cba~mcls and rollwa, tbc pump shuft runs vrrtically frw tbr"agh"et its length. `This nllaws tbc sbnft t" bc guided without significant fricticln nnd cnablcs the mill t" dcliwr full power on tbc mmprcssion stroke, a nccesmry c""dition fur the "se "f a double-acting pump.

`The c"nnccring rc,d appxatw als" makes p"ssiblc greatI!, bxwascd Icn~th "! stmkc - wbicb allows high

\`"I"",c from a ,,atri,\\ bore p"mp. 3. The tower is designed witb bowed ETOSSbraces

under comprcssivc load and a second platform These eliminate tbc nerd fear the iaerior guy wirer. Tower work is tbw more c""w"iew
4. Power i" low winds is i"crc,sed by rbc addition "f OM mast if""< instead of three) znd by a wn-

siderablr incrcasc in sail area f"r each sail. 5. Sail design is impruvrd: - KY tbc wc of j.Ct,tmce DacmniR! rails instcad of

3.8.""ncc. which slwuld increase the !ife "f the sail. Inrrc:~d of a root b"om, as on carlicr NAl mills. or a \im/~lc shock cord as "n the m"st rwcnt one, the Grrrn Gulch Sailwing has a fiber&us batte" sew" into tbc root cdgc of tbc sail. `The shock cord is c"nnccrcd dire+ t" this and to the sail, and an additional cord leading "ut to the snm:t of the r"t"r from this point balds the sail root permanently at 22" (except when high winds increase the angle) and keeps the batten bowed to create a correct airfoil in faint breezes. - A catenarv cun-e is reintroduced (early NAI sailwings had it) and a firming Dacron( Rj cord is EC~II

into the nailing edq

6. A tip boom mechanism was designed for the New Alchcm> Ssiiwiq tbrce years ago, which was to alluw the tips a steep pitch out of the plsne of the llisc for incresscd start-up turque, then would, as centrifugal force iwreased, move the tiP int" the wind for best

operating angle. The mechanism was nut reliable as canstructed. although the concept may not have bceit faultnq. In addition, it wzs found that in the brisk and c&c winds of Cape Cod B wind-spilling tip mecbanism wzs of gtestet USCtbz z" efficient tip ang!e

so the mechanism \va~ turned backwards t" increase the angle out of tbr plane of the disc with increased

Fpccd. However, this mechanism wa, in all events.

too complex.

The ~tos~"picall~sonuolled

tip booms, designed

for the Green Gulch Sailwing. cuuld prove to be an

advance. They are simpler in concept and design,

with only "ne moving part. Final experimental determinatmn of the correct weight for the tip and

construction of the weight-muunting device were "at achieved as of this writing. This was simply due to the abscncc crf windmill pcrs~:nncl f"r the first month after initial testing was t" have begun. lt was "at possible to determine the correct weight fr,r the gyroscnpic lever without rctual tests "11 the machine. Experiments with a, small-scsle model of the tip-boom mechanism weti:: successful.

The tip booms as constructed can als" be set manuaily at various angles, either for greatest efficiency and power or for different atno"nts "f

drag t" prevent averspeeding.

7. With tbc i~~twdwtiw~ "f tbc four-martcd rotor. it bcc"mcs possible to rcinf"rcc each Imast laterally by connecting it t" the prcccding and :,IICcccding omxts with l/X-inch wire ro~c at a p"inr 2,3 mast length fnrm the r""t.
8. `Tbc dwble-acting p"mp is an imprwemcnt over all pumps Net\ Alchemy had trictl with the sailwiny up to tbc spring "I 1977. It all"ws lbighhad pumping with low back prcrsuc for start"(1 and a'l"\vs il range of \."I""ICS per stroke from 0.14 gallons LO0.32 gallons. It is tailored to the ki"d of power this sailning produrcs Beth in form and intensiq. It is specifically dcsignrd f"r biphhead pumping.
Ihc to absence of pcrsmmcl after tbc c"mplerion of the windmill c~~~str,~ti"~t. "nly minimal rating has been dune. It was diswvcrcd that, with an 8inch stroke, pumping nine fcrr up fmm the pond and directly "ut into the squash patch at the bare of the tower, tbc Inill began pumping in 6 mph winds and continued pumpiq in an average of 3 mph. It "red not bc said that this is a" extraordinarily light wind fur tbc pcrformi"g of any useful work.
`Testing of revolutions per minute on two occasions h.u shown that, with a modrratc load (12.inch stroke pumping 35 feet above the pond) the tip sl'eed ratio was 3.6 and did not change as the windspeed changed. Tip speed wtio is the ratio of the speed of the wing tips t" the speed of the wind. In this case. the tips were travelling 3.6 times faster than the wind. This is a somewhat higher ratio than normal f"r a water pumper, bat such speed is P disadvantage only when it translates int" 1"" little torque for light wind start-ups "I wbcn it sends tbc mill into nverspced in bigb winds. Witb the acrodynnmic brake (tips set ii: lncfficient mgle) and sb"ck ~0x1 spillin;: of cxccss wind, the mill should prove t" bc protcctcd against overspeed in winds up to 40 mph. As f"r starting torque, we have already see" that it is excellent.
It is to be expected that, when the mill is pumping up 50, 80 or eve" 150 feet, the tip speed rati" will Lv reduced to nurmal levels, between 2 and 3. Tested tip speed ratio with no load (pump disconnected) was 4.7
Informal observrtion ix winds cstimnted between 40 and 50 mph reveals that the rotor rpm reaches a peak at around 55 and then riows down as the winds rise further. This is due to the loss of airfoil wbcn the shcrck cords are significantly stretched.
A rough idea "f tbc wlumes of water flow to bc erpcctrd in various winds can be bad frum tests with a 12.inch stroke pumping 35 feet above rho water *""Kc:

68.42 9.27
56.69 4.77 18.83

New Alchemy Hydrowirtd Development Program
The innovatiw aspect of the Hydrowind sysrem is the use of a hydraulic pump af rhc rot> of the mill to receive power from the rotor and deliver the tmwx in the form of txcssurizcd hydraulic flaw to groundbased equipment. Of the many altcrnaivc means of using hydraulic power OP !!K ground, electricity gencration was chosen for simplicity. versatility and comutcmcntaritv fo mosf cquitmxnt aused in modern Lttings. '
To gcrwatc electricity. a hydraulic mofor driven by the flow of fluid from tbc top of the mill turm a permanent field, brushtess electric generator. The alternating ~orrcnt from the gcncrator wwics in both wltagc and frcqucncy with windspccd changes and is incompatible with the fixed volwgc (115 volts) fixed frcqucncy (60 cycles per second) utility lines. To overcome this incompntibili~, an electronic ~ynct~ronous hwcrtcr trnnsfonns the wind.gcncratcd etcctricicy to tbc proper vattagr and frequency and combines it with electricity from the usiliq. The installation obviously is not thcrcforc designed for stand-alone operation, but instead substitutes wind power for utility power in the amounts available from tbc milt. When wind generation exceeds the Ark's consumption, surplus power goes out through the utility tines and becomes an input fo the power grid. Under these conditions, the windmitl is like the many other electric gcncrators linked together by the power grid, each adding its contribution fo mea the overall demands of houses and industries tied 10 the grid.
To limit maximum rotor speeds, the machine uses a t~~drauticatt~-dri*cn piston co control the pitch of the rotor blxtcs. Tbc piston. in furn, is confroltett by P bydmutic valve op~rz~t~d by centrifugal weights (Figure 1) in a mcchanisnl anatopr;os to the flyball va1vcs that vent srcam fo govern a stcilm engine.

To gain pcrspcctive 011tbc rclativc merits "f tbc existing Hydrewind mxbinc sml pr"p"scd modifications to it, we nrcd fu cxplorc some "f tbc cfficicncy iltltl comrol characteristics "f borizonnl axis windmills. Power available from the wind varies constantly from t"" little t" use t" cnuugb t" destroy a wind mac!line. Therefore, cfficictq sod c""tr"l design problems split into two categories. \\`ben windspeed is insufficient t" awe structural damage or power pbwt overlord, emphasis falls on efficient extraction of the iagest possible fraction of the pawer potentially available in the wind moving past the rotnr. For higher winds. design emphasis shifts tu proming the r,,t,,r, ruwer and power plant (gceerator, pump, etc.) frum orerload. As a xc""dary emph.xir, once rafcty is assured, the wind plant shoukl remover frown high winds the largest possible fracriw~ of f!~e maximum power it can safely handle.
C~msidering first tLe high-wind design qocrtions, tbcrc arc f"ur important meth"ds "scd singly or in combinati"" t" limit: a) r"t"r speed or power and b) disc drag, the axial wind farce that bends r"t"r blades and the t"\wr.
1. Blade feathcrirzg, the gwcming mechanism of Hydrowind 1, is t:le m"st commonly used method on large windmills. At full power "yeration, the flats of the t"t"r blades lie nearly in the plant of the rotor disc (Figwe 22). cutting across the wind and catching its full force. Part of the aerodynamic bl& force lies c3ngcntial to the r"t"t disc and generates

Wrquc~ wbilc tbc Other. gcnrr;dly Iargcr, disc drag CwIIpww~X c*crts its f":y dowmvind parallel f" fbc wtw aris. delivering 110power (Fip"rr 2;~). I:catbcring the blades back (Figure 21,) rcduc~s tbc zcr". dynamic pitch angle. which rcduccs blade lift, thus dccrcasitlg b"th r"wr drag and torque. If pitch is adjusted t" maintain c""sta"t r"t"r speed and turque with an increased wind tbruugb the r"t"r (Figure Zc), an increarcd fracti"" "f the total blade lift vcct"r becomes torque, so that tutal blade force becomes smaller. and bc,th blade bending forces and t"wer fumes actually decrease with increasing wind at constant power. R!ade feathering has the potential of minimizing blade and t"wcr stresses while maintaining full puwcr, thus ideally fitting b"th criteria for governing. However, if r"t"r blades feather in response f" rotation speed 3l""f, gllsts can El"W severe StresSCs.
Suppose that a r"t"r is operating slightly below gowrning speed when tbe wind iwreascs suddenly. Rotor drag immediately increases, but at first the extra torque goes into speeding up tbc r"t"r. Thus, r"t"r speed and shaft p"wcr lwels at first remain within safe limits as the r"t"r accelerates waking up the extra wind power, but the blades and mwer experience stresses ass"ciatcd with tlw greatly incressed total wind power.
There are tw" ways in which a blade pitch c"ntr"l gowrnor can prevent high gust stresses. One method is to use a conversion plant, alm"st always a synchronous operation clcctric generatar tied directly int" an AC purer grid, that forcer the r"t"r t" turn at rigidly c""sta"t rutati"" speed. With a synchronous generator, the frequency of the power grid enforces a lock-step c""sta"t speed r"tati"n on the generator, and tbc generator. in turn. cnfarces a proportional c"nsfant speed (depending on gear ratio) on the wwr. blnde pitch is then controlled by feedback t" maintain c"nstant generator power output, which am~u"ts to regulating pitch fur c"nstant rutnr torque. Arty change in torque CBUSESonly a brief transient chtmge in r"t"r speed after which the pnwer lines reimpose fixed r"tati~n yeed at a slightly different phase angle reMYe t" power line phase (Figure 3). Because the rotor is held so rigidly t" fixed speed, any wind change is reflected with little delay in a change of generator power. This makes possible rapid feathering response f" gusts.
Hydrowind I is ""t a synchranous "pcrati"" windmill, so Comtanf speed power guverning is inapplicable. The only alternative is t" make blade pitch responshe t" wind force on the blades "r, through sume mechanism, responsive t" measured wind "ear tbr r"t"r ccmcr The New Alchemy sailwing windmill responds t" f"rcc "n its sails through the elastic muunting "f the sails. so that they give W'L~t" high wind f"rces and spill air. But the Hydrowind I blades arc not constructed and pivoted t" tend t" feather passively under stress, so

,hc only oltcrmti\c \\""id be 3 romplex I,"t""3tiC

~omrol system tcrponsive to both *minion speed 3nd

citber biade stress or windspecd. Such il mccbaoism would be difficult to make reliable and would oat be

worth the expenrc. Confronted with this problem,

we opted to derarc the govcrt~or-limited mx4mum

rotatiotl speed of the mill to a valoc where a second

govcming mechanism would protect the blades am!

to~ver: aerodynamic blade stall.

2. .Ie+m,r,,ic

blade rrez// is the second of the

iour major governing methods over

a) speed or power and

1~)rotor dr3g

\Vbco lift-gcocratiegflow which fol!owsairfoil contows brcakr down at high snglcs of a&k, stall results (I'igorr 4). As airfoil angle of attack increases op to

the stall region, lift increases in proportion to angle of attack while airfoil dmg (not to be confused with

rotor drag, which results primarily from airfoil lift forces) is very low. As flow separation Sets in, airfoil

dmg increases sharply while lift &creases. Airfoil drag operates in such a direction as to reduce the

tangential ot torque-producing force of an airfoil.

Thus, sta!l of all or a part of the length of windmill Mates will reduce torque and rotor drag. If rotor

speed is held constaot~ then increasing wind speed wiil cause increasing blade angle of attack and ul-

timately blade stall. (Note Figures Zb and 2~. where blade angle of attack wuuld have increased going

from 2b to 2c if the blade had not pitcbed back to

maintain constzmt torque.) \Vhen blade stall progresses through a rotor, torque ma?; increase. remain about

the same, ot decrease, depending on blade shape.

If a rigid, fixed pitch rotor has any starting torque,

then sufficiently high winds xvill ultimately result in unsafe torque levels, hot care in design can assure

,,that "sofficicntly high winds" means winds unlikely

to occur during the life of the mill. Rotor drag in,creases are slowed by blade stall, but, as with torque,

sufficient winds will produce rotor drag exceeding

levels ctxotmtered before stall. Thus, design of swll-

governed rotors most conform with the built-in limits of the stall mechanism. That design within these

constraints can be successful is borne oat by two

well-rested stall-governed windmills, tbc 200 KW Danish mili at Gedrer. and the 150 KW mill being

tested at Cuttyhunk Island, Xlasracbusetts. For aerodynamic blade stall to operate, some other

mechanism most first limit rotor speed. Otherwise, the rotor blades would continue to speed up with in-

creasing wind. which would prevent increases in blade

~' Angie of attack. The usual mechanism for lnrge tiitidmills feeding into a power grid is the phase-lock

&wrty of generators connected to an AC grid, as

`\~a$ discussed in &tion to power governing by

: blade'pitch cqntr@. However, there are mnny types

of rind energy conversion plants that can be designed

,*,; ,A

,...I ~, (~",,,

,.A.,. .,/....,

..+ ril..l



._., _,h_.

3. i,.,>


to present a steeply increasing load on the rotor above some predetermined rotation speed. One such alterwtive can be accomplished through adjustments in the Gemini synchronous invcrter such as Hydrawind I uses. There are nomcroos possibilities for backtorque speed governing in stand-alone units. Some of these possibilities are the subject oi future New Alchemy research plans.
3. Rotoryaw control is the third governiog mcchanism, rihercby a rotor disc is turoed out of perpendicularitv to the wind to reduce the compooent of wind cm&g the rotor disc. (Figure 5 depicts a simple passive yaw cootrol mechanism.) There is a major limitation to the effectiveness of rotor yaw control: the gyroscopic inertia of a spinning rotor limits the speed with which it is possible or safe to reorient the rotot. Winds cm change direction faster than a rotor coo reorient to woid the perpendicular force, so there are times when blade forces are limited only by stall. Provided a rotor is strong enough to withstand maximum forces, yaw control cm govern rotor speed, though such a system will allow significant speed variation during the time lag between windspeed changes and adjustments in totor heading.
4. Aerodyrwnic spoilers reduce rotor airfoil cfficienties by disrupting flow and causing turbulence and blade drag. A very small tab automatically extending from a rotor blade upper surface and cutting straight

actoss tbc wind tlow cm suffice to slow a rotor. Automiltic spoilers cm gowm rotor torque or rotor speed but generally not rotor drag, so spoiling relics on blade stall to limit rotor drag.
Hwiog considered high-wind protection and goveming, we mnv come to the area of efficiency maximization in safe winds. Anv windmill will be most efficient at one particular ro&ion speed for soy given windspeed Uprimom rotor speed always varies almost exactly in proportion to windspeed. Define tipspeed ratio, t; by r = rotor tangential speed
windspced Then constant optimum propo&ionalit); between tipspeed and Godspeed implies that oprimtrm efficiency occurs at a fixed tipspeed ratio. Define coefficient of performance,Cp, by

c recovered power P = nr~ewer in the wind
where "power in tbc wind" means kinetic energy per unit time contained by the air moviog undismrbcd past ao area cqoal to rotor disc area. Empirically, C, is a function only of Z (see Figure 6 for examples 1 and IS virtually independent of Godspeed at fixed Z. Betz's ,, limit, 5926, is a widely accepted theoretical upper limit to achievable CP and .4 is considered a good practical Cp
Cp is zero at zero tipspeed ratio because a windmill that IS not turning (Z =O) cannot deliver power. Hawevcr, starting ti;tqze of a rotor may be critical for getting a load unstuck and moving in order to make blade power transfer possible. Positive displacement pumps (whether air, water or tefrigerimt) operating against 3 pressure


4 ;I


> ' z
"5 #~:f a'L



-..* n


d,d'OWi~+Om, md urn& pd`h G k

I ~~~~~i~ i~.~~-..~i~--~,;a

T;p<Fccd Rot,o

FIG. 6

Iwad ahwyr present spcci;d low rpccd rorquc rrqoirrmcots for a windmill, with tbc rcsolt tbst mills designed for pumping tend to look qoitc diffcrcnt from. say. electrici~-gencr=ting mills. High starting torque reqoires large blz~dc atcn with the blades pitched back typically at least 30". If tbc pitch of such blades is fired, tbcn torque and Cp rcacb xro at 2 =2 or less. \Vith pitch cootrol designed to flatten blade pitch as Z incrcascs, a high srurting rorqoc rotor can operate efficicntlv up to tipapeed ratios of 5 or 6. Thor. a pitch op&iring windmill is potentially versatile at delivering power to loads with differing torque rcquirements.
The blades of Hydrowind I bavc sufficient area to develop good starting torque when pit&cd back. Tbc blades always come to foil feather when the mill stops or when, for any rcasoo, hydraulic pressure drops to near zero at the pomp outlet. Thor. this mill coo stzrt high starting torque loads. And, as long as load torque increases with rotation speed. this mill will smoothly flatten its blade pitch with increasing windspeed, rotor speed and load torque. until it achieves the maximum
allowed pitch angle. (See Figure 1 for the hydraulic pitch mechanism.) At maximum pitch, Hydrowind I's blades perform efficiently over a broad range of tipspeed ratios (Figure 6). In fact, the proportions of Hydrowind I's blades are ideally suited for performance optimization through pitch control with "difficult" loads such as pumps and compressors. ,ilso, Hydrowiad would fit well into most electricity generating loads if blade pitch attglc remained fixed at the highest pitch setting.
Despite Hydrowind I's high potential versatility, there arc nvo cornmoo land types for which it is difficult or impossible to ncbicvc optimal pitch cootrol. First, there are problems with loads of virtually constant back torque at any 4tafr speed. Constant back torque implies constant hydraulic pressore, which, in turn, implies constant pitch (Figure 1). No adjustment to tbc pitch control mechanism can cause pitch to vary in an efficrcncy-producing way for such a load, which is characteristic of load cwves for most pumps and compressors. The second prablem arises with loads that produce very little low speed back torque. Such loads include centrifugal wt~ter pumps operating to overcome a significant static head and electrical generators operating into some loads; particularly rbroogb rectifiers to hatteries. With little back torque and back hydraulic pressure, the rotor blades rcmaio at full feather, so the rotor cannot possibly turn fast cvccpt in storm winds. Without turning fast, the system cannot devclop hydraulic pressure. So, the system is stuck at full feather. It is possible to rcmcdy either of these problems through hydmulic controls at tbc bottom of the mill. For cxamplc, an appropriately controlled servo-vahc could provide hydraulic back-pressure at


Fully dch"gJed, the hydraulic pitch control mechanism o" H?;dr"wind I has cost far more than the hydra"!ic power transmission. As we hwe soggestcd. rci-lstiicf:, ;inqde cooirols in the transmirsion itself
m mow toward eliminatiog the need for pitch con,"I. On this bzwis. we intend to abandon pitcb-con~wllingmcchamisms on foturc projects and cooeentrate II
it) developing an optimum rigid rotor system and h) de\;eloping load adaptations that aliow a rigid
rotor to operate efficiently md be speedgovcrncd by the load. Iydraulic power transmission should play a key role n some but probably not all ioad adapmtion prob:ms we can anticipate. While some special tasks might
E accomplished best without h\-draolics. we suspect hat hydraulic power transmission will be the best solution for an adaptable, economical general-purpose vindmill.
In the rigid rotor development area we are examning the advantages of tapered rotor blades. The broad Ilade tips of Hydrowind I contribute starting torque at !"I1feather, but at fixed maximum pitch, which would NJthe o.dy pitch for a rigid rotor, the large width of :he tips detracts from aerodynrmic efficiency at all peeds. Nor onI!> would mtrrow tips work more cf. ?icicntly (Figore 6), they would devclop lower peak rouices at m&mum lift angle simply becaosc of reiuced tip area. This implies significantly less bending stress in the blade roots as well as in the tower. The blade constrocti"" of Hy~drowind I is not adaptable to tapered plan form. The most likely candidate for tapered blade construction is fiberglass lay-up.
In the load adaptation area, we are examining several nonelectric tasks adaptable to wind tcchoology: wner pumping. air pumping, heat pumping (i. e., refrigerant pumping), refrigeration and heat geoeration by friction. Air and tefrigera"; pumping have the greatest need of hydraulic transmission approaches. \Pater pumping is probably best accent-

plirhcd I,), mcchaoical drive throogb it vcrtiwl shaft

to a rcntrifogal pomp drsigncd to porrro wiodmill

rpccd hy ;l rrecply rising torqw curvy. Friction hear-

iog oscs fluid friction sod a similur crntrifognl pomp

S~CC~ g0v~r110r ~0ncept. `rhc ppett?

tht h~k

great imerest to these particular thermal aml writer

pornping windmill tasks is storage. Pumped water

cao be stored for later use in an elevated or prer-

surizcd rescwoir. Hear cm be stored iu low gmde

form in water and in higher grade form in hent-of-

fusion of paraffin wax or sulfur. Cold can be stored

in hcat."f.?osio" of substancrs like cthylcne glycol,

so that food freezers can be kept well below freezing

during periods of no compressor power. Eiecrriciq

remains expensive to store. so storage of other forms

of energy holds grat interest.

WC crpcct "or next major windmill coostroction

project to result in a rigid tapered rotor to bc con-

trolled b!, B redundant combination of load control

of rotor speed, yaw control of rOtor speed, a me-

chanical brake to control rotor speed, and stall to

control stress. The design should be much simpler

and more economical than Hydrowind 1.

In all the designs we are considering. we aotici-

pate a fusion of high and intermediate technologies.

Components like hydraulic pomps and transistors are

available from industry and there seems no good

reason to forego their "se in domestic designs, but,

for an economical windmill that can be built and rc-

paired in shops of modcrate capital invcstmeot, the

high technology components must be standard pro-

duction "nits that fit, in modular fashion, into a

relatively simple stroctoral framework. Further,

the workings of the mill should be easy enough to

see and understand that a windmill "wner/"pcrat"r

with minimal tmining can keep the mill running. Some

of tbc mech~niams of Hydrowind I have required pre-

cision machined components not available off-thc4elf,

and this presents n formidable economic barrier to es-

tablishing small scale production. The machine has

failed to bc an encouragement to inooratorr who have

come co examine and learn from it. Instead, it has

taken o" that inscrotable complexity of technologies

like aotomotiw emission control equipment with the

unwritten message in the very stroctore (often reflect-

ed in the label): "Hands Off. Kefcr Service to Qualified Yr"fessi"nals Only." We have learned that there are simpler ways to make B wi"dmi!l. ways that cncourage a br"nder.based, more diverse and stsblc participatory technology.

Acrobat Capture Server 2.0 ilovepdf.com

Search Any Device: