An experimental set-up for energy generation using balanced kites

An experimental set-up for energy generation using balanced kites Mathias Clinckemaillie

Thesis voorgedragen tot het behalen van de graad van Master in de ingenieurswetenschappen: werktuigkunde Promotoren: M. Diehl D. Vandepitte Assessoren: E. Van den Bulck G. Pipeleers Begeleider: K. Geebelen

Academiejaar 2011 – 2012

© Copyright K.U.Leuven

Without written permission of the thesis supervisors and the author it is forbidden to reproduce or adapt in any form or by any means any part of this publication. Requests for obtaining the right to reproduce or utilize parts of this publication should be addressed to the Departement Werktuigkunde, Celestijnenlaan 300, B-3001 Heverlee. A written permission of the thesis supervisors is also required to use the methods, products, schematics and programs described in this work for industrial or commercial use, and for submitting this publication in scientific contests. Zonder voorafgaande schriftelijke toestemming van zowel de promotoren als de auteur is overnemen, kopiëren, gebruiken of realiseren van deze uitgave of gedeelten ervan verboden. Voor aanvragen tot of informatie i.v.m. het overnemen en/of gebruik en/of realisatie van gedeelten uit deze publicatie, wend u tot het Departement Werktuigkunde, Celestijnenlaan 300, B-3001 Heverlee. Voorafgaande schriftelijke toestemming van de promotoren is eveneens vereist voor het aanwenden van de in deze masterproef beschreven (originele) methoden, producten, schakelingen en programma’s voor industrieel of commercieel nut en voor de inzending van deze publicatie ter deelname aan wetenschappelijke prijzen of wedstrijden.

Abstract

’Een testopstelling voor energie-opwekking met behulp van gebalanceerde vliegers’ beschrijft het ontwerp van een nieuw systeem om windenergie te capteren met verankerde vliegtuigen. Dit concept maakt gebruik van vliegtuigen die als een vlieger, via een kabel bevestigd zijn aan de grond en zo een generator kunnen aandrijven. In kader van het Highwind-project is de voorbije jaren onderzoek gedaan naar het opstarten en landen van verankerde vliegtuigen om elektrische energie te produceren. De huidige opstelling biedt de mogelijkheid om één vliegtuig een rotationele start en landing te laten uitvoeren. Om de ontwikkelde algoritmes voor controle en positiebepaling echter ten volle te testen is een nieuwe opstelling vereist. Het eerste deel van deze masterproef beschrijft het onderzoek naar enkele nieuwe ideeën die deze nieuwe installatie kunnen verbeteren. Zo wordt het kantelen van de opstelling bekeken, alsook het mobiel maken van de installatie en het concept van gebalanceerde vliegers. Hierbij worden twee vliegtuigen via één kabel met de generator verbonden. Hierop volgt het conceptuele ontwerp van een opstelling die voldoet aan alle gestelde voorwaarden. Het tweede gedeelte omvat het gedetailleerde ontwerp van een opstelling voor één vliegtuig, die klaar is om uitgebouwd te worden tot een systeem voor gebalanceerde vliegers. Deze nieuwe carrousel is op een aanhangwagen geplaatst zodat deze te vervoeren is naar verschillende testlocaties. Er wordt aandacht besteed aan het ontwerp van de verschillende componenten en een structurele analyse van het geheel. Tot slot bespreekt deze masterproef het experimentele bewijs van een rotationele start en landing vanaf een vliegtuigsteun. Met behulp van de huidige opstelling is het bewijs geleverd dat een ongecontroleerd vliegtuig probleemloos kan opstijgen vanaf een steun, wat het gebruik van vliegtuigen met grotere spanwijdtes zal mogelijk maken.

i

Abstract

ii

’An experimental set-up for energy generation using balanced kites’ describes the design of a new set-up to capture wind energy using tethered airfoils. This concept uses airplanes connected to ground with a cable to drive a generator. The past years the Highwind-team has performed experiments to launch and land tethered airplanes to generate electricity. The current set-up makes it possible to make one airplane perform a rotational start and landing. To further test and validate the produced algorithms for control and pose-estimation a new test set-up is needed. The first part of this master thesis describes the investigation of some new ideas to improve the system. This includes tilting the whole structure, increasing the mobility of the platform and the concept of balanced kites. In this concept two kites are connected to the ground using one single cable. Next a conceptual design is proposed that complies with all the requirements. The second part consists of the detailed design of a set-up for one kite, which is ready to be expanded to a balanced kites configuration. This new carousel is mounted on a trailer to be able to transport it to different locations. The design of the different components is described, as well as a structural analysis for complete system. A last chapter delivers the proof-of-concept for a rotational start and landing from an airplane support. Using the current indoors set-up an uncontrolled airplane was launched from a so-called cradle. . This principle will enable airplanes with a larger span to be tested.

Artikel

Inleiding Dit artikel beschrijft het ontwerp van een nieuwe testopstelling voor energie-opwekking met vliegers. In 1980 publiceerde [Loy80] voor het eerst idee om een vliegtuig met een kabel aan een generator op de grond te verbinden. Door het vliegtuig als een vlieger te gebruiken, kan het de kabel afrollen van een trommel die de generator aandrijft. Deze systemen voor elektriciteitsproductie vergen een lagere initiële investeringskost dan een conventionele windmolen. Bovendien kunnen de verankerde vliegtuigen op grotere hoogte opereren, daar waar de grootste dichtheid aan windenergie te vinden is. De vliegende windenergie-systemen zijn momenteel volop in ontwikkeling, zowel binnen universiteiten als in commerciële instanties. Naast theoretische studies om optimale vliegtrajecten te bepalen, vereist de technologie ook nog onderzoek naar een snelle en nauwkeurige controle van de vliegtuigen. Een testopstelling is hierbij noodzakelijk om voldoende experimentele data te bekomen en de modellen te valideren. Onder de noemer van het Highwind -project is aan de KU Leuven de voorbije jaren een opstelling gebouwd om dit theoretisch onderzoek te ondersteunen. De zogenaamde carrousel laat toe om vliegtuigen met een spanwijdte van 1m te testen tijdens een rotationele start en landing (fig. 1). Dit houdt in dat het vliegtuig voldoende luchtsnelheid krijgt, door het aan het uiteinde van een arm rond te draaien. De huidige installatie kan enkel binnenshuis opereren en vormt zo een beperking voor de uitgevoerde tests.

Doelstellingen Om over te gaan tot het testen van grotere vliegtuigen, gecombineerd met de invloed van de wind is een groter, weerbestendig systeem nodig. Voor het ontwerp van deze opstelling zijn enkele concepten onderzocht die de prestaties kunnen opkrikken: de opstelling kantelbaar maken, zorgen voor voldoende mobiliteit en het gebruik van twee gebalanceerde vliegers. De besluiten van deze analyses leiden tot het gedetailleerde ontwerp van een nieuwe opstelling die de komende jaren als test-platform dienst zal doen. iii

Onderzoek naar nieuwe functies

(a)

iv

(b)

Figuur 1: Een carousel geeft het vliegtuig voldoende snelheid om op te stijgen.

Als de spanwijdte van de gebruikte vliegtuigen toeneemt, vergroot ook het risico met de carrousel bij het opstarten. Om deze reden gebruikt de nieuwe opstelling een vliegtuigsteun voor de start en landing. Een dergelijke steun is bevestigd aan het uiteinde van de arm en houdt het vliegtuig vast tot het voldoende snelheid heeft. Met behulp van een eenvoudig uitgewerkte vliegtuigsteun is het bewijs geleverd dat dit ook in werkelijkheid kan uitgevoerd worden.

Figuur 2: Een vliegtuigsteun bevestigd aan het einde van de arm.

Onderzoek naar nieuwe functies Kantelen van de opstelling Een rotationele start en landing van een verankerd vliegtuig buitenshuis ondervindt invloed van de wind. De snelheid van het vliegtuig is nu niet meer constant langsheen het traject. Figuur 3) toont de invloed van de wind op de resulterende luchtsnelheid van het vliegtuig. Dit fenomeen stelt een extra


v

eis aan de controle-algoritmes. Om het effect te verzwakken kan het rotatievlak van de carrousel gekanteld worden. Zoals te zien is aan de hand van de snelheidsvectoren daalt de invloed van de wind. Een cijfervoorbeeld met 36 km/h wind leert dat er een snelheidsverschil heerst tussen beide richtingen van 1,9 maal de tangentiële snelheid. Een kanteling van 20° zorgt voor een daling tot 1,4. Het kantelen van de opstelling heeft dus een voordeel voor het ontwikkelen van de

Figuur 3: Snelheidsvectoren bij de aanwezigheid van wind. Rood: wind, blauw: tangentiële snelheid, groen: resultante.

controle-algoritmes, maar introduceert ook heel wat complexiteit in het mechanisch ontwerp. Er zijn verschillende opties om deze functionaliteit uit te werken (fig. 4): de hele opstelling kan gekanteld worden, een kraan of hoogtewerker met kantelbare arm kan als basis dienen of de armen van de carrousel kunnen als de bladen van een helikopter worden aangestuurd. Vanwege de sterk dynamische belasting bestaat is er geen platform te vinden dat hiervoor geschikt is. Het uitwerken van deze optie zou dan ook teveel complexiteit introduceren zonder een gegarandeerde winst.

(a) Source: Hapert aanhang- (b) Source: JLG Lift equipwagens ment

(c)

Figuur 4: Mogelijkheden om het rotatievlak van de opstelling te kantelen.

Mobiliteit Om het roterend systeem te testen is een locatie nodig die afgesloten is voor voorbijgangers en voldoende ruimte biedt aan de vliegtuigen. Bovendien moet een dergelijke locatie voldoen aan de reglementering voor radiobestuurde vliegtuigen en moet vanaf 60m hoogte een aanvraag ingediend worden bij de autoriteiten bevoegd voor het luchtruim. Het nieuwe systeem moet dus makkelijk verplaatsbaar zijn van de ene locatie naar een andere om aan ten allen tijde aan deze voorwaarden te voldoen.


vi

Een aanhangwagen is de meest aangewezen optie vanuit praktische overwegingen. Ook hier moet aan een reeks voorwaarden voldaan zijn om de gemonteerde carrousel reglementair over de openbare weg te transporteren. Gebalanceerde vliegers Om de capaciteit van een energieproductiesysteem met vliegers te verhogen, kan één enkele kabel twee vliegtuigen met de generator verbinden (stap 3 in fig. 5). In deze optiek doet het Highwind-project onderzoek naar de zogenaamde gebalanceerde vliegers. De reden voor de efficiëntiewinst zit in de relatieve vermindering van de luchtweerstand op de kabel. Deze schaalt met het kwadraat van de luchtsnelheid en ten opzichte van twee aparte vliegers, beweegt de centrale kabel hier trager. Er is een theoretische verbetering mogelijk van 10MW naar 14MW voor twee vliegers1 [HD07].

Figuur 5: De verschillende stappen in de lancering van de twee gebalanceerde vliegers.

De installatie moet nu aangepast worden aan de specifieke noden van de gebalanceerde vliegers. Zo moet de verbinding tussen de centrale kabel en de kortere stukken bestaan uit zowel een lager om de krachten op te nemen, als uit een elektrisch sleepcontact om de signalen door te geven naar de vliegtuigen. De roterende opstelling op de grond vereist nu ook verschillende componenten om de start en landing in goede banen te leiden. Figuur 5 toont de verschillende stappen in de lancering van de twee vliegers. Bij het opstarten is elke vlieger verbonden met zijn uiteinde van de arm totdat de gewenste snelheid bereikt is. Dan kunnen beide vliegtuigen verder stijgen tot de korte kabels volledig ontrold zijn (±10m). De volgende stap vereist dat deze korte stukken kabel lostgelaten worden zodat de enige verbinding met de grond bestaat uit de centrale kabel. Gedurende de vlucht vergt deze kabel ook ondersteuning ter hoogte van de carrouselarm (groene pijl op figuur 5). Het ondersteunende element mag de bewegingen van de kabel echter niet verhinderen. 1

Deze waarden zijn bekomen v´ oo ´r toepassing van de Betz-limiet

Ontwerp van de testopstelling

vii

Voor elke van de benodigde onderdelen is een conceptueel ontwerp uitgetekend (fig. 6). Maar vanwege de grote hoeveelheid aan nieuwe stappen in deze procedure, is de nieuwe carrousel voorlopig voorzien op één vliegtuig. De verschillende elementen zullen stap voor stap worden ingevoerd om grote praktische problemen bij het testen te vermijden.

Figuur 6: Een overzicht van de nieuwe componenten op de carrousel.

Ontwerp van de testopstelling In het huidige systeem is een kleine 400W-lier voorzien om de kabellengte te variëren. Om de vliegtuigen verder van de arm te laten vliegen en de werkelijke energieopwekking te testen, moet de nieuwe installatie voorzien zijn van een grotere lier. Naast een nieuwe mobiele carrousel is daarom een ontwerp gemaakt voor een lier die 130m kabel kan stockeren in één laag en een maximumvermogen heeft van tien kilowatt. De gebruikte kabel bestaat uit dyneema 2 -vezels en enkele geleiders in het centrum. Deze geleiders voeren stroom en signalen tot bij het vliegtuig. Aangezien de opstelling aanvankelijk met slechts één vliegtuig zal werken, zal deze geen gelagerde verbinding bevatten. Om het twisten van de kabel te beperken bij de rotationele start en landing moet de lier nu meedraaien met de carrousel. De plaatsing van de lier op de opstelling heeft een invloed op het pad dat de kabel aflegt om het einde van de arm te bereiken en op de dynamisch belasting die de carrousel ondergaat. De trommel kan bovenop de arm geplaatst worden, maar ook concentrisch zijn met de as of onderaan de as hangen. Vanwege de betere stabiliteit en de mogelijkheden om over te schakelen naar een configuratie van gebalanceerde vliegers is gekozen om de lier onderaan de carrousel te bevestigen.

2

http://www.dyneema.com


viii

De opstelling (fig. 7) bestaat uit een statief met vier poten dat op de aanhangwagen geschroefd wordt. Het statief bestaat uit geëxtrudeerde vierkante aluminiumprofielen, met bout verbonden, zodat het geheel als een vakwerk fungeert. Onderaan het statief is voldoende ruimte voorzien om de lier op te hangen en te laten roteren. Het draaiende gedeelte bestaat uit een centrale as (φ100mm) met twee lagers, waarop de arm en de lier gemonteerd zijn. Deze lagers brengen het gewicht van de roterende onderdelen en de belastingen van het vliegtuig over naar het statief. Om de axiale krachten efficiënt op te vangen is het bovenste lager een kegellager. Onderaan vangt een kogellager de voornamelijk radiale krachten op.

Figuur 7: De ontworpen opstelling, zonder de aanhangwagen.

Een asynchrone motor en sturing zijn voorzien om de opstelling te laten roteren. Via een getande riem is deze met de as verbonden. Het benodigde vermogen en koppel van deze motor is bepaald aan de hand van de opgewekte weerstandskrachten van de vliegtuigen. Om de verliezen in de transmissie en de lagers in rekening te brengen is deze waarde met 20% verhoogd. Figuur 8 toont aan dat een motor met een vermogen van 1.1 kW vereist is. Om verschillende testlocaties te bereiken is het statief op een aanhangwagen geplaatst. De aanhangwagen is voorzien van vier uitschroefbare steunen verbonden aan twee stalen steunbalken, zodat de opstelling tijdens het draaien niet op de ophanging van de aanhangwagen rust. Er zijn geen gegevens beschikbaar over de stijfheid van de aanhangwagen, daarom wordt de carrousel bevestigd op de stalen balken waar wel gegevens voor beschikbaar zijn. Om de stabiliteit van de hele constructie te controleren is een eindige elementenmodel samengesteld. Het model benadert


ix

200 torque [Nm] power [x10 W] 150

100

50

0

0

10

20

30 40 n [rpm]

50

60

Figuur 8: Vereist koppel (-o- Nm) en vermogen (-x- 10W) van de carrouselmotor.

(a)

(b)

Figuur 9: Het gebruikte 1D-model en de uitwijking bij een statische belasting.

de werkelijkheid door de constructie te beschouwen als 1D-balken verbonden in knooppunten (fig. 9a). Elke balk krijgt hierbij de werkelijke dwarsdoorsnede en materiaaleigenschappen mee. De uitwijking en optredende spanningen zijn bepaald voor de maximale statische belasting. Aangezien er ook sterke dynamische lasten te verwachten zijn bij het draaien van de arm, zijn ook de eerste resonantiefrequenties van de structuur bepaald. De statische uitwijking blijft beperkt tot 1,8mm voor de top van de constructie. De dynamsiche analyse toont dat de eerst resonantiefrequentie van de carrousel op

Rotationele start en landing vanaf een vliegtuigsteun

x

de aanhangwagen rond vier hertz ligt. Dit is vier maal de maximale draaisnelheid. Indien bij het testen blijkt dat de opstelling toch teveel oscilleert in deze modevorm, kan men de stijfheid makkelijk verhogen door het aanbrengen van goedkope steunen aan de aanhangwagen. De lier (fig. 10) bestaat uit een synchrone servomotor via een riem verbonden aan een trommel (φ 200mm). Om het open afrollen van de kabel ordelijk te laten voorzien, is ook een mechanisme voorzien om elke winding naast te andere positioneren. Dit mechanisme bestaat uit een elektrisch aangedreven lineaire geleiding waarop twee katrollen gemonteerd zijn. De elektrisch aandrijving zorgt ervoor dat het systeem makkelijk uit te breiden is naar meerdere kabeldiameters.

(a)

(b)

Figuur 10: De lier, met en zonder frame.

Rotationele start en landing vanaf een vliegtuigsteun Met behulp van de huidige testopstelling is het experimentele bewijs geleverd dat een ongecontroleerd vliegtuig kan opstijgen van een vliegtuigsteun (fig. 11). Voor deze experimenten is een goedkope steun geassembleerd (fig. 11a), die zowel in lengte als in hoek verstelbaar is. In de Orocos-software omgeving van de opstelling zijn veiligheidsfuncties toegevoegd om schade aan het vliegtuig te vermijden. Deze veiligheden bestaan uit een limiet op de liersnelheid en op de maximale lengte van de kabel. In het geval van een bruuske beweging zou het vliegtuig niet langer een stabiele baan volgen en een te lange kabel resulteert onvermijdelijk in een botsing met de wand van kooi waarin deze staat opgesteld. Verschillende tests hebben aangetoond dat het vliegtuig enkel met controle van de rotatiesnelheid en de kabellengte terug geland kan worden op de steun. Zelfs na een botsing van de kabel met de steun bleef de uitwijking van het vliegtuig beperkt. De beste resultaten zijn behaald met de steun 30° gekanteld ten opzichte

Conclusie

xi

(a)

(b)

Figuur 11: Een rotationele start vanaf de vliegtuigsteun.

van de horizontale. Onder deze hoek bevindt het vliegtuig zich ver genoeg onder de arm om stabiel te zijn, terwijl de hoek toch groot genoeg is om aan hogere rotatiesnelheden te landen. Hoe groter deze rotatiesnelheid, hoe kleiner de invloed van de wind zal zijn.

Conclusie Vanuit de noodzaak van een nieuwe testopstelling voor buitenshuis gebruik is onderzoek gevoerd naar mogelijkheden om het systeem te verbeteren. Het kantelen van de gehele opstelling kan het hoofd bieden aan wisselende vliegsnelheden bij een rotationele start, maar is achterwege gelaten vanwege de mechanische complexiteit. Een conceptuele oplossing is geboden voor het lanceren van gebalanceerde vliegers. Deze opstelling zal via incrementele verbeteringen van een installatie voor één vliegtuig bereik worden. Deze nieuwe carrousel (fig. 7) is volledig ontworpen en is momenteel in productie. De vraag naar meer mobiliteit is aangepakt door de structuur op een aanhangwagen te monteren, die ter plaatse vastgezet wordt met uitdraaibare steunen. De resultaten van dit werk geven het team de mogelijkheid om uitgebreidere testen uit te voeren en het onderzoek naar vliegende windenergie-systemen voort te zetten. In de toekomst zal een verdere ontwikkeling van de algoritmes voor controle en pose-bepaling ertoe moeten leiden dat de vliegtuigen aan deze testopstelling succesvol energie kunnen produceren. Uit experimenten zal moeten blijken waar de installatie nog kan verbeterd worden, voordat overgegaan wordt naar een systeem met gebalanceerde vliegers. De optimalisering van het elektrisch gedeelte van het systeem is het onderwerp van een nieuwe masterproef in 2013. Net zoals het ontwerp van een geschikt vliegtuig voor dit systeem.

Preface

Before I started this master thesis, I had no idea about the possibilities and potential of airborne wind energy systems. During the course of the past year, I have not only been surprised by the number of innovative projects running across the world to launch these systems, but most of all by the enthusiasm and pioneering spirit of the people involved. First of all, I would like to thank my promotors, prof. Moritz Diehl and prof. Dirk Vandepitte, for their advice and feedback during this project. Their technical knowledge and enthusiasm always put the Highwind-project back on the right track. Special thanks to prof. Diehl for introducing me to different people in the industry. Most of all, I want to thank Kurt Geebelen for his daily support and answers to all of my questions. Together with Hammad Ahmad and Andrew Wagner, he lead the experimental research in this project to where it is today. Their constructive feedback and advice helped me a lot during this thesis. Also I would like to thank the whole Highwind-team for their warm welcome. My thanks also go to Corey Houle for sharing all his experience and his helpful suggestions. Through many emails and during his visit, he pointed out the critical parts in the design and shared his solution for the problem. The nice winch drum and pulleys he designed will absolutely help to turn the outdoors carousel into a success. Finally, I want to thank my family and friends for their support and encouragement, not only during this thesis but for the past five years. Thanks to Alexander Clarysse for his drive to finish all of our projects in style and the hours he spend correcting this text. Last but not least, thank you, Tom D’Hoore, Sanne Radoes and Lies Willaert, you know why! Mathias Clinckemaillie

xii

Contents

Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Artikel . . . . . . . . . . . . . . . . . . . . . . . . . . . . Inleiding . . . . . . . . . . . . . . . . . . . . . . . . . Doelstellingen . . . . . . . . . . . . . . . . . . . . . . Onderzoek naar nieuwe functies . . . . . . . . . . . . Ontwerp van de testopstelling . . . . . . . . . . . . . Rotationele start en landing vanaf een vliegtuigsteun Conclusie . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

i

. . . . . . .

iii iii iii iv vii x xi

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xii

List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xiv

List of Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xv

1

Introduction . . . . . . . . . . . . . . . . 1.1 Airborne wind energy . . . . . . . 1.2 Launch methods for AWE-systems 1.3 ERC Highwind . . . . . . . . . . . 1.4 Master thesis goals . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

1 1 3 5 6

2

Determination of design 2.1 Specifications . . 2.2 Tilted platform . 2.3 Test locations . . 2.4 Mobility . . . . .

specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

7 7 9 11 12

3

Balanced kites set-up . 3.1 Alternatives . . . Double cable 14, 3.2 Selected balanced 3.3 Conclusion . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Single cable 16, Winch location kites configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . .

14 14

. . . . . . . . . . . . . .

21 24

. . . . .

. . . . .

. . . . 18 . . . .

xiii

Contents 4

xiv

Design of the single kite set-up . . . . . . . . . . . . . . . . . . . . . . . . .

25

4.1

Carousel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Frame 25, Shaft 28, Bearings 29, Arm 30, Drive 31, Weatherproofing 32, Analysis 33

25

4.2

Trailer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Selected trailer 38, Installation of the carousel 38, Analysis 39

38

4.3

Winch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Drum 42, Level wind mechanism 43, Winch drive 44, Frame and integration 45

41

4.4

Electrical equipment . . . . . . . . . . . . . . . . . . . . . . . . . . .

46

4.5

Cost estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48

5

Experimental proof of rotating start and landing . . . . . . . . . . . . . . . .

50

6

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

52

6.1

Evaluation of the project goals . . . . . . . . . . . . . . . . . . . . .

52

6.2

Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

52

Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A1

A Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A2

A.1 Belgian road traffic regulations . . . . . . . . . . . . . . . . . . . . .

A2

A.2 Calculations of profile lengths and forces . . . . . . . . . . . . . . . .

A3

A.3 Trailers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A6

A.4 Cost estimates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A7

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

A8

List of Figures

1 2 3 4 5 6 7 8 9 10 11 1.1 1.2 1.3 1.4 1.5

Een carousel geeft het vliegtuig voldoende snelheid om op te stijgen. . . iv Een vliegtuigsteun bevestigd aan het einde van de arm. . . . . . . . . . iv Snelheidsvectoren bij de aanwezigheid van wind. Rood: wind, blauw: tangentiële snelheid, groen: resultante. . . . . . . . . . . . . . . . . . . . v Mogelijkheden om het rotatievlak van de opstelling te kantelen. . . . . . v De verschillende stappen in de lancering van de twee gebalanceerde vliegers. vi Een overzicht van de nieuwe componenten op de carrousel. . . . . . . . . vii De ontworpen opstelling, zonder de aanhangwagen. . . . . . . . . . . . . viii Vereist koppel (-o- Nm) en vermogen (-x- 10W) van de carrouselmotor. ix Het gebruikte 1D-model en de uitwijking bij een statische belasting. . . ix De lier, met en zonder frame. . . . . . . . . . . . . . . . . . . . . . . . . x Een rotationele start vanaf de vliegtuigsteun. . . . . . . . . . . . . . . . xi Airborne wind energy systems can profit from higher wind speeds at altitude. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Two different methods for power generation. . . . . . . . . . . . . . Different launch methods used in airborne wind power projects. . . . A rotating arm to give the airplane enough relative speed to climb. . The concept of balanced kites.[HD07] . . . . . . . . . . . . . . . . .

. . . . .

. . . . .

2 3 4 5 6

2.1

Speed vectors and graphical representation of the created speed difference and effect of tilting. Red: wind, blue: tangential speed, green: resultant speed vector. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Possibilities to implement tilting in the new carousel. . . . . . . . . . . . 2.3 The lower airspace near Leuven which is inside Brussels CTR(0 to 1500ft).Source: Belgocontrol AIP . . . . . . . . . . . . . . . . . . . . . . 2.4 Available trailers. Source: cfr. appendix A.2. . . . . . . . . . . . . . . .

12 13

3.1 3.2 3.3 3.4

14 15 16 17

Two options for the balanced kites configuration. . . . . . . . . . . . . . Conceptual drawing of a tether clamp. . . . . . . . . . . . . . . . . . . . A pulley assembly at the top of the arm to support the cable. . . . . . . A schematic representation of the components of an airborne swivel joint.

10 11

xv

List of Figures 3.5 3.6 3.7

Different steps in launching airplanes in a balanced kites configuration. . Angular velocities and resulting moment. . . . . . . . . . . . . . . . . . Different positions to mount the winch on the carousel.(Drum = gray, motor = orange) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.8 The winch drums mounted concentric with the carousel shaft but with an independent rotation speed. . . . . . . . . . . . . . . . . . . . . . . . 3.9 Step 1: landing the airborne connection. . . . . . . . . . . . . . . . . . . 3.10 Step 2: gripping the short tethers. . . . . . . . . . . . . . . . . . . . . . 3.11 Step 3: Bringing the attachment points to the end of the arm, reeling in the tethers and landing the kites. . . . . . . . . . . . . . . . . . . . . . . 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14 4.15 4.16 4.17 4.18 4.19 4.20 4.21 4.22 4.23 4.24

5.1 5.2

The designed carousel with a red line indicating the path followed by the tether. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Two alternative connections for the profiles. . . . . . . . . . . . . . . . . Detail of the adjustable supports and feet of the carousel. . . . . . . . . Shematic representation of the central shaft (rotated 90°). . . . . . . . . An exploded view showing the locknut, bearing, bus and aluminium structure element. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . An view of the assembled pieces and exploded view of the carousel arm. Required torque and power in function of carousel speed. . . . . . . . . 1D-beam models used for statical and dynamical analysis. . . . . . . . . Results of the statical analysis at maximal loading. . . . . . . . . . . . . Displacements of the arm divided by the applied force, in function of the load frequency (FRF). . . . . . . . . . . . . . . . . . . . . . . . . . . . . First displacement mode of the structure (8Hz). . . . . . . . . . . . . . . Results of the statical analysis without arm: stress (colors, MPa) and displacements (exaggerated). . . . . . . . . . . . . . . . . . . . . . . . . Frequency response function for the top of the carousel (no arm). . . . . First displacement mode of the structure without arm (28Hz). . . . . . A front and side view of the trailer with support beams (red). . . . . . . Displacements of the structure under statical loads only. . . . . . . . . . Frequency response function for the top of the carousel and trailer. . . . First deformation mode of the structure under dynamic loads (around 4Hz). Example of a cheap extra support, to be clamped if needed to the trailer. The designed winch with and without frame. . . . . . . . . . . . . . . . The level wind mechanism: a linear guide mechanism moves pulleys along the drum. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The winch mounted on the carousel shaft. . . . . . . . . . . . . . . . . . Overview of the electrical equipment (power, no signals), coloured components are optional. . . . . . . . . . . . . . . . . . . . . . . . . . . A through bore slip ring to connect the static controller to the rotating conductors in the tether. . . . . . . . . . . . . . . . . . . . . . . . . . . . The concept and experimental set-up for an airplane cradle. . . . . . . . The airplane taking off from the cradle seen from the carousel arm. . . .

xvi 17 19 20 21 22 23 23

26 27 27 28 30 31 32 34 34 35 35 36 37 37 38 39 40 41 41 42 44 46 47 48 50 51

List of Figures A.1 Definitions used in the geometrical relations. . . . . . . . . . . . . . . . A.2 Trailers within the 2x4m-range available in the vincinity of Leuven. . . . A.3 A list of the ordered components and their cost estimates. . . . . . . . .

xvii A3 A6 A7

List of Symbols

v ω larm S m Cl ρ L φ r E φdrum Fcent e ωwinch Ldrum Iyy Pcr Ealu Iprof ile lprof ile T Galu Jshaf t y k

speed [m/s] angular velocity [rad/s] length of the carousel arm [m] surface [m2 ] mass [kg] lift coefficient of an airfoil [-] density [kg/m3 ] lift force [N] tilt angle of the structure [°] tether length [m] gliding ratio [-] drum diameter [m] centrifugal force [N] eccentricity [m] angular velocity of the winch drum [rad/s] angular momentum of the winch drum [kg · m2 /s] moment of inertia [kg · m2 ] critical loading for buckling [N] Young’s modulus of aluminum [Pa] bending moment of inertia [m4 ] length of a component [m] torque [Nm] shear modulus of aluminium [Pa] torsion constant of the shaft [m4 ] deflection of a beam under a bending load [m] spring constant [N/m]

xviii

Chapter

1 Introduction

Renewable energy is a fast growing branch of the global energy production. In order to find economic feasible alternatives and additional production systems for our current energy sources, research must be conducted to find innovative energy harvesting methods. Next to solar power, wind energy provides us with an enormous potential. Windmills convert the momentum of moving air near the earth’s surface into electricity, but also at higher altitudes consistent, high velocity winds can be found. In 1980 Loyd was the first to publish the idea of harvesting this power using tethered airplanes [Loy80].

1.1 Airborne wind energy Mankind has been using wind to drive mills and sail across oceans for centuries. In the last decades windmills have been optimised to generate electricity. Since the aerodynamic forces on the turbine blades are proportional to their relative speed squared, the power generated by the wind mill is proportional to the wind speed cubed. An increase in wind speed can be observed with higher altitudes, because friction with the surface and obstacles slow wind down (fig. 1.1a). Because of this wind gradient modern windmill towers are higher than ever before. In Belgium, the tallest operational windmill can be found in Estinnes. The blade tips of this Enercon E-126 1 reach just below 200m. Higher constructions are unlikely to be built due to the bending moments on the tower and the increase of initial investment cost. But from collected data [AC08] (fig. 1.1b), it can be seen that the available power density still increases, never reaching a maximum below 500m. Due to the dependency to relative wind speeds, the outer parts of the rotating blades contribute the most to the power delivered on the shaft. This means that a big part of the blade and the whole tower need to be erected merely to support the fast moving blade tips of the turbine. 1

http://www.enercon.de/de-de/66.htm

1

1.1. Airborne wind energy

2

3,500 3,000

altitude [m]

2,500 2,000 1,500 1,000 500 0

0

10 20 30 wind speed [km/h]

(a) Wind speed in function of altitude, Brus- (b) Altitude of maximum wind power densels 30-05-2012. (Source: Belgocontrol) sity(95th percentile of the annual average).

Figure 1.1: Airborne wind energy systems can profit from higher wind speeds at altitude.

The disadvantages of a conventional windmill can be cancelled out by using an airfoil connected to the ground by a tether. By doing so, the number of opportunities to harvest the enormous potential of renewable energy increase. These energy production systems are referred to as airborne wind energy systems (AWE). The kites can fly higher because no tower needs to be built. With appropriate control algorithms they can follow optimal trajectories. In this way the airborne wind energy system acts as a wind turbine of which only the efficient outer part is used. To gain electricity from the movement of the kites, there are main strategies. A first option is to use the flight wind passing the kite to drive a turbine, mounted on the kite (fig. 1.2a). The on-board generated power is then transmitted through the tether and fed to the grid. The second method(fig. 1.2b) is to place a single generator on the ground. This generator/electric motor is connected to a drum on which the cable is wound. The lift force of the airfoil is used to exert a force on the cable, reeling it out and driving the generator. Of course the cable length is finite and should be pulled in again, in this stage the generator is used as a

1.2. Launch methods for AWE-systems

3

(a) On-board generation: due to the flight speed a turbine is driving the airborne generator, there is no need to vary the tether length.

(b) A pumping cycle: reeling out the cable using maximal lift forces(green) and pulling it back in at minimum drag (red) results in a net energy gain.

Figure 1.2: Two different methods for power generation.

motor, delivering energy to the kite. By controlling the kite, this can be done using less energy than produced in the first phase. The result of this pumping cycle is a net gain in electricity.

1.2 Launch methods for AWE-systems Starting an airborne wind power system requires the launch of the airfoil. Because of the wind gradient, speeds near the surface are often too low to generate sufficient lift on the wing to make it climb to its operating altitude. All over the world commercial and scientific projects are running to improve this kind of energy production system. The different approaches to the problem result in different launch methods for the kites or tethered airplanes. Skysails, a company fabricating kites to tow big ocean ships, uses a crane to bring the kite up to a certain height to maintain clearance from the ship. This approach is not adequate to start up a static system on land because of several reasons. Sufficient relative wind speed can more easily be generated using the vessels forward speed and the wind gradient is smaller above water. Moreover the kites can only be controlled in a uniform airflow, while wind above land is known to be more turbulent. In the situation where on-board generators are used (fig.

1.2a), the teth-

1.2. Launch methods for AWE-systems

4

ered aircraft is equipped with a turbine. To deliver the needed energy for take-off, the generator can be used as an electric motor. The turbine now acts a propeller and delivers thrust for a conventional or even vertical take-off. Figure 1.3b and 1.3c show the implementation of this principle by Makani and Joby Energy 2 .

(a) A winch launch inspired by gliders, used by Ampyx Power.

(b) Vertical launch using a catapult and propellers.

(c) Vertical take-off using propellers.

Figure 1.3: Different launch methods used in airborne wind power projects. If a pumping cycle is used to generate the electricity, the electric winch is an important component in the design. As this powerful winch is available, the kite or airplane can be launched as is done with gliders (fig 1.3a). First the cable is reeled out and connected to the airplane standing at the other end of the field. When the cable is now winched in, the airplane gains speed and can climb following a parabolic shaped flight path. This approach requires a rather large field for the launch, but no extra components. To reduce the needed launch area, the linear movement of this manoeuvre can be turned into a rotation. If an airfoil is fixed at the end of a rotating arm (fig. 1.4), it experiences a relative airspeed v = ω · larm . Once the stall speed 3 of the kite is reached it can take-off from the arm. The rotating platform or carousel keeps delivering energy to the airplane until it is sufficiently high to be powered by the wind. This climbing manoeuvre necessitates a tether length control, usually a winch. It is referred to as reversed pumping, because it consists of a reeling in phase at high tether forces and a phase in which the tether is reeled out again at lower forces. This results in an energy flow from the winch to the airplane. [GGG+ 12] 2 3

Joby Energy and Makani recently merged and are now working under the name of Makani. The stall speed of an airfoil is the minimal (steady-state) speed needed to maintain an attached airflow and produce lift. Below this speed, the boundary layer separates from the profile and no lift is generated. Practically this corresponds to the lowest speed at which an airplane can stay airborne.

1.3. ERC Highwind

5

Figure 1.4: A rotating arm to give the airplane enough relative speed to climb.

1.3 ERC Highwind At KU Leuven research is done to improve the performance of power generating kites. The Optimization in Engineering Center (OPTEC) tries to optimize the performance of the kites by choosing the flight path producing maximal energy and improve the control during the different stages of the flight. The control algorithms for these systems must be fast and able to cope with sudden wind speed changes. Since 2011 the project has been supported by the European Research Council (ERC ) and is called Highwind. The team decided to focus the research on the rotational start and landing method with the aim of a 24-hours autonomous flight at the end of the project. An automated carousel should be able to execute any necessary landings and launches without human interaction. This master thesis is a part of the Highwind-project and continues on work of the past years. A first master thesis was written in 2008 [ES08] concerning methods to improve tethered airplane control and the concept of a carousel. To gather data about kite behaviour and for the purpose of validation of the models and controllers, a test set-up is indispensable. In 2010 the first carousel was built and models of the airplane behaviour and pose determination described in a new thesis [GGD+ 10]. This indoors test set-up is still in use to collect test data. After the construction of a small winch [CE11] the opportunity arises to get experience with rotational starts and landing. To increase the performance of the system, more than one kite can be integrated in a single system. For the Highwind-project the option of so called balanced kites is chosen. Figure 1.5 shows the principle: a single line leads to a connection of two shorter tethers to which a kite is attached. The kites are controlled in such

1.4. Master thesis goals

6

Figure 1.5: The concept of balanced kites.[HD07]

a way that the central line is always moving at relative low speeds. Due to the quadratic relationship between drag and airspeed, the total losses in the system are significantly lowered. Two kites, separately capable of producing 5 MW, generate up to 14MW when joined in a balanced kite set-up (before applying correction with Betz-factor)[HD07].

1.4 Master thesis goals The goal of this master thesis is to continue the development of a test set-up. The design should use the experiences gained with the current indoor carousel. The first part of this thesis consists of the determination of the best concepts and configuration to use. In a first chapter,the project specifications are described as well as the qualitative investigation of some new ideas. The next chapter discusses which concept to chose to combine the rotating start and landing with the balanced kites concept. The results of this part set the goals for the remaining years of the Highwind-project. The second part incorporates the actual design and construction of a single kite carousel, ready to be changed in a balanced kites launching platform. As wind is of the essence in this research, the new system should be able to perform on an outdoors test location. This means mobility and weatherproofing are determining factors in this design. A secondary goal is to deliver the proof-of-concept for the rotational start and landing from an airplane support, often called a cradle. The results of this experiment can be found in the fifth chapter of this thesis.

Chapter

2 Determination of design specifications

This chapter covers the determination of the project goals. To do this a list of specifications is put together and some interesting ideas are investigated to determine their feasibility.

2.1 Specifications After consulting the team working on the modelling and optimization the needs of a new carousel were lined out. Next to the obligatory improvements, such as an enlarged clearance for the kites or the mobility of the platform, the list contains options worth investigating. Functions Automated rotational start/landing of two tetherd airplanes(2m wingspan) in balanced kite configuration. Control of tether length and possibility to generate energy using a pumping cycle. Sensors and communication to control the airplanes and energy production (not part of this masters thesis). Tiltable plane of rotation to increase controllability (cfr. 2.2).

General Outdoor use: protect from rain, dust, wind and if possible lightning Mobility:

– transport from/to and de/installation at location: maximum once a day, limited number of people(±3) – stand alone: own energy production for start-up and testing Robust design: avoid delays due to mechanical failure (off the shelve components and reserve) Stability of the construction Minimise energy losses, but not as a main goal Safety: protection for operators at all time

7

2.1. Specifications

8

– cage/net – emergency stop Dimensions and loads Airplanes: wingspan: 1-2m The lift force from the airfoil and the exerted centrifugal force during rotation can be estimated as follows:

vmax = 45 m/s S = 0.5 m2 mmax = 5 kg

Clmax = 1.5 ρ = 1.22 kg/m2 1 2 · ρ · S · vmax · Clmax 2 = 926N Maximum lift force, 1 kite

Lmax =

Fcent = ω2 · r mmax = (2π)2 · 5 = 20g

(2.1)

Fcent = 990N max. centrifugal force A good upper limit for the airplane forces on the structure is 1000N. This already incorporates a certain safety, because they will not be flown at maximum speed and maximum lift coefficient. Also the acceleration of 20g due to the rotation is above the structural limit for the planes. Structure – rotation speed: 60rpm – arm length: 2x2 = 4m – normal road transportation: maximum heigth 4m, maximum width 2.50m Winch specifications

Electrical winch 6 m/s real-out speed (generator), max force by kites = 1800N (10kW) 8 m/s real-in speed at ±500N (motor) 100-150m of dyneema+copper cable (3-5mm diameter) on the drum Only one layer of cable on the drum during the power generating phases: in case of multiple layers the cable can be pulled between separate windings and damage the system. Cable bending radius larger than 5cm: to avoid a large decrease of the cable life or damage to the copper conductors the bending radius of the cable must always remain above a certain value. Manufacturers for steel cables provide guidelines for this radius: the pulley diameter should at least be 30 times the cable diameter (30 x 3mm = 9cm) [Sav06]. The dyneema tether is less stiff than a steel cable, but to avoid wear it is wise to maintain the same minimal bending radius.

2.2. Tilted platform

9

Dump the produced energy: the main goal of this project is to test control algorithms and optimization models, not to produce energy. The cable is wound uniformly on the drum using a level wind mechanism: make this adaptable to new cables. Volume: 1x1x0.8m (order of magnitude) Budget: maximum e20 000 Whole winch rotating with carousel in one kite configuration Line coming out upwards, ready to go through main shaft in case of one kite

Power supply Provide electrical power for the winch, carousel motor and other electronics at their respective voltages. Prevent damage due to voltage surges or too high currents For the first tests in or near the lab the set-up can rely on a connection to the grid. But later on the system should be able to run autonomously at different locations.

Budget There is no exact budget limit for the whole test set-up but every expense must be well-motivated. As an order of magnitude e50 000 is mentioned as a maximum for the whole set-up.

2.2 Tilted platform One of the proposed functionalities of the new carousel is to make the plane of rotation tiltable. Driving the tethered airplanes by means of the rotating arm in a windy environment creates a difference in relative speed on the upwind and downwind side (fig. 2.1). The proposition of a tiltable structure originates in the reasoning that adding a tilt angle φ diminishes this speed difference. A numerical example:

n rad/s 60 = larm · ω = 12, 6m/s

n = 60rpm vtan

ω =2·π·

(2.2)

vwind = 36km/h = 10m/s Horizontal: φ = 0° ∆ v = 26.6-2.6 = 24 m/s ∆v vtan = 190%

Tilted: φ = 20° ∆v = 22.3-4.7 = 17.6 m/s ∆v vtan = 140%

The example shows that the magnitude of the speed difference is lowered. But this improvement comes with other disadvantages. As can be seen from the graphical representation, not only the magnitude of the resulting vector changes but also its angle with the flight path. Therefore

2.2. Tilted platform

10

Figure 2.1: Speed vectors and graphical representation of the created speed difference and effect of tilting. Red: wind, blue: tangential speed, green: resultant speed vector.

adequate control of the airplanes angle of attack remains of the essence. The carousel structure is limited in height because of road transport regulations (cfr. appendix A.1). If the airplanes fly a tilted circular trajectory the bottom part of this path approaches the ground with increasing tilt angle. The risk of a collision increases. In reality the wind speed and direction change during the course of the day. A continuous measuring of and adjusting to the wind is necessary during operation to maintain the benefits of the tilt. Mechanically this functionality increases the complexity and introduces new dynamic loads due to the gyroscopic effect of the rotating carousel. There are different options to implement this principle. A first is to tilt the whole structure or the platform on which it is mounted. Trailers with a tiltable platform are commercially available to simplify the load process(fig. 2.2a). As they are not designed for dynamic loads, they lack stiffness and they cannot be turned if there is a change in wind direction. One could also provide each support of the carousel with a length adjustment system. By changing the height of every support differently, the structure would obtain an inclination. To limit the production cost and time it is wise to look for existing equipment that can be modified to meet the requirements. Examples of machines incorporating a swinging arm are cherry pickers (fig. 2.2b), small cranes and fairground attractions. There are a lot of possibilities on the second hand market. The problem with this kind of solution remains the dynamic loading of a structure designed for static operations. By simply looking at some working cherry pickers, it can easily be seen that the arm tends to swing at frequencies around a few hertz, which will pose a problem for a carousel rotating at one hertz. Fairground attractions combining a rotating load at the end of an arm are too big and costly for

2.3. Test locations

11

this project. A last option is to keep the structure and main shaft stationary, only angling the upper part. This can be done by putting a cardan between the shaft and the upper part or by controlling the arms of the carousel like the blades of a helicopter (fig. 2.2c). Both require the design of a complex coupling and control system carrying all the loads applied to the system. The cost of the system increases with a decreasing reliability. Briefly, implementing the tilt function in the new set-up comes with relative big disadvantages with respect to the advantages. This functionality will not be included in the outdoors set-up. The fluctuations in wind speed are part of the requirements for the controllers and state estimating algorithms.

(a) Source: Hapert aanhang- (b) Source: JLG Lift equipwagens ment

(c)

Figure 2.2: Possibilities to implement tilting in the new carousel.

2.3 Test locations Another new item in the specifications list is the increased mobility of the platform. The current test set-up is located inside a darkened safety cage in the lab. To be able to perform tests in a real-life environment the whole system needs to be transported to locations where there is enough wind and space. The ideal test location is hard to find and will most probably vary from time to time. A large empty area is needed in order not to harm people in case of a tether failure. By increasing the wingspan and tether length the test set-up comes in the range of certain regulations. In Belgium radio controlled airplanes are only allowed to fly on certain fields. When reaching an altitude of 60m the Belgian federal government imposes a duty to report this obstacle to Belgocontrol, the national air traffic control authority [evD06]. Locations near highways, rivers or airports are most probably to get rejected. The city of Leuven is located inside the control area (fig. 2.3) of the

2.4. Mobility

12

national airport in Zaventem. When exceeding the 60m limit, it is possible that testing will need to take place further away from the lab. For future reference, a study of the Brussels university VUB, is mentioned which discusses suitable sites for wind energy harvesting in Flanders [CDL00].

Figure 2.3: The lower airspace near Leuven which is inside Brussels CTR(0 to 1500ft).Source: Belgocontrol AIP

2.4 Mobility To transport the set-up to different test locations a mean of transportation is needed. To avoid long times to mount or dismount the set-up at the location, the specifications state to have the carousel build on a mobile platform. A rented trailer would require a complete disassembly of the carousel because an unfixed structure can not be transported safely over public roads. This design must satisfy the regulations for road traffic. The maximum width of a vehicle is 2.50m, the maximum height 4m. Other specific rules concerning length and turning radii can be found on-line [KB887] (cfr appendix A.1). Depending on the maximum allowable weight of the movable platform the driver needs a certain driving license. The cost and practical problems of obtaining such a license must also be taken in consideration. Figure 2.4 shows the possible range of trailers. Only a few small trailers can be towed by a person with a normal driving license. Once the dimensions increase a new license (B+E) is needed. It is possible for the team to work with such a trailer because some members have this permit. If future tests require an additional person with this license, the cost of e800 is acceptable. Following these trailers in size are the small trucks. Their disadvantage is the cabin which reduces the clearance between kite and structure. The selected trailer must not only be strong and large enough but also able to resist the dynamic loading of the rotating carousel. If any trailer would be left on its wheels, the suspension will allow big deflections of the system. Therefore a

2.4. Mobility

13

support system must be found which can lift the whole structure from the trailer suspension. After consulting local dealers, it is clear that no such off-the-shelf supporting systems are available within the range of trailers, towable by a normal car. The chassis of a trailer is certified and any modification, even only drilling a hole, could lead to a rejection during the obligatory inspection. Therefore the final choice fell on a car-transport of 2x4m fitted with steel support beams (cfr. 4.2).

Figure 2.4: Available trailers. Source: cfr. appendix A.2.

Chapter

3 Balanced kites set-up

Based on the specifications, different concept alternatives are described and the final selection is made. The main functionality of the designed carousel is to launch, control and land two tethered airplanes. To do this, not only the connection between the long and shorter tethers is to be determined, but also the location of the winch and the procedure to land the planes on their cradle.

3.1 Alternatives In this section some possible conceptual designs are proposed together with their advantages and disadvantages. A first distinction between these alternatives is the use of one or two cables.

(a) Two separate cables joined at a cer- (b) A single central cable connected to tain point. the shorter tethers.

Figure 3.1: Two options for the balanced kites configuration.

3.1.1 Double cable To simplify the connection of the two airplanes, each kite can have a separate cable which are then connected to each other at the appropriate length by a rod (fig. 3.1a). 14

3.1. Alternatives

15

This is of course an approximation of the theory of the balanced kites and will have a larger cable drag than a true single cable. The use of two separate cables implies also that there are two cable drums on the winch.The implementation of this principle consists of a simple rod with two clamping mechanisms at the end. Clamping mechanism The open clamps allow the cables to be reeled out while the kites climb further away from the carousel. At the appropriate length the clamps should be closed and grip the cables. By doing so the rod climbs together with the tethers and serves as the airborne connection. To avoid to have an actuator on the bar, the clamping force would be delivered by a strong spring. To open the clamp, actuators mounted on the carousel arm are activated. Comparable cable grips can be found on chairlifts where they secure the gondolas to the cable. The tether used in the Highwind-project consists of woven dyneema fibres for strength and some copper conductors in the center to send power and control signals to the airplanes. For this reason, the pressure on the cable should be limited and distributed over a certain length. To increase the grip on the cable the surface of the clamp can be given waving contour (fig. 3.2). This increases the contact length and the pressure on the cable (∝ 1/radius). A last important factor to keep in mind is the bending radius of the cable. Due to fragility of the conducturs inside the tether must always have a bending radius of at least five centimetres.

Figure 3.2: Conceptual drawing of a tether clamp.

Evaluation An advantage of this configuration is that the spring mechanism itself, as well as adding an extra drum is quite simple. But the disadvantages are numerous. Catching the bar with the clamps attached and opening the mechanism with actuators on the carousel will be hard to control. The winches in the double cable approach always need to rotate with the carousel to avoid cable twist (cfr. 3.1.2). The setup will always be tested first with one kite, when upgrading to two kites the same cable would be used. This leads to a cable weigth and cost that are twice that of a single cable. Also the bar and the gripping parts must be lifted by the airplanes, leading to extra efficiency losses.

3.1. Alternatives

16

3.1.2 Single cable A first challenge to tackle in a single cable configuration is the point where the cable leaves the carousel (green arrow in fig. 3.5). The tether is continuously reeled in and out and should be supported and redirected to the winch at this point. This support must consist of a pulley equipped with bearings to keep the friction losses and cable wear minimal. To allow the cable to angle in all directions with respect to the carousel shaft, these pulleys must be able to rotate around the shaft direction (fig. 3.3). In the concept the pulley assembly is fitted with some rollers to smoothen the contact of cable. As soon as the angle of the tether changes and it touches these rollers, the whole assembly will turn to realign the pulleys and the cable.

Figure 3.3: A pulley assembly at the top of the arm to support the cable.

Twist When the airplanes start flying circular trajectories, the cable will twist. Due to wear of the fibres and the conducters this effect must be minimised. One option is to use only flight paths shaped like an eight, in which turns in both directions are made and the resulting twist is zero. Limiting the possible flight trajectories in the mechanical design is no option because one of the main goals of the Highwind-project is to optimize their shape. Two solutions are possible: make the whole winch turn simultaneously with the kites: the airborne connection between the cables is a simple rigid connection. connect the main line to the small tethers using a swivelling joint and slip ring: the winch is static and placed under the carousel, power and signals go to the airplanes through the slip ring (fig. 3.4).

3.1. Alternatives

17

Figure 3.4: A schematic representation of the components of an airborne swivel joint.

Start and landing The rotating start requires to drive the airplanes by connecting the cable to the ending point of the arm. When this is not the case, the tethered planes will not be accelerated to the required speed to lift off (step 1 in fig 3.5). To fly the kites in a balanced configuration, they need to have a seperate, short tether each of at least ten times their wingspan. This allows them to fly the optimal trajectories without the risk of a collision. Specific for this carousel each airplane needs to be separated at least by ten meters of cable from the connection point to the single cable (step 3 in fig. 3.5). To go from step 2 to 3 and the other way round for the landing, requires a mechanism that can release and reattach the shorter tethers to the end of the arm.

Figure 3.5: Different steps in launching airplanes in a balanced kites configuration.

The kites can only be driven by the carousel as long as their separate cables remain attached to the arm. This means that the maximum height realisable with

3.1. Alternatives

18

the carousel is not only depending on the arm to tether length ratio (larm /r) and glide ratio[GGD+ 10], but also by the length of these cables. Once these are fully reeled out, the only way to climb further is by applying reversed pumping (cfr. 1.2). During the landing phase the first action is to reel in the cable until the central connection lands on the carousel. At that point the cable can not be pulled in any further, but neither can the carousel drive the planes by rotating the arms. This is a critical phase where either extremely adequate control or an external energy source will be required. The performance of the airplane controllers and the speed of the mechanism on the carousel to fix the cables back to the end of the arm must first be determined. The time between the landing of the central connection and the kites reaching the level of the arm can be calculated starting from an estimated gliding ratio. The kites will be optimised to have a large gliding ratio E as it has a quadratic influence on the maximum power they can generate for a given wind [Loy80]. The maximum value of E is around 40, but to incorporate imperfections and the tether drag, the calculations are done using a gliding ratio of 10. Emax = 10 VEmax = 40 km/h VEmax /3.6 Vv = E 10m t= = 9s Vv

Maximum gliding ratio of the kite: lift/drag speed at wich this E is reached vertical speed of the descending kite

(3.1)

time to react

The developed system must bring the cable attachment points to the end of the arm and drive the kites within 9 seconds. The option of an extra energy source remains open as well: if the technique of on-board power generation is to be investigated using the outdoors set-up, small propellers will be fitted on the airplanes. Otherwise cheaper solutions such as compressed air are under consideration. Evaluation The single cable configuration incorporates all advantages of the balanced kites theory. Another benefit is the prospect of a single winch with only one cable. The problem of supporting the cable under any angle with the carousel can efficiently be solved (fig. 3.3). Disadvantages in this case are the problems with reeling in and out the shorter tethers during start and especially landing.

3.1.3 Winch location The last major decision in determining the configuration of the new carousel is to appoint a location for the winch. The electric winch is the heart of the system since it both provides power to the airborne kites and is used to generate electricity. For the placement of this winch we need to consider the ease with which the cables can be guided to the end of the arm and the consequences for the stability of the whole system.

3.1. Alternatives

19

The specifications (cfr. 2.1) state a cable length of 100 to 150m and a maximum power of 10 kilowatts. Based on the density of aluminium and a drum diameter of φdrum 200mm the weight of the drum can be estimated around 45 kg. A 10kW-servomotor1 weighs around 45kg. This makes a total of around 100 kg for the complete winch assembly. The influence of the position of this mass must be considered, especially when the winch is rotating with the carousel. The centrifugal force of an eccentric rotating object is proportional with the mass: Fcent = m · ω 2 · e with e the distance to the center of rotation. The displacement response of the structure depends on its rigidity between the input of this force and the considered point, but is also dependant on the excitation frequency. Rotating the working winch with the carousel creates also a bending moment. When reeling the tether in and out, the drum of 45 kg rotates at ωwinch . The carousel rotation means changing its angular momentum Ldrum . The change in this angular momentum requires a new moment M (fig. 3.6). −−−→ dLdrum dt −−−→ ! −−−→ dL0drum − = + → ω × L0drum dt −−−→ − ωdrum =constant = Idrum · α + → ω × L0drum     0 0 =  0  ×  Iyy · ω  ω 0 − = −ω · ωdrum · Iyy · → ex = 2π · 80 · 0.5 · 45 · 0.102 + 0.092 = 205N m

M=

Figure 3.6: Angular velocities and resulting moment.

1

Lenze MSC190P series

(3.2)

3.1. Alternatives

20

(a)

(b)

(c)

Figure 3.7: Different positions to mount the winch on the carousel.(Drum = gray, motor = orange)

Possible locations Figure 3.7a shows the set-up in the current carousel. A small 400W-winch is mounted on top and enables the operator to vary the tether length during indoors experiments. For the outdoors set-up, the winch will have a mass that is ten times higher. The detrimental influence on the stability, when the winches are mounted with the same error in balancing, would also increase with a factor ten. Keeping this big mass on top of the carousel has two disadvantages: it is as far away from the structure’s supports as possible. The consequence is that the displacements due to the dynamic loading of a little eccentricity will create large displacements. The second problem is that the single cable configuration becomes impossible to implement because there is no place left to mount the necessary components on top of the carousel. By moving the winch down the carousel (fig. 3.7b) the moment delivered by the centrifugal force with respect to its supports on the ground is lowered. It can be mounted on a rotating platform underneath the structure or hang from the main shaft of the carousel. This option also leaves space open on top of the carousel to mount the pulley assembly which can support the main cable in a single cable system. The winch in this location can be mounted with the drum standing upright or lying down. In the upright position the carousel must be made higher. Another option worth mentioning is to mount the drum of the winch concentric with the carousel shaft (fig. 3.7c). The connection between the shaft and the drum consists of a bearing, providing an independence in rotating speeds for the drum and carousel. Both the winch and the carousel are driven by their own motor (fig. 3.8). The tether length can be controlled by varying the difference in rotation speed between drum and carousel. When they rotate at the same velocity, the tether remains constant, when the winch spins faster then the carousel the cable is reeled out and vice versa. The advantage of this configuration is the cancellation

3.2. Selected balanced kites configuration

21

of any forces introduced by changing the angular momentum of the drum because the rotation axes are concentric. The downside of the set-up is the working point where the tether length remains constant. This is the case when on-board power generation is tested. The winch motor must continuously rotate the drum at the same speed as the arm to maintain the length constant. A good control of the difference in speeds also requires better sensors and control.

Figure 3.8: The winch drums mounted concentric with the carousel shaft but with an independent rotation speed.

3.2 Selected balanced kites configuration Decisions Both the single and double cable approach contain some insecurities in the handling of the landing phase. The double cable alternative shows no real advantages over the single cable configuration but does increase cable and winch costs. It is decided to further developed the single cable concept. The single cable leads to a single winch. This winch is mounted on the bottom of the carousel. By doing so a lot of possibilities remain available. The system can be used with a single kite, the winch is then rotating with the carousel. When used with two kites the winch either remains hanging under the carousel for the fixes connection. Or in case a good swivelling joint is developed, the whole winch construction can even be placed static under or next to the carousel. An overview of the taken decisions and possibilities with the selected configuration: No tilt Mounted on a trailer (category BE) with steel supports Single cable Winch mounted below carousel

– single kite: rotating with the carousel – balanced kites ◦ fixed connection: rotating with the carousel ◦ swivelling connection: winch static under or next to the carousel


22

Conceptual design The figures below show the configuration worked out to concept level. These images do not represent a technical implementation but show the approach for the balanced kite system as the basis to start building a single kite system (cfr. chapter 4). Next, the landing of two balanced kites is described to illustrate the function of each component. While the kites are steered down, the tether is reeled in and wound evenly on the drum. Once the cable is reeled in completely, the swivelling joint lands on the central support point (fig. 3.9). By providing a gradual thickening from the cable to the joint a smooth landing and good fixture can be guaranteed. The following moment is critical. Until now, even in the case of zero wind speed, the kites were driven by the winch which was reeling in the cable. Once the joint landed, the tether length is constant. As described earlier, the next action must be completed within 10 seconds (cfr. 3.1.2).

Figure 3.9: Step 1: landing the airborne connection.

The next action (fig. 3.10) is to fix the short tethers back to the end of the arm so that the carousel can drive the airborne kites until they have landed their supports. To increase the reliability of this step it is important that there is a solid part at the connection point. These solid pieces have a fixed orientation


23

with respect to each other unlike the free moving cables. By rotating the arm at the same speed as the solid top of the swivelling joint two donut-shaped holders can enclose the shorter tethers. Once closed, these holders are moved along the arm to become the outer attachment points for the tethers (fig. 3.11). These donut-like components are quite voluminous. This is due to the requirement that the bending radius of the tether must always remain above 5 cm. In addition to this, the attachment at the end of the arm must allow a change in tether length.

Figure 3.10: Step 2: gripping the short tethers.

The carousel is now rotating and driving the two kites. The final step is to bring the kites back in and land them. Both short tethers must be pulled and the carousel rotation speed controlled until the planes are right in front of their support. Pulling in the short tethers requires a last new component: a drum on which the cables are wound. To use minimal space and material the drum is interpreted here as two upright cylinders (fig. 3.11). By rotating the plate on which these cylinders are mounted with respect to the carousel arm, the remaining length of the cable will become shorter with every rotation until they are reeled in completely.

Figure 3.11: Step 3: Bringing the attachment points to the end of the arm, reeling in the tethers and landing the kites.

3.3. Conclusion

24

3.3 Conclusion The result of the first part of this thesis is the conceptual design of a balanced kites system. Different feasible options are compared and the most likely problems are covered. In comparison with the current indoors carousel, the new set-up sets some new challenges: a full-scale winch, transport, weatherproofing, rotating start and landing with kite-supports and the balanced kites configuration. To avoid an accumulation of practical problems during the production and initial use of the new set-up, it is preferred to implement these features one at a time.

Chapter

4 Design of the single kite set-up

The balanced kites set-up is the goal for this project, but this shall be achieved by incremental improvements. In order to assure a working carousel at the end of the project (2016), the number of uncertain steps must be limited. At this moment the aim is to design a working single kite carousel with a working winch. This chapter gives an overview of the designed carousel structure and the winch.

4.1 Carousel Figure 4.1 shows the new carousel with the winch hanging from the shaft. The arm, shaft and winch rotate as one piece and are supported by the four legs. The red line indicates the routing of the cable: starting from the drum, the tether is guided by a level wind mechanism which leads it upwards through the main shaft. On top of the carousel another pulley leads it towards the end of the arm where it leaves the carousel supported by either a block of nylon or a swivelling pulley assembly.

4.1.1 Frame This section covers the most important design decisions made for the carousel as well as an analysis of the complete structure. This frame is designed to stand by itself or be mounted on the trailer. It consists of a truss-like frame, a long central shaft surrounded by a static cylinder (fig. 4.1). The frame is made of aluminium extruded profiles with a hollow square cross-section. These tubes combine a low cost and good corrosion resistance. The connections of the different tubes exist of M12 bolts and nuts put trough holes in the tube wall and tightened once the truss is assembled. This requires a selection of different sized profiles to fit in each other. Another possibility is to fabricate components that serve as a connection (fig. 4.2). The material (e7/kg) and production cost (e7/h internal ESAT-rate) of nearly 50 of these pieces would largely exceed the extra cost of buying different profiles (e20 difference per 6m-tube). 25

4.1. Carousel

26

Figure 4.1: The designed carousel with a red line indicating the path followed by the tether.

As can be seen from figure 4.1 the frame consists of four legs in stead of three. For a ground based carousel a tripod is a logic option: the production costs are lower and it stands firmly even on an uneven surface. The design of the new carousel on the other hand is strongly linked to the trailer it is mounted on (cfr. 4.2. The available trailer surface is 2 by 4 m and mounting a tripod with 120° spacing between the legs results in two legs fixed at only 1m from the center. The shorter this distance, the smaller the reaction moment a leg can exert, resulting in a lower strength and stiffness in this direction. By using four legs, each one is fixed at the same distance from the center and the forces are distributed more evenly in all directions. This might cause problems when it is placed on an uneven surface or when the parts are not all exactly equal due to the fabrication process. To cope with these differences the support at the end of each leg can be slightly rotated to become parallel with the floor (fig. 4.3). The off-the-shelf rubber feet are connected to these supports with an M16-threaded bar and hence the exact height of each leg can be adjusted separately.

4.1. Carousel

27

(a) Detail of profiles ready to be con- (b) Solid aluminium pieces serving as nected with bolts. connectors between the beams.

Figure 4.2: Two alternative connections for the profiles. In order to reuse some of the materials and to benefit from the gained experience, the new structure uses some of the same techniques and materials of the current indoors carousel. This because of the gained experience of building the first set-up and to use the rest of the material(6m-profiles) bought at that time.

Figure 4.3: Detail of the adjustable supports and feet of the carousel.

It is important to make sure that the designed frame fits on the trailer and to limit its height to 4 meters once mounted. The trailer-platform has a height of 720 mm and because of road regulations the height of the structure must be around 3 meters (total <4m). Based on estimations of the winch dimensions also requirements for the space under the structure are established. There is a height constraint for the lower bearing holder (green part in fig. 4.1) of 1300 mm and a minimum width of 900mm. Because of the iterative character of the design process geometric relations were used to automise the calculation of the different profile lengths and a first order estimate of arising forces (cfr. appendix A.2). These forces are calculated by considering the frame as a perfect truss, hence the tubes are only loaded in axial pressure and tension. The highest force under a load of 2000N (cfr. 2.1) at the top of the shaft is around 13kN for the lowest profile in the construction. The distribution of the fitting profiles is chosen in accordance with the maximal appearing forces.

4.1. Carousel

28

Because of the appearance of pressure forces in the slender beams, a check for the risk of buckling and crippling is performed: Pcr =

π 2 · Ealu · Iprof ile 2 lprof ile

π 2 · 70e9 · −1.159e−6 1.942 = 274 kN with typical safety factor 4 = 68 kN

(4.1)

=

It can be seen that for the profile with the highest loading and respective largest cross-section (80x80x4) the critical loading for buckling is not reached.

4.1.2 Shaft The central shaft is a hollow cylindrical profile. The arm is fixed on top of it and the winch will be hanging underneath. Two bearings are fitted on this shaft to connect the rotating shaft to the frame. The loads working on this component are the weight of the pieces attached to it, forces exerted by the cable, centrifugal effects and the torque applied by the motor (fig. 4.4). The torque T on the shaft is maximal when the carousel is braked to zero rotation speed and the airplane exerts its maximal force (1000N) at the end of the arm(2m). To limit the angular displacement of the shaft to 2° a tube bigger than the current one (100x5mm) must be used. The internal diameter must now be equal to 8 cm, resulting in a thickness of 1 cm. T · Lshaf t = 2° = 0.035rad Galu · Jshaf t T · Lshaf t = Galu · ∆φ 2000 · 2.5 = = 5.7143e−6 m4 25e9 · 0.035 = π/32(0.1004 − d4in )

∆φ = Jshaf t

Jshaf t

(4.2)

din = 0.08 m

Figure 4.4: Shematic representation of the central shaft (rotated 90°).

The natural bending frequency of the shaft with the 100kg-winch hanging underneath is estimated. The maximum displacement for a beam with a load at a distance a

4.1. Carousel to the left from the first support is y =

29 F ·a2 (Lshaf t +a) . 3·E·Ishaf t

The bending stiffness can be

determined as the force needed to a unit displacement of the load: k = Fy . Combined p with the mass of the winch this gives a natural frequency of ωn = k/mwinch = 65 Hz. The carousel will be rotating at 1 Hz maximum so there is no risk for the shaft to be excited at its resonance frequency.

4.1.3 Bearings The rotating shaft must be supported to stay upright but at the same time, it should be able to rotate around its longitudinal axes. Two bearings are used to make this connection in an efficient way and to guide the forces on the shaft to the static carousel structure. The loads that determine the choice of these bearings are represented in figure 4.4. The winch and the arm represent a weight of more than 100kg putting a downwards force on the shaft. In the case of a single kite system the winch is hanging on the shaft and the cable will add a compressive force. For balanced kites, the winch is static and the cable puts only a downwards force on top of the shaft. Because of these downward axial forces a tapered roller bearing is selected. This specific type of roller bearing is designed to perform well under a combination of axial and radial loads. The lower bearing consists of a cheaper ball bearing because it mainly takes radial loads. To improve the connection between the bearings and the shaft two buses are mounted on the shaft at the location of each bearing. The outer diameter of the extruded aluminium shaft is not precise due to its production process. By adding smaller buses, which can easily be finished on a turning lathe, the needed dimensional tolerances are obtained. The outer part of the bearing rests in a piece of aluminium designed to be a tight fit around the bearing. The bus fits tightly in the inner part and is mechanically secured by a locknut. A part of the bus is threaded (M110x2) so that the locknut can be screwed on the bus. When an axial force is applied to the shaft, large shear stresses are put on this thread. Its strength is checked using an empirically proven method [Bee04]. An estimate of the shear area Ath is made using the diameter d0 and the thread depth. By dividing the applied force with this surface, the shear stress is obtained. The result of this calculation show that occurring stress remains well below the yield stress for aluminium. Ath = 0.5 · π · d0 · Le with Le = length of the threaded part 3000N τ= = 840 kP a Ath

(4.3)

The ball bearing at the lower end of the shaft is mounted upside down. The bearing is now put in the support from below and screwed up with the locknut. It will take only radial loads to counteract the moment introduced by the planes pulling at the top of the arm and some of the compression forces in a single kite configuration. When a large upwards force is put on the shaft, this way of construction will prevent

4.1. Carousel

30

the rotating part of the carousel to be pulled out of the structure.

Figure 4.5: An exploded view showing the locknut, bearing, bus and aluminium structure element.

4.1.4 Arm There are several reasons to extend the length of the carousel arm. First of all it increases the clearance of a plane hanging from the end of the arm. Secondly it limits the rotation speed of the carousel. A longer arm makes it also possible to drive the airplanes at longer tether lengths, as was shown in [GGD+ 10]. This makes that higher altitudes can be reached before another energy source is needed to power the airplanes. It also reduces the accelerations acting on the airplanes during the startup phase. For these reasons the new arm will be four meters, this means two meters at each side. The arms must be easy to disassemble in case of long road trips or to replace them by another type. The arm consists of square hollow aluminium profiles connected to an item 1 -profile which serves as connection to the shaft. Because the set-up will first work with one kite only the cable will always remain attached to the end of the arm. During operation the total forces exerted by the plane will be put on the end of the arm by the tether. The cable will apply a vertical and horizontal force component and because these forces are not applied in the center of the tube also some torsion will occur. The bending moment is determining for arm dimensions and results in a 100x100x10mm profile. The effect of the torsion on this arm is calculated using shear-flow in a thin-walled profile and the deflection remains under one degree. 1

http://www.item.info

4.1. Carousel

31

To mount the long arm to the central shaft, two solid aluminium pieces are screwed together. The item-profile is bolted to a plate, which at its turn is fixed by sunken screws to a cylindrical piece. This cylinder contains holes with thread by which it can be hooked up to the shaft. Because the connection between plate and cylinder is critical, eight M12 bolts are used and the length of the thread in the cylinder is checked the same way as for the bearing holders (cfr. 4.1.3).

Figure 4.6: An view of the assembled pieces and exploded view of the carousel arm.

4.1.5 Drive To select a motor to rotate the carousel an estimate of the required power is needed. The carousel is rotating mainly to power the airplanes. Extra torque will be needed to compensate losses in the transmission and bearings and some aerodynamic friction of the arm. The drag that the airplane can exert on the arm depends on its relative velocity squared and so the required torque scales with with the rotation speed squared. The drag is calculated for an airplane at 5 meters from the central shaft giving it a maximum airspeed of 31m/s. The drag coefficient is based on wind tunnel experiments performed in 2011 on the current airplane [CE11]. Figure 4.7 shows the required torque and power delivered to the shaft when flying an airplane with 2m wingspan and 20% losses.

4.1. Carousel

32 200 torque [Nm] power [x10 W] 150

100

50

0

0

10

20

30 40 n [rpm]

50

60

Figure 4.7: Required torque and power in function of carousel speed.

It can be seen that a motor of at least 1 kW must be chosen to drive one plane. To avoid implementation problems, a suited motor and drive are selected with the same supplier (Lenze 2 ) as the larger motor for the winch (cfr. 4.3.3). Because the speed variations of the carousel are lower and there is no need to have a very accurate control of the motor an asynchronous motor is chosen. It has a lower cost but also a higher inertia, meaning a less quick response, than a synchronous servomotor. To control the motor an electronic frequency converter with a power of 1.1kW or 2kW can be installed. In case the configuration is changed to a balanced kites system, only this drive must be switched from 50 to 87Hz and the motor will be able to turn at higher speeds. Changing the belt radius or gearbox, the same motor can now deliver 1.8kW to the carousel instead of 1.1kW. The motor is connected to the lower bearing holder using a piece of aluminium. A belt pulley is attached to the output shaft and the power is delivered through a belt to the shaft.

4.1.6 Weatherproofing Because this set-up is designed for outdoors use, the design incorporates certain measures to make it weatherproof. The structure is not designed to remain on the test field for weeks during bad weather, but while operating the set-up will encounter wind, rain, dust and possibly lightning. Almost every component is made of aluminium, which is highly corrosion 2

http://www.lenze.be

4.1. Carousel

33

resistant. Only for the bolts connecting the different elements stainless steel is used. The stronger and stiffer steel is chosen because these connections are critical and carry the loads through small cross-sections. Nonetheless connecting steel to aluminium is not a good practice in a wet environment. Because of their difference in corrosion potential the corrosion of aluminium would be accelerated. Therefore stainless steel bolts are use which can be connected to aluminium in a non-saline environment without problems [Hat84], certainly because the surface of stainless steel is a lot smaller than the one of aluminium. To avoid that large amounts of water or dust make their way to the moving components provisions are made to cover the whole carousel. The upper bearing holder has straight edges and some spare holes to fix a cheap plastic cover. The bearings themselves have seals to keep the inner parts clean and lubricated. Lightning is also a concern, certainly when the airplanes are airborne during upcoming or unforeseen thunderstorms. When a lightning strikes the set-up mostly the plane, the tether and the electronic components are in danger. Because an adequate ground connection will be required and implemented for the electric system, it is mentioned here that both the carousel and the trailer must be connected to this circuit to provide a low resistance path for any current to the ground.

4.1.7 Analysis To select the different components, hand calculations and first order estimations are made considering strength and stiffness. But once all these pieces are assembled they function as a structure to support the tethered airplanes. The effect of the forces applied by the kites on the whole system must be determined. Because the carousel is a rotating device it is important not only to do this for a constant pulling force of the kites, but also to look at the dynamic behaviour of the carousel. First the occurring displacements, stresses and dynamic modes are calculated for the structure with the arm. Because the arm is easily replaceable if a stiffer arm is required for some experiments the analysis is also carried out for the frame only. Figure 4.8 shows the two models used for the finite element analysis. The model consists of 1D-beams connected in nodes. Each element is associated with the properties of the materials and cross-sections used in that specific part. The figures shows the profiles with their defined cross-sections, but this is for presentation purposes only, the calculations are performed on the unidimensional meshes. The Siemens-NX software is used to construct the model and to visualise the results, the actual calculations are performed by Nastran.

4.1. Carousel

(a) Model of carousel mounted on the trailer

34

(b) Model of carousel only, including arm

Figure 4.8: 1D-beam models used for statical and dynamical analysis. Static loading The loads exerted by the planes are already determined (cfr. 2.1). The static analysis is carried out for the maximal load of 1000N under an angle of 45° at the end of the arm. Also the weight of the winch is taken in consideration. Figure 4.9 shows the occurring stress and displacement under this loading. Note that the deformation of the structure is exaggerated to allow easy interpretation. The displacements are acceptable and the stresses remain below the yield stress of the used aluminium.

(a) Stress in MPa, max: 24MPa.

(b) Deformation in mm: max 13mm. Exaggerated representation

Figure 4.9: Results of the statical analysis at maximal loading. Dynamic loading To assess the dynamic behaviour of the structure the frequency response function (FRF) of the beam-model was determined. Nodal forces were

4.1. Carousel

35

applied at the end of the arm and on the winch position. Also the 100kg winch is modelled because its mass might lower the resonance frequencies of the system. Figure ?? shows the FRF of measured at the end of the arm. This location is chosen because it is most probable to have the largest deformation. The first eigenfrequency of the structure can be found at 8Hz. At this mode the arms of the carousel are flapping up and down (fig. 4.113 ). This poses no problem for operational use because the loads are expected to remain at 1Hz.

Figure 4.10: Displacements of the arm divided by the applied force, in function of the load frequency (FRF).

Figure 4.11: First displacement mode of the structure (8Hz).

3

The color gradient represents displacements but has no useful physical meaning because these displacements scale with the applied force.

4.1. Carousel

36

Without arm If this frequency does pose a problem in the future the arm can easily be replaced by one with a higher stiffness, increasing this frequency. Once using the carousel for the balanced kites configuration, the cable will be supported at the top of the shaft in stead of at the end of the arm. For these reasons the analysis is conducted without the arms and the loads placed on top of the structure. Figure 4.12 shows the results of the static analysis in the weakest direction. The occurring stresses are far below the limits.

Figure 4.12: Results of the statical analysis without arm: stress (colors, MPa) and displacements (exaggerated).

The dynamic analysis shows the first resonance frequency of the structure can be found around 28Hz. The deformation mode at this frequency is a torsion of the whole structure. This is to be expected because the legs have no interconnections. The FRF for frequencies between 0 and 50Hz is show in figure 4.13and a representation of the first mode in figure 4.143 . Because of this relatively high frequency and good static results, no action is required to change the behaviour of the structure.

4.1. Carousel

37

Figure 4.13: Frequency response function for the top of the carousel (no arm).

Figure 4.14: First displacement mode of the structure without arm (28Hz).

4.2. Trailer

38

4.2 Trailer Following the reasoning made in section 2.4 different trailers were compared based on their dimension, cost and possibilities to mount a support system. Details can be found in appendix A.2. This section describes the selected trailer and gives the results of a structural analysis of the carousel mounted on the trailer.

4.2.1 Selected trailer The final decision fell on a trailer from a dealer experienced in mounting custom made steel supports by welding them to the frame. The trailer is a double-axled car transporter with a maximum allowable mass of 2500 kg and outer dimensions of 4 by 2m. Underneath two steel profiles of 60x60x4 are mounted, leading to four support legs.

Figure 4.15: A front and side view of the trailer with support beams (red).

4.2.2 Installation of the carousel To connect the carousel to the trailer there are different options. The trailer consists of a large steel frame covered with smaller support beams and a wooden or metal floor plate which can be removed. The plate is not a good option to use as basis for the carousel. It is not designed to take high discrete loads and has a low stiffness. The two other options are to use the trailer frame or the newly added support beams. The four carousel legs must be put as far apart as possible for stability and stiffness. Neither the dealer nor the producer of the trailer could provide us with a detailed drawing of the trailer. By fixing the trailer to the support beams with know properties, the unknown factors in the carousel design can be avoided. For these reasons it is decided to fix the carousel to the new support beams. These beams run along the whole width of the trailer and are mounted at a distance that can be chosen. The beams are placed at a distance of 2900mm from each other and have a length of 3000mm, this creates an almost symmetric support area in all directions. The carousel will be fixed to the trailer by removing the installed feet and using the holes to bolt them to the support beams. A rubber or plastic sheet between the steel supports and the aluminium frame is recommended to avoid electrical contact and the galvanic corrosion resulting from it.

4.2. Trailer

39

4.2.3 Analysis Static loading The finite element model is extended with the steel support beams and the screw driven legs to estimate the dynamic behaviour of the carousel once it is mounted on the trailer. The arm is not integrated in this model because of the reasons mentioned earlier (cfr. 4.1.7). Due to the lack of data about the trailer its stiffness in different directions cannot be plugged into the model. Because the mainframe of the trailer exists of two steel beams running from the tow bar to the end, it is considered to be able to maintain the carousel legs in the front and in the rear at a fixed distance. The mass of the trailer is taken equal to 500kg and is distributed over the attachment points of the carousel to the trailer. The static loading of two kites is applied on top of the structure. The stresses and displacements remain well within the limits. It can be seen that support beams are bending slightly under the applied loads. The maximal deflection is 1.8mm.

Figure 4.16: Displacements of the structure under statical loads only.

Dynamic loading The results of the analysis with dynamic loads can be seen in figures 4.17 and 4.18. The first natural frequency is around 4Hz. This is four times higher than the carousel turning speed, but the margin is not very big. It must be noted that the model does incorporate the trailers weigth but not all of its rigidity. The experiments will need to show what the exact behaviour is and most of all at which frequency the kites load the trailer. The previous analyses show that the stresses are within limits in the different cases. In case the displacements of the

4.2. Trailer

40

trailer and its supports are not satisfactory, cheap extra supports (fig. 4.194 ) can be added. These metal bars have no mechanism to lift the trailer, but once it is on the four main legs, these extra supports can be clamped into place. They will distribute the weight of the system and limit displacements in different directions. This increased stiffness will increase the margin to the first resonance frequency.

Figure 4.17: Frequency response function for the top of the carousel and trailer.

4

Source: online advert: http://gratis-zoekertjes.onlinertjes.be/d-1572/230517/aanhangwagens-enkarren-steun-met-klembeugel.html

4.3. Winch

41

Figure 4.18: First deformation mode of the structure under dynamic loads (around 4Hz).

Figure 4.19: Example of a cheap extra support, to be clamped if needed to the trailer.

4.3 Winch The current set-up is equipped with a small 400W-winch to allow testing with a variable cable length. This is of course not enough to test airborne wind power at its full capacity. As can be seen from the specifications a 10kW-winch is needed. The

4.3. Winch

42

aim is to have as many off-the-shelf components as possible, so different companies and AWE-projects were contacted with the question if they can deliver such a winch. Most winches used in daily life are designed to lift high loads at relative low speeds. The engine drives the drum through a mechanical transmission, often a worm gear, and delivers the requested tension on the cable. The systems are never used in the opposite way, the load never drives the engine to produce electricity. The worm gears have a self-breaking effect and cannot be used in this direction. Only some large scale cranes use regenerative braking, but these systems are to big to be taken in consideration. Just like the pulley sheaves most commercial products are designed to work with steel cable of diameter of over 5mm, whereas this project will be using the 3mm dyneema. The lower bending radii and bigger radii of supporting grooves would largely decrease the cable life. So a custom solution is needed. Due to the lack of commercial solutions, it was concluded that the design of the winch could no longer be postponed. This sections describes possibilities, the resulting design and issues to consider the following years of the project. The main components for a winch are the drum, the motor and transmission, a level wind mechanism and the frame (fig. 4.20).

(a)

(b)

Figure 4.20: The designed winch with and without frame.

4.3.1 Drum The tether is wound on a cylinder or drum, which is fitted on a shaft provided with bearings. The current winch has a grooved drum to make sure every winding lays next to the other. Since the results of this system were not satisfactory, the new drum must have a smooth surface combined with a level wind mechanism. The combination of diameter and length of the cylinder determine the maximum cable length that can be wound on a single layer. Starting from the maximum reel-in speed, the diameter of the drum also determines the needed rotation speed.

4.3. Winch

43

ltether · φcable estimation of needed drum length φdrum · π vin = · 60 required rotation speed φdrum · π

Ldrum = ndrum

(4.4)

The drum could be made from steel, aluminium and (carbon fibre) composites. The evaluation between these materials must be made by comparing weight and inertia, corrosion resistance, availability and cost. The weight and following inertia is important because it is determining the reaction speed of the winch. A drum with a large inertia around its rotation axis will be harder to brake or accelerate during operation, making it harder to maintain the tension on the cable. Aluminium scores better than steel regarding weight and corrosion resistance. A composite solution would even be better but tubes with diameters between 200 and 500mm and a length of less than 1m are hard to find. Also the direction of the fibres in such pieces are important. The drum will be loaded with a radial pressure due to the tensioned cable. If the composite tube was designed for another load case, it would not be able to resist the pressure. Aluminium tubes can only be bought with a maximum diameter of 250mm, however there are more options for thickness for the 200mm tubes. The material cost of the tube itself is quite high (± e600) because the tubes can only be bought starting from 3m while less than 1m is needed. Several AWE-projects with experience in building winches were contacted. Finally an agreement was made with Swiss Kite Power 5 , a project of three non-profit organizations and Alstom. The team was producing a new winch themselves, by sharing the material and production cost, a complete drum mounted on a shaft is bought for around 1000 euro.

4.3.2 Level wind mechanism The level wind or spooling mechanism is a guidance system for the tether during the operations of the winch. If the winch would reel the cable in without any guiding mechanism the different loops would lay on the drum in a random way. Parts of the tether laying on top of the other creates high pressure regions in the cable and its conductors. It makes it harder to reel the cable back out at a uniform speed and it increases the risk for the cable to get entangled. Ensuring the level winding of the cable can be done by simply guiding it back and forth accordingly to the drums rotation speed (fig 4.21). A well-placed pulley can be moved back and forth using a linear actuator. The control of this movement must exactly match the speed of the winch so every loop lays exactly next to the previous one. The easiest way to link both speeds together is by connecting the drum and the level wind mechanism through a mechanical transmission. 5

http://www.swisskitepower.ch/

4.3. Winch

44

A second way to do this is to electronically control the speed of the linear actuator. A good winch speed sensor is now required combined with an accurate electric control of the spooling mechanism. The reason to use a similar set-up is because of the flexibility in the control. A simple mechanical system is fixed to one speed ratio. The speed ratio in system driven by an electromotor can be changed by simply changing a software parameter. This means that it can be tweaked until it exactly matches the best speed ratio for a given tether. Changing the tether, adding or removing conductors in the core can be done without any problem. This flexibility also offers possibilities to use longer tethers without any change. The pumping of the kites can still only be done on the last layer on the drum to avoid cable wear and spooling problems, but an additional layer can be put on top of this layer because the forces are lower during start-up and landing. The electromotor can just switch direction at the end of the drum and put the next layer in the other direction. Due to the advantages, an electrical level wind mechanism will be used. There are several linear actuators available on the commercial market, so there is no need to design one. Both belt and screw mechanism are available which can withstand the maximum forces of 2000N. Due to elongation of the belt a screw seems more appropriate in this application. A system of Hiwin and Hepco-Motion were compared. Both have prices in the range of e1400 for the mechanical parts going to e3500 for a complete system. The biggest advantage of the Hepco-Motion system is that is completely closed, whereas the Hiwin-variant can only be closed by an additional folding bellow. Finally, the SDM-module of Hepco-Motion is chosen.

Figure 4.21: The level wind mechanism: a linear guide mechanism moves pulleys along the drum.

4.3.3 Winch drive The winch motor/generator is in fact the heart of the winch. It drives the winch drum and generates electricity while it is driven. Some large winches, mainly on ships, use hydraulics to drive the winch drum. In this case that would mean an extra hydraulic system must be inserted between the electric generator and the drum.

4.3. Winch

45

This brings extra complexity and energy losses. The electric motor must have a maximum power of 10kW and a rotation speed which is maximum 3 to 4 times higher than the rotation speed of the winch. An accurate and fast speed control is required to have a good control of the tether speeds and tension. The exact length of the tether is not critical but the speed at which it is reeled in or out is important for the control. To maintain the cable tension a torque-controller is certainly an advantage. To have a good control performance a synchronous servo motor is preferable over a cheaper asynchronous motor. These AC-motors cannot just be plugged into the net. The control is done by motor drives that consist of power electronics and logic to control the current through the windings following different parameters. It is also this device that can switch the motor within its four quadrants. When the torque inverses at the requested speed, the driver puts the energy of the generator on the outputbus. The cost of these drives is often the same or more than the cost of the motor itself. To order the motor, decisions need to be made concerning the isolation class, the mounting position (for oil levels and sealing) and gearboxes. The gearbox cannot contain a worm transmission because the motor must work in all four quadrants. The ratio of the gearbox can be chosen in a certain range. The choice depends on the cost versus the possibilities of putting the ratio in the belt drive. Belt transmission The transmission between the motor and the drive consists of a belt transmission. This is a cheap and easy way to make this connection and insert a speed ratio. Due to volume restrictions the motor is placed parallel to the drum with the belt transmission on the bottom of the winch. To prevent the belt from slipping and introducing a difference in speed between motor and drum, a timing belt is used. The drum, delivered by SwissKitePower, is provided with a Madler HTD-belt pulley. Pulleys for the motor and tensioner are ordered at the same company to have no problems with interchangeability.

4.3.4 Frame and integration All the functional winch components are put in a frame of Item MB -profiles. These extruded aluminium profiles are more expensive than the ones used for the carousel structure, but allow larger flexibility. Their specific cross-section allow several sorts of connections which can be moved later. To make the winch rotate with the carousel it can either be put on a separate platform or hang on the main shaft. A rotating platform requires a third bearing or several small guiding wheels along the edge to support it. This platform must then mechanically be connected to the rotating carousel shaft to be driven by

4.4. Electrical equipment

46

the carousel motor. This method requires an extra bearing and a coupling which allows production tolerances for the alignment of the shaft and the platform. It is easier to hang the winch underneath the carousel, a tapered roller bearing is used to counteract the increased axial load on the shaft. A solid aluminium piece provides the connection between the circular shaft and the winch frame. The top is the frame is provided with four profiles which can easily be loosened and shifted in any direction (fig. A.3). The exact position of the center of mass will change when components are added or replaced. The slideable connection allows repositioning of the winch under the central shaft. The frame is made in such a way that it can easily be covered from the outside. A plastic foil or plate can protect the winch from dirt and rain. The profiles allow addition of extra framework if this is required later.

Figure 4.22: The winch mounted on the carousel shaft.

4.4 Electrical equipment The aim of airborne wind power systems is to harvest the energy in the atmosphere and transform it into electrical power. The aim of this carousel is to perform experiments and not to produce electricity. Nonetheless the motors and electronics need to have a power supply and any produced electricity shall be dissipated using dump resistors. This section gives a short overview of system this carousel will use. Once the system is running, it might be expanded to increase its performance. Improving the electrical side of the set-up will be one of the goals of new master thesis in 2012-2013. Figure 4.23 shows a schematic representation of the needed components.

4.4. Electrical equipment

47

Both the winch and carousel motor are controlled by a motor drive. These drives run on a three-phase AC network (3x230V) or 600V DC. As long as the set-up is tested in the vicinity of its storage building, they can be connected with three-phase plug to the grid. Once a remote test-location is found, an independent power source is required. There is a wide range of diesel and gasoline small-scale generators providing a three-phase and/or single phase output. When the carousel is switched to such a power source it is advised to also integrate a high-capacity battery (red components in fig. 4.23), because the generators have a less stable output than the grid. To combine this battery with the three-phase or 600 V line a (10kW) two-way converter is indispensable. A similar type of converter is either hard to find off the shelf or very expensive. Some international partners (e.g. SwissKitePower) have shown interest in developing such a converter in co-operation to reduce the costs.

Figure 4.23: Overview of the electrical equipment (power, no signals), coloured components are optional.

Next to the three-phase supply the on-site computers and other accessories require a normal 230V AC-connection. In a first stage this power can come from a nearby connection to the grid. On a test-location an extra 230V-line is needed. To protect these more delicate electronics from voltage surges this line needs to be protected by a voltage regulator. The green component in figure 4.23 represents this one-way transition from the power supplied by the generator to a secured 230V-line. It is also possible to buy a more expensive generator with a build-in

4.5. Cost estimates

48

voltage regulator. To avoid computers or controllers shutting down because of generator problems during operation, a back-up battery must be used for this line. When the tethered airplanes go into a pumping mode, they will drive the winch. This will force the winch motor to go in generator mode. In this case the motor controller notices an increased voltage and puts it on a DC-output bus. The easiest solution is to dissipate this energy by connecting a dump or brake resistor. These resistive elements transform the electricity into heat which they pass to the surrounding air by convection. Once the outdoors carousel is running successfully, the resistors can be replaced with power electronics to put the produced electricity in batteries or on the grid. Sliprings In the single kite set-up there are two connections in the electrical system that require power and signals to go from a static conductor to a rotating one. First we have the winch rotating with the carousel while most of the electrical components are static. A second problem arises due to the conductors in the tether. Control signals for the airplane must be send through these conductors but the whole cable is spooled on the rotating drum. The technical solution for this problem is a so-called slip ring. These connectors use a sliding contact, liquid mercury or induction to transmit the signals from the stator to the rotor. Depending on the requirements, different types can be chosen. In this stage the design provides space to mount these slip rings. The shaft of the winch is ready to mount a through bore component (φ40mm) (fig. 4.24). The space between the bottom of the winch and the trailer can be used to mount different slip rings according to configuration of the electrical equipment.

Figure 4.24: A through bore slip ring to connect the static controller to the rotating conductors in the tether.

4.5 Cost estimates For the detailed design of the single kite set-up the costs of the different elements have been estimated and monitored. Appendix A.4 shows a detailed overview of the cost estimates based on material prices, work hours (internal workshop rate) and

4.5. Cost estimates

49

quotes from suppliers always rounded up. This results in a maximum cost of e24000, which is within the maximum stated in the specifications (2.1). An important remark is that this total has a high sensitivity to the price of the electrical equipment and motors for which no exact price is available yet.

Chapter

5 Experimental proof of rotating start and landing

The indoors carousel has experimentally proven that tethered airfoils are stable during the rotation of the arm and variation of the tether length. However during start and landing it happens sometimes that the airplane collides with the structure. The new carousel will have a higher arm to account for this, but the proof was needed to show that the planes can also land and take-off from a support. This section describes the experiments that were performed on the current carousel. Set-up To deliver this proof-of-concept a small cheap airplane support or cradle was build. It consists of an angled beam ending in a plate with a foam protection. A V-shaped cut-out allows to catch the cable when it comes back down. The length of the beam and the angle under which it is placed is made adaptable to fine-tune its position during the experiments. To perform the experiments both the carousel rotation and the winch need to be controlled. Both motors are controlled through an Orocos-software environment. Within the framework of this thesis, the code of the winch controller was expanded

(a) Conceptual view of the tethered air- (b) The cradle used for the first rotating plane on its cradle. start and landing experiments.

Figure 5.1: The concept and experimental set-up for an airplane cradle.

50

Chapter 5. Experimental proof of rotating start and landing

51

with safety functions, consisting of limits for tether length and winch speed. Because of the cage in which the carousel is rotating, the winch cable must remain below a certain length. Otherwise the airplane would collide with the walls. The speed must be limited because the small winch on top can operate at high speeds when there is no high load on the cable. Sending a wrong command results in a very abrupt movement of the winch drum. During operation this would cause high disturbances of the kites trajectory and a high risk for crashing it. For these experiments the airplane is not controlled. As long as it remains under the arm, it flies a stable trajectory. So these landings can be executed even if controllers have failed. Results With these safety features implemented, some experiments were performed. The winch limits and structure were checked by a simple mass connected to the tether. Next, the plane was installed. The carousel was rotated at half of its nominal speed and by releasing the tether, the airplane separated nicely from the cradle. During the landing phase the used airplane proofed to remain stable even when the cable landed next to the V-shape due to a sudden deceleration of the carousel. A simple acceleration sufficed to get the airplane above the cradle again and land it smoothly. The experiment is repeated with the cradle under different angles. A higher angle corresponds to the airplane landing at higher speeds, which is ideal to minimize the effect of the wind on total speed vector. The kite can then land at a high rotation speed while the relative speed difference at the up- and downwind side remains small.

(a) Cradle angled at 50°from horizontal.

(b) Cradle angled at 30°from horizontal.

Figure 5.2: The airplane taking off from the cradle seen from the carousel arm. Conclusion The behaviour of the kite was satisfactory in all the performed experiments. It is proven that a tethered airfoil can take-off and land from a cradle in a stable way without any control action on the airplane itself. The next step is to perform similar experiments outdoors to examine the influence of disturbances.

Chapter

6 Conclusion

6.1 Evaluation of the project goals This master thesis has taken the next steps in the experimental research of the Highwind-project. These steps consist of proving the concept of a rotating start and landing and investigating new aspects such as increased mobility, finding a suitable test-location, tilting the platform and using a balanced kites configuration. Using a simple airplane cradle, the rotating start and landing was executed successfully at different rotation speeds on the current set-up, proving the principle with an uncontrolled aircraft. After the investigation of the different ideas, the different functions of the new carousel were formulated as a more detailed list of specifications. Starting from this document a conceptual design of a balanced kites set-up was made. The structure is mounted on a trailer to move it from one test-location to another but cannot be tilted. The double kite design incorporates a number of new functionalities and to avoid practical problems during the production and operation the concept was translated into a detailed single kite design. Materials and components have been ordered and technical drawings are made to produce this set-up. Some adjustments are made to comply with the limits of the production facilities. The set-up is expected to be operational within a few months. The results of this thesis are part of a 5-year project. The new test set-up gives the team the possibilities to perform more extensive tests and to push airborne wind energy to an economic feasible energy production method.

6.2 Future work Once every component is assembled the carousel should be tested standing alone as well as mounted on the trailer. After fixing any practical problems that may arise from these tests, the system is ready to perform the first outdoor test of the Highwind-project. The next goals are to test in windy conditions with longer tethers 52

6.2. Future work

53

and to perform the transition from a rotating start to a wind powered trajectory. To increase the produced power a new plane will be developed next year. With a lager wingspan and customization of the control surfaces and weight distributions, this new aircraft can be optimized for the application. The new outdoors carousel can be equipped with airplanes up to 2m wingspan. A new master thesis will also be written concerning the electronics and electrical equipment of the carousel. With a two way connection to the power grid the carousel might proof the real possibilities of airborne wind energy.

Appendices

A1

Appendix

A Appendices

A.1 Belgian road traffic regulations Trailer dimensions and turning radii This is selection of the regulations publicated for vehicles used on Belgian public roads. Only criteria considering a towed vehicle or important for the set-up are mentioned here. Source: http://www.ejustice.just.fgov.be/ File number: 1987-01-09/30 Date: 10-02-1987 Total lenght of car and trailer ≤ 18m The width of vehicles (expect motorised transport vehicles) ≤ 2.5m, this is without lights and tires The height of a vehicle cannot be more than 4m. Turning radii:

– when turning the vehicle in a turn with radius 12.5m, no part of the car or trailer can exceed the tangential to this circle with more than 0,8m – the total width of the track cannot exceed 7.2m Driving license A driving license of the type B+E is required in the following cases: the maxiumum allowable mass of the trailer is higher than 750kg AND

– the combined maximum allowable mass of the car and trailer exceeds 3500kg OR – the maximum allowable mass of the trailer is higher than the empty mass of the car

A2

A.2. Calculations of profile lengths and forces

A3

A.2 Calculations of profile lengths and forces Due to the iterative character of the frame (fig. A.1) design, the length of each profile is determined using a script. The same geometric relations are used to estimate the occuring forces in the truss. This way, the profiles with the largest cross-section are used at the places with the highest loading.

Figure A.1: Definitions used in the geometrical relations.


A4

L2 = chosen L3 = chosen x1 = ltip−center − lsupport − hclearance − rbearing x1 L1 = cos(alpha) L5 = hclearance − rbearing L22 − L12 − L32 ) −2 · L1 · L3 φ=β−α

β = acos(

x3 = x1 − L3 · cos(φ) y3 = h1 + L3 · sin(phi) x6 = lplane − rbearing p L6 = (y3 − y6)2 + (x3 − x6)2 q 2 L4 = (y3 − h2)2 + (x3 − rbearing ) h2 − y3 L4 = asin(y3/L2) − α x6 − x3 λ = acos( L6 γ = acos(

For each node in the truss, a force equilibrium delivers one of the equations to determine the beam forces.   − cos (α) − cos ( + α) 0 0 0 0 0 1 0 0 0     − sin (α) − sin ( + α) 0 0 0 0 0 0 1 0 0       0 cos ( + α) − cos (φ) − sin (Γ) 0 − cos (λ) 0 0 0 0 0       0 sin ( + α) sin (φ) − cos (Γ) 0 sin (λ) 0 0 0 0 0      cos (α) 0 cos (φ) 0 −1 0 0 0 0 0 0      0 − sin (φ) 0 0 0 0 0 0 0 0  A=   sin (α)    0 0 0 0 1 cos (λ) 0 0 0 0 −1       0 0 0 0 0 − sin (λ) −1 0 0 0 0       0 0 0 sin (Γ) 0 0 0 0 0 −1 0       0 0 0 cos (Γ) 0 0 1 0 0 0 0    0 0 0 0 0 0 0 0 0 2980 h1


A5



0



             X=             

0

                          

0 0 0 0 (L3) (L4) (L1) (L2) (−2980 · L1 − L3 · h1)

F = A/X Vector F contains the beam forces

A.3. Trailers

A6

A.3 Trailers Different trailers in the range of 2x4m were compared. An image can be found in figure A.2. The requested drivers license, the price and the options for supports can be seen in the table below. The selected trailer is the Eduard 72cm that can be seen in figure A.2e.

(a) Source: Vermeir aanhangwagens

(b) Source: Vermeir aanhangwagens

(c) Source: JEK

(d) Source: Group Cé

(e) Source: Eduard

Figure A.2: Trailers within the 2x4m-range available in the vincinity of Leuven.

Type AMT 2500 AMT 3000 SARIS SARIS zonder zijborden MSX 1500 MSX 3000 EDUARD - 72 cm EDUARD - 56cm

license BE BE BE BE B BE BE BE

supports both(sliding) both(sliding) both(sliding) both (flip down) both(flip down) both (flip down) large large

price 3473,86 euro (+ 24,20) 3974,85 euro (+ 24,20) 3012,67 (+ 24,20) 2473,61(+ 24,20) 3522,38 euro (+ 96,80) 3522,38 euro (+ 96,80) 3049,13 + 411 euro borden = 3460,53 3200,3 + 411 euro borden = 3611,78

A.4. Cost estimates

A7

A.4 Cost estimates This appendix contains a list of the ordered components and their cost estimates.

Stuklijst Highwind Carousel statief + as profiel A profiel B profiel C profiel D profiel E profiel F profiel poot lagerhouder lagerhouder boven omhulsel as lagerbus lagerbus - boven as poot: schroefdraad+rubber groefkogellager kegellager lager zetschroef motorhouder moeren, bouten en rondellen arm profiel 2 m bevestiging arm - plaat bevestiging arm-as vliegtuigsteun middenstuk item

aantal stukprijs 4 22 4 22 4 40 4 45 4 15 4 15 4 15 1 200 1 200 1 15 1 53 1 53 1 201 4 12 1 1 2 1 100 1 250

prijs

Subtotaal

88 88 160 180 60 60 60 200 200 15 53 53 201 48 0 0 0 100 250

1916

Totaal

2 1 1 2 1

110 80 80 0 0

220 80 80 0 0

420

5 4 1 1

0 0 50 500

0 0 50 500

550

motor riem riemschijven controller

1 1 3 1

1000 100 300 1000

1000 100 900 1000

3000

elektronica dieselgenerator invertor voltagecontroller communication energy dissipation

1 1 1 1

500 400

500 400 0 0

winch motor trommel riem riemschijven lineaire geleiding kar geleiding frame motor geleiding bevestiging winch controller

1 1 1 3 1 1 1 1 1 1

6000 1000 200 100 3500 100 300 1000 100 2000

6000 1000 200 300 3500 100 300 1000 100 2000

14500

Trailer

1

3500

3500

3500

cable management katrolschijven lagers staande houder assembly einde arm

23886

aandrijving

2?

Figure A.3: A list of the ordered components and their cost estimates.

Bibliography

[AC08]

C.L. Archer and K. Caldeira. Atlas of high altitude wind power. 1:1, 2008.

[Bee04]

A. Van Beek. Ontwerpen op bedrijfszekerheid, 2004.

[CDL00]

Y. Cabooter, L. Dewilde, and M. Langie. Een onderzoek naar mogelijke locaties voor windturbines. Technical report, Vrije Universiteit Brussel, 2000.

[CE11]

D. Cosaert and K. Elst. Ontwerp van een lier voor & modellering en controle van, de verankerde vlucht van een modelvliegtuig. Master’s thesis, KU Leuven, 2011.

[ES08]

K. Evers and B. Struyven. Modellering en controle van verankerde kites en vliegtuigen. Master’s thesis, K.U. Leuven, 2008.

[evD06]

FOD Mobiliteit en vervoer and DGLV. Richtlijnen betreffende de bebakening van hindernissen voor de luchtvaart., 2006.

[GGD+ 10] K. Geebelen, J. Gillis, M. Diehl, J. Swevers, and D. Vandepitte. Modelling and control of rotational start-up phase of tethered aeroplanes for wind energy harvesting. Master’s thesis, KU Leuven, 2010. [GGG+ 12] J. Gillis, J. Goos, K. Geebelen, J. Swevers, and M. Diehl. Optimal periodic control of power harvesting tethered airplane. In Proceedings of the American Control Conference 2012, 2012. [Hat84]

J. E. Hatch. Aluminium, properties and physical metallurgy. ASM, 1984.

[HD07]

B. Houska and M. Diehl. Optimal Control for Power Generating Kites. In Proc. 9th European Control Conference, pages 3560–3567. 2007.

[KB887]

Koninklijk besluit inzake de massa’s en de afmetingen van de voertuigen welke in het internationaal verkeer tussen de lidstaten van de benelux economische unie zijn toegelaten., 1987. A8

Bibliography

A9

[Loy80]

M.L. Loyd. Crosswind Kite Power. Journal of Energy, 4:106–111, 1980.

[Sav06]

Design info - pulley guidelines, 2006.

An experimental set-up for energy generation using balanced kites

Recommend Documents