Dynamics And Control Of A Multi-Tethered Aerostat .

2y ago
18 Views
2 Downloads
9.59 MB
184 Pages
Last View : 1y ago
Last Download : 3m ago
Upload by : Fiona Harless
Transcription

Dynamics and Control of a Multi-TetheredAerostat Positioning SystembyCasey LambertDepartment of Mechanical EngineeringMcGill University, MontrealCanadaOctober 2006A thesis submitted to Mc Gill Universityin partial fulfillment of the requirements of the degree ofDoctor of Philosophy Casey Lambert, 2006

1 1Library andArchives CanadaBibliothèque etArchives CanadaPublished HeritageBranchDirection duPatrimoine de l'édition395 Wellington StreetOttawa ON K1A ON4Canada395, rue WellingtonOttawa ON K1A ON4CanadaYour file Votre référenceISBN: 978-0-494-32203-1Our file Notre référenceISBN: 978-0-494-32203-1NOTICE:The author has granted a nonexclusive license allowing Libraryand Archives Canada to reproduce,publish, archive, preserve, conserve,communicate to the public bytelecommunication or on the Internet,loan, distribute and sell th es esworldwide, for commercial or noncommercial purposes, in microform,paper, electronic and/or any otherformats.AVIS:L'auteur a accordé une licence non exclusivepermettant à la Bibliothèque et ArchivesCanada de reproduire, publier, archiver,sauvegarder, conserver, transmettre au publicpar télécommunication ou par l'Internet, prêter,distribuer et vendre des thèses partout dansle monde, à des fins commerciales ou autres,sur support microforme, papier, électroniqueet/ou autres formats.The author retains copyrightownership and moral rights inthis thesis. Neither the thesisnor substantial extracts from itmay be printed or otherwisereproduced without the author'spermission.L'auteur conserve la propriété du droit d'auteuret des droits moraux qui protège cette thèse.Ni la thèse ni des extraits substantiels decelle-ci ne doivent être imprimés ou autrementreproduits sans son autorisation.ln compliance with the CanadianPrivacy Act some supportingforms may have been removedfrom this thesis.Conformément à la loi canadiennesur la protection de la vie privée,quelques formulaires secondairesont été enlevés de cette thèse.While these forms may be includedin the document page count,their removal does not representany loss of content from thethesis.Bien que ces formulairesaient inclus dans la pagination,il n'y aura aucun contenu manquant. Canada

AbstractThe focus of this study is the dynamic behaviour of a novel aerial positioning system, whichconsists of a lighter-than-air aerostat attached to a series of actuated tethers in the form ofa tripod. The objective of the positioning system is to achieve accurate station-keepingof a payload at heights up to 500 m. This work encompasses comprehensive dynamicsmodeling, experimental investigation, and advanced controller development and analysisof the tri-tethered aerostat system. The system's characteristics and positioning objectivesare targeted to receiver placement during the operation of a large-scale radio telescope.The experimental test facility, built at one-third scale of the proposed radio telescope,was constructed for the dual purpose of measuring the precision of the positioning systemand providing a basis for validating our computational model. The main components ofthe experimental apparatus are a 18-m helium aerostat, three synthetic braided tethershundreds of meters in length, an instrumented payload, and three computer-controlledwinches. The dynamics model of the tri-tethered aerostat system is achieved by discretizing it into a series of lumped-masses, and estimating the aerodynamic properties of each ofits physical e1ements. The disturbance input to the model is provided by a wind model thatincludes stochastic turbulent gusts.A series of experimental fiight tests were conducted over a two-year period to studyboth the uncontrolled and controlled response of the system. This leads to an incrementalmodel validation process where the passive elements of the model, such as the tethers andthe aerostat are validated prior to the verification of the entire closed-Ioop system. Thetether model shows excellent agreement with the measured response at a temporal level,while the validation of the aerostat model is conducted at a statistical level due to theuncertaintyassociated with recreating the actual test wind conditions.The controlled experiments were performed using basic proportional, integral, andderivative (PID) feedback gains applied to the position error of the payload. The closedloop system effectively reduces the standard deviation of the payload defiections to lessthan 10 cm over the range of operating conditions tested. A comparison of the predicted

c1osed-Ioop behaviour of the dynamics shows favourable agreement to the experimentalresults, indicating that the tether actuation system is modeled appropriately.To improve on the PID controller tested in the field, various optimal control strategiesare investigated that exploit the availability of our validated dynamics model. The optimalcontrollers, which are designed and simulated using linear time invariant versions of thedynamics model, result in approximately double the precision of the PID controller. Thecontrol system is improved further by incorporating a feedforward control input basedon measurements of the primary disturbance force acting on the system. In general theadvanced control techniques, tested in simulation, offer encouraging results that suggestthe positioning system should exceed the requirements of the radio telescope application.11

RésuméCette thèse porte sur l'étude de la dynamique d'un nouveau système de positionnementaérien, composé d'un aérostat attaché à une série de câbles actionnés montés enforme detrépied. L'objectif du système de positionnement est de stabiliser une charge utile à unehauteur de 500 mètres. Cette étude présente la modélisation dynamique complète, unephase d'expérimentation, ainsi que le développement et une analyse avancée du systèmede positionnement aérien. Les caractéristiques et la position du système sont définies enfonction d'un récepteur d'un radiotélescope à grande échelle.Un prototype du système à l'échelle 1:3 a été développé dans le but d'évaluer laprécision du système et de valider notre modèle mathématique. Les composantes principales de l'appareil expérimental sont: un aérostat gonflé à l'hélium de 18 m; trois câblestressés synthétiques de plusieurs centaines de mètres de longueur chacun; une charge utileavec des capteurs; et trois treuils commandés par ordinateur. Le modèle dynamique estobtenu en discrétisant le système en une série de masses ponctuelles et en estimant les caractéristiques aérodynamiques de chaque composant physique. La perturbation du modèleest définie par un modèle éolien qui tient compte de rafales turbulentes stochastiques.Une série d'essais expérimentaux de vol a été effectuée sur une période de deux anspour étudier la réponse du système avec et sans commande. Ceci nous a mené à un procédéde validation incrémentaI, les éléments passifs étant validés avant la vérification du systèmecomplet. Le modèle des câbles montre un excellent accord avec la réponse mesurée à unniveau temporel. Par contre, la validation du modèle de l'aérostat est effectueé à un niveaustatistique dû au fait de l'incertitude liée à la modélisation des conditions réelles du vent.La commande du système a été réalisée au moyen d'un asservissement de type PlO(Proportional-Integral-Oerivative) appliqué à l'erreur de position de la charge. Le systèmeréduit considérablement l'écart-type des déplacements de la charge à moins de 10 cm pournos conditions d'essai. Une comparaison du comportement prévu du modèle dynamiquemontre un accord favorable avec les résultats expérimentaux, indiquant que les câbles actionnés sont modélisés convenablement.iii

Afin d'améliorer l'algorithme de commande primaire testé sur le terrain, plusieursstratégies de commande optimale ont été étudiées, nécessitant entre autres un modèledynamique validé de notre système. Les contrôleurs optimaux, qui sont développés etsimulés en utilisant des versions linéaires du modèle dynamique, donnent une précisionapproximativement deux fois meilleure que notre contrôleur antérieur. Le système decommande est amélioré davantage en utilisant un asservissement de l'entrée basé sur desmesures de forces de perturbation primaires agissant sur le système. En général, les techniques avancées de commande testées en simulations offrent des résultats encourageantsqui suggèrent que le système de positionnement devrait dépasser du radiotélescope.IV

Acknowledgements1 am sincerely grateful for the assistance and guidance 1 have received from numeroussources, which have helped make my studies a pleasure instead of a chore. This was mysecond tour with my supervisor Meyer Nahon and 1 again would like to thank him for ajob well done. He has a knack for supervising with a fortuitous balance of guidance andfreedom, while also offering encouragement and criticism in appropriate doses.1 am fortunate to have collaborated with the fine folks at DRAO in Penticton, B.C.,and 1 would like to personally thank Dean Chalmers, Richard Hellyer and Peter Dewdneyfor their tireless dedication to the LAR project. 1 appreciate their efforts, often starting at3:00 AM, to collect the ftight data 1 relied upon for my research.During my research it often felt like 1 was walking around with my head in the c1ouds,but 1 am grateful that 1 was not alone. 1 would like to acknowledge those who playedwith balloons alongside me inc1uding Alistair Howard, Philippe Coulombe-Pontbriand,Jonathan Miller, Francois Deschenes, Gabriel Meunier, Samuel Bouchard, Alexandre Boyer,Etienne Frenette, Dr. Hamid Taghirad, Dr. Wen Bo, Dr. Benoit Boulet and Dr. ClementGosselin. Special thanks to Dr. Gabriele Gilardi who was always willing to help.1 have only to look around my office to find several others who 1 would like to thankfor their scholastic companionship. Thanks to Melita Hadzagic, Alessio Salerno, ShawnArseneau, and Dany Dionne for inspiration, ideas, laughs and chocolate.1 would like to thank my Mc Gill colleagues inc1uding Aki Sato, Aaron Saunders,Chad Abbey, Brian Wong, Philippe Cardou, James Smith, Jason Evans, Shivakumar Ranganathan, Yuwen Li, Hiba Agha, and Lianzhen Luo. Special thanks to Stéphane Caro andMathias Legrand for all their help. 1 would also like to thank Jan Binder, Cynthia Davidson,Donna Morgan, and Joyce Nault for their good moods and administrative assistance.1 thank NSERC, FQRNT, Mc Gill University, and Professor Nahon for providing financial support during my studies and the NRC and NSERC for supporting the LAR project.Finally, 1 would like to thank my family members-Mom, Dad, Rachel, Russel, Bridget,Chris, Haillee, and Parker-for sending their support and friendship across this vast country.v

ContentsAbstractiiiRésuméAcknowledgementsvList of FiguresxList of Tables1xvi1Introduction1.1Lighter-than-Air .11.2Multi-tethered Aerostat System.11.3Large Adaptive Refiector21.4Related Research.51.4.1Tethered Aerostats51.4.2Multi-Tethered Aerostats .61.4.3Cable Manipulators . . .81.4.4Variable Length Tethers91.4.5Tether Control101.5Primary Contributions101.6Thesis Outline . . . . .111.6.1Experimental System .121.6.2Dynamics Model . . .12VI

2.1.6.3Passive Response of System1.6.4Controlled Response of System1.6.5Advanced Control.121313Experimental System142.114General Features of the LAR2.2 Scaling Factor .162.3Aerostat182.4 Tethers.202.5Instrumentation222.6 Winches . . . .242.7Control System262.8Testing Pro gram .282.8.1Flights Summary282.8.2Flight Procedures .29313 Dynamics Model3.131Model Overview32 .3.2 Tether Model .3.2.13.3Tether Damping Coefficient35.383.3.1Equations of Motion393.3.2Aerostat Aerodynamics .403.3.3Added Mass and Added Moment of Inertia443.3.4Summary of Physical Parameters45Aerostat Model3.4 Instrument Platform .453.5Control System463.6 WindModel .483.750Numerical Solutionvii

3.843.8.152Linear Results544.1Flight Tests . . . . . .544.1.1Passive Flight Results564.1.2Motion Reduction564.1.3Varying Geometry584.1.4Summary of Passive Flight Tests .594.3650Passive Response of System4.25Linear Model . . . . .Validation of Passive Response .604.2.1Tether Model Only604.2.2Complete Model63Design Application . . .704.3.1Aerostat Comparison .714.3.2Spherical Aerostat Aerodynamics724.3.3Results . . . . . . . . . . . . . .74Controlled Response of System795.1Controlled Flight Tests .795.2Controlled Results and Model Validation.805.2.1Open-Ioop815.2.2Closed-Loop Test Results835.2.3Closed-Ioop Validation88. .5.3Modal Analysis .965.4Design Application98Advanced Control1026.1Introduction.1026.2Linear Model1036.2.1106Incorporating Winch DynamicsV111

6.36.46.56.67Optimal Control . . . .1106.3.1LQR Stability .1126.3.2LQR Results1136.3.3State Estimation1166.3.4Feedback Frequency and Discrete-Time Control1186.3.5Measurement Noise126Feedforward Control.1376.4.1Results with Optimal Control1406.4.2Results with PID Control . . .144Covering the Operational Workspace .1476.5.1Extreme Zenith Angle1476.5.2Non-Local Controller .149152Implementation Issues / OptionsConclusions1537.1157Recommendations for Further ResearchBibliography159IX

List of Figures1.1Existing large radio telescopes: a) the Very Large Array (VLA) in NewMexico, b) the Arecibo telescope in Puerto Rico, c) the Oreenbank telescope in West Virginia. . . . . . . . . . . . . . . . . . . . . . . . . . .1.23LAR radio telescope concept: a) an artist's rendition including details of areflector panel and the receiver, b) general schematic displaying the maintelescope features. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42.1a) Layout of LAR multi-tethered positioning system: b) overhead view.152.2Layout of the One-Third Scale Experimental System.172.3Outdoor test site at DRAO south of Penticton, B.C. .182.4The aerostat BOB: a) during move from hangar to launch site, b) in profileduring early launching stage, c) in flight looking up a support tether. . .2.55-mm diameter Plasma tether next to 12-mm aerostat leash which has interior con duc tors and a protective sheath.2.619.21The instrument platform a) under-side with array of sens ors and ND modules b) top-side with ultrasonic wind sensor and OPS receiver.232.7Mechanical components of winch system. . . . . . . . . . . .242.8Winch a) frame during assembly b) installed in the field with level-wind2.9and fair-lead accessories. . . . . . . . . . . . . . . . . . . . . . .26Architecture for PC-based measurement and motor control system.272.10 a) Block diagram of overall control system with 10-Hz position feedbackb) system plant with 500-Hz internaI motor current and velocity feedback.x28

3.1Reference frames for dynamics model. . .323.2Lumped-mass visco-elastic tether model. .333.3Experimental3.4Comparison of oscillatory damping tests for tether lengths of 9 m on theset upfor damping estimation .36left and 19 m on the right. . . . . . . . .373.5Aerodynamics forces acting on aerostat.413.62-D representation of tether geometry the payload's CUITent position anddesired position. . . . . . . . . . . . . . . . .473.7Comparison of the turbulent intensity profiles.493.8Comparison of simulated linear and nonlinear zero-input response of platform position. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .524.1a) Layout of LAR multi-tethered positioning system, b) overhead view.554.2Comparison of the motion of the aerostat (le ft axis) and the instrumentplatform (right axis) about their mean positions during Flight #4. . . . . .4.3Statistical coherence between aerostat position and platform position forFlight #4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .aerost t4.4LAR system with4.5Measured and simulated platform position to validate the tether model forremoved and replaced by leash force.Flight #3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.6626466Comparison of experimental and simulated results for aerostat position andorientation for Flight #3. . . . . . . . . . . . . . . . . . . . . . . . . .4.961Comparison of experimental and simulated results of platform position andwind speed for Flight #3. . . . . . . . . . . . . . . . . . . . . . . . . .4.858The power spectral density for the measured and simulated horizontal turbulence, along and transverse to the wind direction, for Flights #1 and #3.4.75767Comparison of experimental and simulated results of tether tension forFlight #3. . . . . . . . . . . . .68Xl

4.10 Power spectral density of simulation and experimental results; a) leash tension b) horizontal and vertical position (results are also included from Section 4.2.1 for the tether model). . . . . . . . . . . . . . . . . . . . . . .694.11 Simulation results for platform (left) and aerostat (right) for a spherical andstreamlined aerostat for test case #1. . . . . . . . . . . . . . . . . . . .744.12 Simulation results for tether tension for a spherical and streamlined aerostatfor test case #1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5.1Bode plot comparison of simulated and experimental payload position inresponse to input at winch #1 with Oze 0 . . . . . . . . . . . . . . . .5.27582Experimental results for platform position during May 6 test with and without PID control at Oze 0 . . . . . . . . . . . . . . . . . . . . . . . . .845.3Power spectral density of platform position during May 6 test at Oze 0 .855.4Platform position during instability on March 22 with K p 0.75865.5Power spectral density of platform position during May 27 test at Oze 50 . 875.6Comparison of experimental and simulated results for platform position forçl. . .trial A, a) time history, b) PSD. . . . . . . . . . . . . . . . . . . . . . .5.7Comparison of experimental and simulated results for platform position fortrial B, a) time history, b) PSD. . . . . . . . . . . . . . . . . . . . . . .5.891Comparison of experimental and simulated results for platform position fortrial C, a) time history, b) PSD. . . . . . . . . . . . . . . . . . . . . . .5.99092Comparison of experimental and simulated results for platform position fortrial D, a) time history, b) PSD. . . . . . . . . . . . . . . . . . . . . . .935.10 Comparison of simulated position error for system at Oze 60 with 3, 4,5, and 6 tethers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1005.11 Comparison of simulated position error for case at Oze 60 with 6 tethersand a base radius, RB 240,320, and 400 m. . . . . . . . . . . . . . .XlI101

6.1Comparison of simulation results with and without control to show thecontroller effects on the aerostat dynamics. Plots are given for the verticalplatform and aerostat position as weIl as leash tension.6.2.105Winch velocity step response-input, measured response and second-orderapproximation. . . . . . . . . . . . . . . . . . . . . . . . . . .1076.3Transfer functions from desired velocity input to position input.1086.4Concatenation of two state-space plants, one for the winch dynamics andone for the tethered system. . . . . . . . . . . . . . . . .1096.5Leash disturbance measured during May 27, 2005 flight.1106.6Comparison of nonlinear and linear simulation results for platform positionwith no control. . . . . . . . . . . . . . . . . . . . .1116.7Block diagram of full-state feedback control system.1126.8Comparison of nonlinear and linear simulation results for platform positionwith an LQR controller. . . . . . . . . . . . . . . . . . . . . . . . . . .6.9114Comparison of linear simulation results with an LQR controller based ondifferent weighting matrices: a) platform position, b) winch velocity during2-second interval. . . . . . . . . . . . . . . . . . . . . . . . . . . . .6.10 LQG regulator with Kalman filter state estimation and LQR feedback.1151176.11 Comparison of simulation results with an LQR controller with full statefeedback and an LQG controller with estimated states.6.12 Block diagram of discrete-time LQG controller and continuous plan

iii . Afin d'améliorer l'algorithme de commande primaire testé sur le terrain, plusieurs stratégies de commande optimale ont été étudiées, nécessitant entre autres un modèle dynamique validé de notre système. Les contrôleurs optimaux, qui sont développés et simulés en

Related Documents:

Business Ready Enhancement Plan for Microsoft Dynamics Customer FAQ Updated January 2011 The Business Ready Enhancement Plan for Microsoft Dynamics is a maintenance plan available to customers of Microsoft Dynamics AX, Microsoft C5, Microsoft Dynamics CRM, Microsoft Dynamics GP, Microsoft Dynamics NAV, Microsoft Dynamics SL, Microsoft Dynamics POS, and Microsoft Dynamics RMS, and

Process Modeling For control applications: Modeling objectives is to describe process dynamics based on the laws of conservation of mass, energy and momentum The balance equation 1.Mass Balance 2.Energy Balance 3.Momentum Balance (Newton’s Law) Rate of Accumulation of fundamental quantity Flow In Flow Out Rate of Production - File Size: 1MBPage Count: 32Explore furtherPDF Download Process Dynamics Modeling And Control Freewww.nwcbooks.comProcess Dynamics and Control - PDF Free Downloadepdf.pubUnderstanding Process Dynamics And Control PDF Download .www.fuadherbal.netA Short Introduction to Process Dynamics and Controlwww.users.abo.fiProcess Dynamics And Control Lecture Notesrims.ruforum.orgRecommended to you b

Microsoft Dynamics 365 for Operations on-premises, Microsoft Dynamics NAV, Microsoft Dynamics GP, Microsoft Dynamics SL, Microsoft Dynamics AX 2012 or prior versions, or Microsoft Dynamics CRM 2016 or prior versions. This guide is not intended to influence the choice of Microsoft Dynamics products and services or provide technical specification.

Microsoft Dynamics Guide Dynamics GP vs. Dynamics 365 (NAV) vs. Dynamics SL . Dynamics 365 BC -1 Premium User 100/month/user (Subscription) 2000 (On-Premise) . Solomon's application became Dynamics SL. Dynamics SL is geared first and foremost for project-based businesses. This makes SL the

Microsoft Dynamics 365 for Operations on-premises, Microsoft Dynamics NAV, Microsoft Dynamics GP, Microsoft Dynamics SL, Microsoft Dynamics AX 2012 or prior versions, or Microsoft Dynamics CRM 2016 or prior versions. This guide is not intended to influence the choice of Microsoft Dynamics products and services or provide technical specification.

This guide is designed to improve your understanding of how to license Microsoft Dynamics 365, Business edition. This document does not apply to Dynamics 365, Enterprise edition, Microsoft Dynamics NAV, Microsoft Dynamics GP, Microsoft Dynamics SL, Microsoft Dynamics AX 2012, or Microsoft Dynamics CRM 2016 or any other prior version.

1 PLLP stands for Microsoft Dynamics Partner Localization and Translation Licensing Program (PLLP). This program authorizes partners around the globe to localize and/or translate Microsoft Dynamics AX, Microsoft Dynamics GP, Microsoft Dynamics NAV and Microsoft Dynamics SL in countries where licenses to these Microsoft Dynamics ERP products are .

Microsoft Dynamics 365 is the next generation of intelligent business applications . on Dynamics NAV, it is cloud only, optimised for 10 - 250 employees . implementing and supporting Dynamics CRM, Dynamics AX and Dynamics 365. Discover our Dynamics 365 services Phone: 01467 623 900 .