You are on page 1of 369

Retrospective Theses and Dissertations

1994

Design of a wing section in ground effect:
application to high speed ground transportation
Christoph Hiemcke
Iowa State University

Follow this and additional works at: http://lib.dr.iastate.edu/rtd
Part of the Aerospace Engineering Commons, and the Mechanical Engineering Commons

Recommended Citation
Hiemcke, Christoph, "Design of a wing section in ground effect: application to high speed ground transportation " (1994).
Retrospective Theses and Dissertations. 10478.
http://lib.dr.iastate.edu/rtd/10478

This Dissertation is brought to you for free and open access by Iowa State University Digital Repository. It has been accepted for inclusion in
Retrospective Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please
contact digirep@iastate.edu.

INFORMATION TO USERS

This manuscript has been reproduced from the microfihn master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may
be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in
reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9" black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly
to order.

University Microfilms International
A Bell & Howell Infornnation Company
300 North Zeeb Road. Ann Arbor. Ml 48106-1346 USA
313/761-4700 800/521-0600

Order Nuiaber 950S563

Design of wing section in ground effect: Application to
high-speed ground transportation

Hiemcke, Christoph, Ph.D.
Iowa State University, 1994

UM-I
SOON.ZeebRd.
Ann Arbor, MI 48106

Design of a wing section in ground effect:

Application to high speed ground transportation

by

Christoph Hiemcke

A Dissertation Submitted to the

Graduate Faculty in Partial Fulfillment of the

Requirements for the Degree of

DOCTOR OF PHILOSOPHY
Department: Aerospace Engineering and Engineering Mechanics
Major: Aerospace Engineering

Signature was redacted for privacy.

Signature was redacted for privacy.
For the Major Department
Signature was redacted for privacy.
For fh^^raduate College

Iowa State University
Ames, Iowa
1994

ii

DEDICATION

For the many teachers and friends

who have been role models and inspiration

to me over the years

Prima di essere You are human

ingegnere voi before being

siete uomini engineers

[Francesco de Sanctis]

geography). Sabine Handschuh. Els Hiemcke. Carola Baumbach. Herr Nieskewitsch (German). Beppie. Herr VVetscherek (mathematics). Karsten Jedamski. Herr Schulz (English. Kai Lehmann. Tibor. Theodor Wilhelm Johann Hiemcke (DePa). Koos Kool. and places along the way: (please forgive misspellings and omissions) Early years: My parents: Anke Berta Hiemcke (Mami. my teach­ ers: Herr Kruger (French). Urgrossmutter). Birgit Hiemcke. Frau Schroder (music). Herr (Friedcl) und Frau Albersdorfer. vieja). Petra Ickels. Frau Harzheim (Englisch). and Annette von ?. Tina Fonseca. and my country: Germany. Herr Thiess (mathematics. Lena and Christoph II. Herr Bock . Heide. Matthias Fiene-Vizy. DeMa). Germany. Beata Annegret Vizy. and Biisum. my grand­ parents: Else Catharina Marie Riihmarin (geborene Schuett. Herr Fiedler (Biology). Betsy Cornelia Hiemcke (geborene Tims. Thomas and Michael Karbe. Hermann Hiemcke (Dad. Teenage years: My school and my surroundings: the towns of Moorrege and Uetersen. the Schmidt-Wiborg's. my school and my surroundings: the towns of Rellingen. my friends and classmates: Anja Schultz. Frau Hagemann (elementary school teacher). other relatives: Theo &. Plattdeutsch). Ramon Marcel Hiemcke. Gesa Melina. Sharon Kool-Hiemcke. Bertha Katharina Schuett (geborene Claussen. my sisters and brother and their families: Ingrid Vizy. teachers. Omi). Otto. lii Friends. Kraffczyk. der Alte. Sabine Holz. Michael &: Juttka Hidy. my French friends Catherine Rabouel and Nadia LeBerre. Gerd and Dirk Sohngen. Tangstedt. viejo). Rebecca. Dr.

the Barani's (Antonio. my friends. Sheila Marco (English. the Asuaga's (especially Louisa and Christina). Felipe Gon­ zalez. especially families Schrijk. Lahni-Jean. Iowa. Kai Wohlers. Ackie. Bradley (principal). Mrs. Jean-Jaccjues. Casey and the other cats. Charles Durlacher. Cindy and Paul Maynes. Spain: My school and my surroundings: my parents and Sharon. and Hoogland. Fiona Rodda. de Alonso (math­ ematics). Jane Barbadillo (principal). . my schoolmates and friends: Tiina Ruuskanen. my host town: Lenox. our zebra finch. Diane Pat­ terson. Year as foreign exchange student in Lenox. Dan Carruthers. Berk. Lisa. IV (sports). the Berlin friends: Joachim. Christine Baker. Bianca and Joep. da Costa. Cati and Lori. Mayka Carmen Mate Perez. Pauline Miller (speech and drama). Rogers (physiology). Tracey Gridley. Dee Ann Christensen. the grandparents Brannon and Barker. teachers. Lonnie Barker. literature). Lisa Round and family. Mrs. Jose Dobon. the state of Andalucia. Mac Hiipferfink. my teachers: Mrs. Mrs. the city of Malaga. Diane Robinson. Ron Bonnett. and Sybille. Kelly Dukes. and acciuaintances: Sue and Jean Fuersteiiau. iny Polish co-worker in the rose fields. van der Keijl. Bonnie Schofield. Yolanda ?. Ingrid and SvenjaBary. la Sefiora Josefa. "Interpares" and its senior inhabitants. Dillenburg (mathematics). the Lake Wobegon of Iowa. my host country: the United States of America. Mrs. Two years in Malaga. Benji. Sari Pajunen. Sunny View (Vista Sol) High School in Torremolinos. Mr. Paul Rainforth and family. and Alex). USA: My host family and their families: Billie-Jean Barker. Todd. Kevin and Millie. Mrs. Henryk Martynowski.

my friends: Elisabeth Waterman Archer ("Liffy") and her family. ISU Center. Antonio Sabariz. Chantal Hoareau. Laura. Catherine Morel. Lorenzo . Joan Grzywa. Sushil and Latha Bhavnani. James Iversen (aerodynamics). gas dy­ namics). Matthew and Rebecca Chackalackal. at Iowa State University: My friends: Audra (my wife) and family (Ken. design). especially my friends Martha Rayner and Jeff Launius. Balsy Kasi. P. Alumni Hall. Andrew de Souza: my teachers: David Robinson (physics. my co-workers at the M-Shop. Zori.J. Ravi. Bill James (propulsion. Alyssa). Sris. M-Shop. John Daniel. Graduate Studies. Kaveh. Pam Goomas. Rich Jorgensen (flying). Omana and Jos Putherikal. adviser) and Professor Breyman (his­ tory). T. J. V Two years at Drake University. the lol)byrats.S. Prakash Mohan. Ahmed Sarhan. Hermann. Mike and Noelle Martin. Guna. Sylvan Senior. Kara. M. Minie and Vikram Yadama. Ganesh Rajagopalan (aerodynamics). Rao Vadali (controls). Steve Hooper (structures. Jerry Vogel (design). my teachers: Professor Hsu. K. Kumaran Kalyana- sundaram. Philip and Elizabeth (Lisa) Theruvakattil. my friends Kurt Lynn. Armando Exposito. Marge. Pushpa Verma. Stacy Maurer. Iowa: My roommate and baritone Matthew Eisenberg. Babak. Trevor Lettering. Bashar Jano. McDaniel (structures). Kleanthis Lambrinides (Leo). L. Professor Apt (French history). places: Friley food service. Magdy Hindawe (structures). The undergraduate years at Iowa State University: My roommates at Alumni Hall. Seversike (astrodynamics). Manoj Dan- dekar. Achilles Avraamides (ancient history). Juan. Des Moines. Alex Dameus.

Gapa. Lidia. Scot Kelchner. Inke. the Courteau's (Dick. Sam Sudam. Mark Summerson (employment counseling). Arne and Mitzku Gaedtkens. Jason and Kelly Gillette. Helmut Maier. Tony Smith (philosophy) and family. Mava. Richard Johnson. Bion Pierson (controls. Gertrud Lippisch. at Iowa State University: My friends: Charles Gratto. VI Maggioni.D. Graduate Student Senate. Wheatsfield . Thelma. my adviser) and family. Kevin and Janet Renze.Jatzek. optimization). Mark and Sara Eversden. Ph. Danielle and Robert Donet- Vincent. Graduate Studies.Joanna Courteau. Pete Barrance. places and institutions: Friley food service (the table of foreigners). Tiffany. Coleen. Habitat for Humanity. vehicle dynamics. Jerry Lamsa. Sam Wormley. Marcelo Pinto. Carl Bleyle (music history). Ali Tiirel (planning in LDCs). Bob Hollinger. David Roochnik (philosophy) and family. and Amabo). Ursula von Godanyi. JefF Huston and family. Kathy Lynott. Blanca Bailey/Lippisch. and Roger). Marek. Renate Vogelsang. my teachers: Alric Rothmayer (theoretical aerodynamics. Bob Hollinger (philosophy) and family. George Work (music). secretaries: Susan. . David Wilson (history of sci­ ence). adviser) and family (Patricia. Henny. Beata. . Professor McClain (art history).Ann Stanley. places and institutions: city of Ames. Dani Prinz. De^nnis and Ronnie Lindemann. Mark Mumford (ar­ chitecture history). Flash. Nancy Eaton (library). my teach­ ers: Jeff Huston (dynamics. Ulla von Dorsten. Stanley Yates (library). Professor Kostelnick (literature). Su- sanne Pollmann. Darcy. Dorota. Sharon Anderson (pre-marital counseling). Sarah. Bernie and Clarice Gerstein. Tankut Atan. Marcia and Ahmet: my students in aerospace structures courses and lab­ oratories. Bill Jensen (technician). Gloria. Sonya Rodolfo-Sioson. Penny. Donald Young (biofluid dynamics). Dee. .

VII Co-Op. Cafe Lovish (Crubsteak). the Farmer's Market. . Bali House. Lee's Eggroll. Cheese'ii Puppets. Little Taipei. the ISU Outdoor Recreation Center. the Varsity Theatre in Des Moines. Student Union Board Films. Cafe Baudelaire.

viii TABLE OF CONTENTS DEDICATION ii KEYWORDS RELATED TO THE DISSERTATION TOPIC xxi GROUND EFFECT MACHINES: COMMON ABBREVIATIONS . THE AERODYNAMIC GROUND EFFECT AND ASSOCIATED FLOW REGIMES 6 Wings Near the Ground or Near Water 9 Weak (classical) ground effect regime 10 Moderate (intermediate) ground effect regime. A SURVEY OF VEHICLES FOR OPERATION IN VARIOUS GROUND EFFECT REGIMES 18 . INTRODUCTION 1 CHAPTER 2. xxiv NOMENCLATURE xxix ACKNOWLEDGEMENTS xxxiii CHAPTER 1. ram-wing regime 14 Extreme ground effect regime 17 Lubrication Regime 17 CHAPTER 3. WIG regime 12 Small-Gap Regime 14 Strong (close) ground effect regime.

. or duorail tech­ nologies) 49 Monorails and other unconventional non-levitating trains 55 Tracked air-cushion vehicles (TACVs) 57 Magnetic Levitation (Maglev) Trains and Linear Induction Motors (LIMs) 64 Technical introduction to maglev and linear induction motors . ix The Categorization Of Ground-Effect Vehicles 18 Wing-In-Ground-EfFect (WIG) Vehicles (Ekranoplans) 19 Ram-Wing Vehicles (Aeroskimmers) 21 Planing. 34 Introduction 34 A Sketch Of Train History 39 Three technological breakthroughs 39 Steam locomotives and vacuum technology 41 Electric propulsion 43 Diesel propulsion 44 Diesel-electric propulsion 44 Post-war population pressure and energy crises 44 Discussion Of The Various Technologies 48 Conventional high-speed rail trains (steel wheel-on-rail. Air-Lubricated Vehicles (Hydroskimmers) 25 Power-Augmented Ram-Wing (PAR) 26 PAR/WIG Vehicles 27 Planing/Ram/WIG vehicles (Aerofoil Boats. . 64 A brief history of maglev technology 67 . A SURVEY OF UNCONVENTIONAL TRAINS . Winged Hulls) 29 CHAPTER 4...

A SURVEY OF THEORIES ASSOCIATED WITH THE DIFFERENT GROUND EFFECT REGIMES 106 Two-Dimensional Mapping Techniques (Analytical) IDS Two-Dimensional Thin Airfoil Theory (Analytical) 109 Three-Dimensional Finite Wing Theory: Lifting Lines And Lifting Surfaces (Analytical) 112 Three-dimensional simple (classical... . 101 A Note on the Aerodynamics of Bluff Bodies (Cars) 103 A Note on the Aerodynamics of Slender Bodies (Trains) 104 CHAPTER 6.. linear) lifting line theory (analytical) 112 Three-dimensional extended (linear) lifting line theory (analytical) . A SURVEY AND CRITIQUE OF TECHNIQUES USED IN GROUND EFFECT EXPERIMENTS 90 Mirror-Image Method 90 Fixed Ground Method 94 Moving-Belt Technique 97 Moving-Model Technique 98 Rheoelectric Analog Method 101 A Comparison of Experimental Results for Wings Near the Ground .. 114 Three-dimensional numerical lifting line method 114 Three-dimensional lifting surface theory (analytical) 115 . X Maglev versus high-speed duorail trains 83 Trains using aerodynamic suspension 86 CHAPTER 5.. 114 Three-dimensional double (linear) lifting line theory (analytical) .

THE STUDY OF WING SECTIONS IN GROUND EFFECT 1. A PANEL METHOD FOR PRELIMINARY WING SECTION DESIGN 132 Motivation for an Analysis Using an Inviscid 2-D Panel Method 132 .30 CHAPTER 8.and Three-Dimensional Numerical Methods for Potential Flow: Gen­ eral Comments 117 Two-Dimensional Panel Methods (Numerical) 118 Two-dimensional source panel method (numerical) 119 Two-dimensional vortex panel method (numerical) 120 Two-dimensional Douglas-Neumann panel method (numerical) .. xi Two... numerical) 124 Three-Dimensional Wings in Close Ground Effect: Early Methods (Analyt­ ical) 125 Three-Dimensional Wings In Close Ground Effect: Asymptotic Linear 1-D Channel Flow Theory (Analytical) 125 Three-Dimensional Wings in Close Ground Effect: Asymptotic Linear 2-D Channel Flow Theory 127 Three-Dimensional Wings In Close Ground Effect: Asymptotic Nonlinear 2-D Channel Flow Theory 128 Three-Dimensional Wings in Extreme Ground Effect: The Growing Influ­ ence of Viscosity 128 CHAPTER 7. 121 Other two-dimensional panel methods (numerical) 122 Three-dimensional panel methods (numerical) 122 Three-dimensional vortex lattice method (VLM.

199 Fabrication of the Pitot Rake 203 CHAPTER 10.. xii Nonclimensionalization 133 Coordinate Systems 135 Derivation of a 2-D Vortex Panel Method 139 Extension of the Panel Method to Ground Effect 150 Discrete Vortex Code 154 Vortex Panel Method 156 Constant-Strength Vortex Panel Method 161 Internal Boundary Conditions and Other Difficulties 169 Constant Strength Vortex Panel Code ..Initial Results 175 Preliminary Design of an Airfoil in Ground Effect 179 CHAPTER 9. CONSTRUCTION OF THE EXPERIMENTAL WIND TUNNEL MODEL 187 Introduction and Initial Ideas 187 Description of the Two Prismatic Wings 191 Fabrication of the Two Wings 193 Fabrication of the Wing Mounting Plug and Traversing Mechanism . EXPERIMENTAL PROCEDURE 208 Description of the Wind Tunnel 208 Data Acquisition Equipment and Instrumentation 212 Calibration Procedure 213 Experimental Procedure 215 Problems with the Experimental Procedure 221 Smoke Visualization 222 ..

CONCLUSIONS AND FUTURE DIRECTIONS . mission The elevated guideway Takeoff. A RAM-AIR TRAIN PROTOTYPE: FEASIBIL­ ITY STUDY 281 Origins of the Ram-Air Train The Concept of the Ram-Air Train Expected performance and possible uses of the ram-air train Prototype design: sizing. RESULTS AND DISCUSSION Introduction Presentation of Numerical Results Presentation of Experimental Results 236 Experimental pitching moment and pressure drag results 264 Discussion of the Ground Effect Phenomenon 275 Suggested Improvements for Experimental Work 277 Suggested Improvements for Computational Work 278 CHAPTER 12. GLOSSARY OF TERMS USED IN THE DISSERTATION BIBLIOGRAPHY 301 . braking and landing The Future of the Ram-Air Train CHAPTER 13. Xlll Pitot Rake Experiments CHAPTER 11. .

84 Table 11.1: NACA 5312... a = 6®. xiv LIST OF TABLES Table 4..1: Milestones in railroad history 39 Table 4. experimental data 265 .2: Energy use and noise levels of transportation systems .

or aerodynamic levitation (AeroLev) train 5 Figure 2.8: Short-stator EMS train with LIM propulsion 69 Figure 4.1: Proposed ram-air train..9: Long-stator EMS with LIM propulsion.6: French Aerotrain 59 Figure 4.1: Some aerospace nomenclature 7 Figure 2.4: French high-speed train (TGV) 53 Figure 4.2: The flow regimes of the aerodynamic ground effect 8 Figure 2.7: Combinations of magnetic levitation and linear motors . 68 Figure 4.3: Wing showing the wingtip vortex 11 Figure 2.. similar to German Transrapid TR-07 70 .2: Trainsets and bogie types 38 Figure 4.1: Categories of ground effect vehicles 20 Figure 4.5: The Piasecki Airtrain 56 Figure 4.4: Profile in ground effect: image method 13 Figure 3. XV LIST OF FIGURES Figure 1.1: Railroad nomenclature 36 Figure 4.3: Categories of trains and tracked ACVs 50 Figure 4..

142 Figure 8. 107 Figure 8.10: Panel-centered geometry 158 Figure 8. similar to Japanese MLU-001 71 Figure 4.3: Rotation of the coordinate system 138 Figure 8.7: Leading edge moment due to panel i 149 Figure 8.1: Theories associated with the Ground Effect Flow Regimes .4: Geometry of panels i and j on the profile 141 Figure 8. . xvi Figure 4.9: Kutta point and numbering system for the image profile .12: German Transrapid TR-07 81 Figure 4.10: Short-stator EDS with LIM propulsion.3: Moving-belt method 98 Figure 5.4: Moving-model method 100 Figure 6.11: Types of Switches 79 Figure 4. .12: Approaching the midpoint from the bottom 166 Figure 8..6: Definition of a wake panel 146 Figure 8.1: Mirror-image method 92 Figure 5.13: Global and local coordinates 168 .8: Coordinate system for the image profile 151 Figure 8.13: Terrawing train 88 Figure 5.11: Approaching the midpoint from the top 165 Figure 8. .1: Dimensional geometry of a profile in ground effect 133 Figure 8.2: Fixed ground-plate method 95 Figure 5. 153 Figure 8.2: Coordinate systems 136 Figure 8.5: Airfoil surface coordinates and panel numbering system ..

.1 . . xvii Pressure distribution for a NACA 4412 177 Pressure distribution for a NACA 4412 178 Design code user interface 180 Effect of varying the maximum camber location 181 Effect of varying the camber 182 Effect of varying the thickness 183 NACA 5312 airfoil shape 185 Pressure distribution for NACA 5312. and data acquisition system . 186 Self-propelled tracked test rig 189 Test rig pushed by an automobile 190 Top (main) wing during assembly 192 CNC mill with wooden wing section 195 The two aluminum center sections during construction . lift versus ground clearance 230 . 209 Front view of model mounted inside the wind tunnel 211 Scanivalve and calibration transducer 213 Diagram of the data acquisition system 214 Aerodynamic resultants acting on panel i 219 NACA 5312.5. = 0. = 1. . . smoke injector. 196 Outside of the model mounting plug 200 Traversing mechanism with slider blocks 201 Model and Pitot rake in the wind tunnel test section 206 Wind tunnel. a — 6^.

1.13: NACA 5312.3: NACA 5312. ^ = 0. a = O''.4: NACA 5312. panel code results.8: NACA 5312. experimental results. isolated airfoil. a = OP^ pressure distributions for various ground clearances 232 Figure 11. a = 6®. pressure distribu­ tions for various incidences 233 Figure 11. experimental pressure distributions for various ground clearances 252 Figure 11.14: NACA 5312. isolated airfoil.6: NACA 5312.11: Smoke visualization image showing flow symmetry 250 Figure 11. q = 6°. match of experimental data from top and bottom wings 249 Figure 11. a = —4''. a = 6®. xviii Figure 11. panel code results. 251 Figure 11. lift curve 246 Figure 11. ^ = 0.1. lift curve with and without ground effect 247 Figure 11. isolated airfoil. panel code results. .9: NACA 5312.12: Smoke visualization showing stagnated underbody flow . experimental pressure distributions for various ground clearances 253 .7: NACA 5312. effect of Reynolds num­ ber on pressure distribution 248 Figure 11. lift curve with and without ground effect 231 Figure 11. panel code. ^ = 0.5: NACA 5312.1. pressure distributions for various incidences 234 Figure 11.2: NACA 5312. effect of the number of panels on panel code results 235 Figure 11.10: NACA 5312. .

a = 6®.25: Three equivalent loadings 266 Figure 11.18: NACA 5312. ^ = 0. a = 12®.19: NACA 5312.1.17: NACA 5312. comparison of experimental and nu­ merical results for three inclinations 262 Figure 11.21: NACA 5312. experimental pressure distributions for various ground clearances 254 Figure 11.1. experimental pressure distributions for various ground clearances 255 Figure 11.26: NACA 5312.20: NACA 5312.1. experimental pressure distributions for various ground clearances 256 Figure 11. a = 6®. experimental pressure distributions for various incidences 259 Figure 11. ^ = 0.1. effect of Reynolds number on the pressure distribution 260 Figure 11. matching experimental and numerical results 263 Figure 11.23: NACA 5312. ^ = 0. ^ = 0.1. comparison of scanivalve and manometer results 261 Figure 11.22: NACA 5312. a = 4". experimental pressure distributions for various ground clearances 258 Figure 11. xix Figure 11.15: NACA 5312. a = 8^. a = 6®. experimental pressure distributions for various ground clearances 257 Figure 11. ^ = 0.24: NACA 5312. a = 16®. experimental leading edge pitching moment ver­ sus incidence 268 . a = 6®.16: NACA 5312.

1: Hovering aerofoil-boat 283 Figure 12. experimental efficiency results 273 Figure 11. a = 6". experimental drag polar 272 Figure 11.2: Early ram-air train 284 Figure 12.33: NACA 5312.28: NACA 5312.30: NACA 5312.3: Wooden model of the AeroLev train 286 . a = 6^. experimental leading edge pitching mo­ ment versus ground clearance 269 Figure 11. a — 6 ^ . experimental pressure drag curve 270 Figure 11.32: NACA 5312.27: NACA 5312. experimental results 271 Figure 11. experimental efficiency results 274 Figure 11.29: NACA 5312. lift contributions from the top and bot­ tom wing surfaces 276 Figure 12.31: NACA 5312. a = G''. XX Figure 11.

xxi KEYWORDS RELATED TO THE DISSERTATION TOPIC advanced ground transport (AGT) aerodynamic ground effect (AGE) aerodynamic levitation aerodynamic suspension aerofoil boat aeroskimmer aerostatic ground effect aerotrain air-bearing vehicles air cushion vehicles (ACV) air levitation air suspension channel flow vehicle compression airplane distributed-suction experimental technique dynamic air cushion vehicle dynamic ground effect dynamic interface vehicle Ekranoplan electrodynamic suspension (EDS) maglev electromagnetic suspension (EMS) maglev flat-plate experimental technique flying fish gap flow ground-board experimental technique ground cushion ground effect machine (GEM) ground effect take-off and landing aircraft (GETOL) .

xxn ground effect vehicle (GEV) ground-plate experimental technique ground pressure vehicles ground proximity ground representation ground simulation ground transportation ground vicinity- high-speed ground transportation (HSGT) hovercraft hovertrain hydroskimmer image-wing technique (experimental and theoretical) Low Boy lubrication theory magnetis levitation (maglev) trains mirror-wing technique (experimental and theoretical) moving-belt experimental technique moving-model experimental technique plajiing boat railroads ram-air cushion ram-wing reflection method technique (experimental and theoretical) reflection-plate experimental technique road vehicles seabirds near the water slot flow static ground effect surface effect ship (SES) surface effect vehicle surface skimmer terrafoil towed-model experimental technique tow-tank experimental technique tracked air cushion vehicle (TACV) underbody flow undersurface flow wing in ground-effect (WIG) winged hull .

XXUl IN GERMAN: Aerodynamischer Bodeneffekt (aerodynamic ground effect) Bodeneffekt (ground effect) Bodeneffektgeraet (BEG) (ground effect machine) dynamischer Bodeneffekt (dynamic ground effect) Flugflaechen-Boot (aerofoil boat) Gleitboot (planing boat) Luftkissen (air cushion) Luftkissenfcihrzeug (LKF) (air cushion vehicle) Magnetschwebebahn (raaglev train) Schwebefahrzeug (hovering craft) Staufluegel (ram-wing) Staufluegelfahrzeug (ram-wing vehicle) Tragflaechenboot (hydrofoil boat) .

U.Advanced Passenger Train . xxiv GROUND EFFECT MACHINES: COMMON ABBREVIATIONS ABL . AMS . D.Boundary-Layer.S.Bundesministerium fiir Verteidigung (West German ministry of defense).A.Advanced Passenger Train.S.Anti Submarine Warfare. AGE . BMVg . Also the Spanish word for bird.). APT-P .A. New Jersey. AGT . AVE .Alta Velocidad Espanola.Aerodynamic Ground Effect. Aero . APT .Advanced Ground Transport. (U.C. Spanish HSR system consisting of TGVs.Air-Cushion Vehicle.Prototype version. . APT-S . BL .Aerodynamics Laboratory of the David Taylor Model Basin. ASW .Service version.Department of Aerospace and Mechanical Sciences. Washington.Atmospheric Boundary Layer. ACV . Princeton University.Advanced Passenger Train .

(U. CFR .Canadian Institute for Guided Ground Transport. DOT .A.A.British Railways. GEM .).Department Of Transportation (U.A. DOT.S. German national railways.Deutsche Bundesbahn. BRB .und Versuchsanstalt fiir Luft.Captured Air Bubble.Ground Effect Take-Off and Landing.und Raumfahrt.). Research & Development Division. D.).ElectroMagnetic Suspension. .S.Code of Federal Regulations (CFR 49 governs railroad). Washington.British Railways Board.S. EMS .Bureau of Naval Weapons (U.Deutsche Forschungs.). .S.A.Department Of Transportation Office of High-Speed Ground OHSGT Transportation (U. XXV BR . EDS . FRA .Ground Effect. United Kingdom.S.Ground Effect Machine.). DFVLR . CIGGT .Federal Railroad Administration (U. formerly Experimental Model Basin. DB . GE . CAB . R&DD Derby. GETOL . BuWeps . DTMB .A.David W. Taylor Model Basin.ElectroDynamic Suspension.C.

Gravity Vacuum Transit.Linear Induction Motor.A.Intelligent Vehicle-Highway System. IVHS . Cambridge. xxvi GEV . ICE . California (U.National Research Council (USA).Operation & Maintenance. NRC .. ISTEA . . 1990.Intermodal Surface Transportation Efficiency Act.IR .). national railways.MAGnetic LEVitation.Motor Industry Research Association (United Kingdom). LIM .International Track Gauge for railroads. JSSS .Japanese Railways.S. IGE . MIT .Douglas Aircraft.Inter-City Express (German HSR). 1. . 0 M . U. MAGLEV .668 meters.7".In Ground Effect. HSST .S. Long Beach.Ground Effect Vehicle.National Maglev Initiative.A. or 5'5. GVT . HSR . 1991.Massachusetts Institute of Technology.High-Speed Rail. NMI . MIRA .Jane's Surface Skimmers. LB .High Speed Surface Transportation. ITG .

xxvn

OGE - Out of Ground Effect.

OST - Office of the Secretary of Transportation (U.S.A.).

OTA - Office of Technology Assessment, U.S. Congress.

PAR - Power Augmented Ram-wing.

PMS - Permanent Magnet Suspension.

ROW - Right-of-Way (land surrounding railroad tracks).

RTRI - Railway Technical Research Institute of Japan.

SEV - Surface Effect Vehicle.

SES - Surface Effect Ship.

SRL - Seiler Research Laboratory, U.S. Air Force Systems Command.

SS/GTS - Super Speed Ground Transportation System.

SST - Super Speed Train.

TAB - Tandem Aerofoil Boat.

TACRV - Tracked Air Cushion Research Vehicle.

TACV - Tracked Air Cushion Vehicle.

TGV - Train a Grande Vitesse (French HSR).

TLRV - Tracked Levitated Research Vehicle.

TLV - Tracked Levitated Vehicle.

TMLV - Tracked Magnetically Levitated Vehicle.

TR - Transrapid, German maglev train.

xxviii

TRACV Tracked Ram Air Cushion Vehicle.

TRB Transportation Research Board, part of NRC.

TRECOM Transportation Research Command, U.S. Army, Fort Eustis,
Virginia.

TSC (John A. Volpe National) Transportation Systems Center of
the U.S. Department Of Transportation at Cambridge,
Massachusetts.

TWIG Tandem Wing In Ground Effect.

UHSGT Ultra High Speed Ground Transportation.

UIC Union of International Railways (regulatory body).

UMTA Urban Mass Transport Administration, part of USDOT.

USACE United States Army Corps of Engineers.

USDOT United States Department of Transportation.

VIA Canadian national passenger rail corporation.

VNTSC Volpe National Transportation Systems Center.

V/STOL Vertical/Short Take-Off and Landing.

WIG Wing In Ground effect.

WSEV Winged Surface Effect Vehicle.

WTOL Water Take-OfF and Landing.

xxix

NOMENCLATURE

number of counts from the data acquisition system.

chord length, distance between the leading and trailing edges.

lift coefficient, two-dimensional.

drag coefficient, two-dimensional.

moment coefficient, two-dimensional.

pressure coefficient.

distance.

unit vector

height above the ground; in this case the distance between the
trailing edge and the ground vertically beneath it.

unit vector along the global abscissa.

unit vector along the global ordinate.

panel length; two-dimensional lift.

two-dimensional lift.

two-dimensional moment.

XXX

N - number of panels representing one airfoil.

p - pressure.

q - dynamic pressure.

r - geometric scale ratio (size of the model divided by the size of the
original).

r - position vector.

Re - Reynolds number.

S - planform area.

s - panel surface coordinate.

t - thickness of the airfoil.

V - velocity.

X - global cartesian coordinate, abscissa.

X - local cartesian coordinate, abscissa.

Z - global cartesian coordinate, ordinate.

z - local cartesian coordinate, ordinate.

a - incidence, angle between the chord line and the freestream.

r - circulation.

7 - vorticity strength.

AC - number of counts corresponding to a pressure differential.

Ap - pressure differential.

xxxi

9 - panel inclination; also angular position of point Q in the
panel coordinate system.

A - temporary variable for facilitating integration.

^ - dummy panel coordinate, similar to x p .

p - density.

$ - velocity potential.

^ - streamfunction.

Subscripts:

ac - aerodynamic center.

AMB - ambient conditions, outside the wind tunnel,

c - camber

CAL - pertaining to the calibration transducer or procedure.

G - gap region, space between the bottom edge of an endplate and
the ground.

GBL - ground boundary layer.

i - general index.

int - internal; inside the airfoil.

j - general index.

LBL - lower surface boundary layer.

LE - leading edge of the profile, also refers to the region between the

xxxii

wing and the ground in the vicinity of the leading edge.

n - normal.

p - panel, panel-fixed coordinates.

Q - refers to a point Q near the panel.

r - radial.

t - tangential.

TE - trailing edge of the profile, also refers to the region between the
wing and the ground in the vicinity of the trailing edge.

U - underbody region, space between the wing and the ground.

UBL - upper surface boundary layer.

V - due to the vortex distribution.

6 - transverse, circumferential (component of a vector).

oo - freestream condition; test section condition far upstream of
the model; also refers to a quantity that is due to the
influence of the oncoming freestream.

Superscripts:

* - denotes a dimensional quantity.

' - denotes a two-dimensional quantity.

xxxiii

ACKNOWLEDGEMENTS

I thank my adviser, Professor JefF Huston, for taking a sincere interest in me

and in my work. His support was in the form of philosophical discussions, technical

discussions, conceptual questions, probing questions, thought-provoking questions,

encouragement, pressure, critiques, honest opinions, informal comments, joking re­

marks, and thorough reviews of my manuscripts.

I am thankful for his respect and tolerance, and for giving me great freedom in

my work. That freedom allowed me to focus not just on technical problems, but to

develop my sense for ethics, social concerns, history, and creativity.

These comments do not just apply to my research, but also to my work in the ISU

Courseware Development Studio, which is directed by Professor Huston. Under his

guidance I learned to better understand my audience, to write proposals and papers,

and to cope with bureaucracy. But most of all, I developed a better understanding

of the professions of engineering and teaching.

I also thank the other four members of my committee, Dean Nancy L. Eaton,

Dr. James D. Iversen, Dr. Jerald M. Vogel and Dr. David B. Wilson, for their time

and support.

I thank Dean Eaton for consistently acting in the interest of the Alexander

Lippisch Collection at ISU's Parks Library.

I am especially thankful for his guidance leading up to my presentation on Lippisch to the Midwest Junto for the History of Science in Chicago in 1993. xxxiv I thank Dr. I thank Dr. Iversen for initiating rny work on the Lippisch Collection. . and for his continued interest in it. Wilson for his support of my work on the Lippisch Collection and of my interest in history in general. Last but certainly not least I thank my wife Audra for her unceasing support and for taking my mind off my work. I thank Dr. Vogel for his advice on panel methods and for facilitating the construction of the experimental model.

1 CHAPTER 1.but perhaps it will require a combination of resource scarcity and immediate health threats to bring people to reason. They would also vastly reduce pollu­ tion and traffic congestion. INTRODUCTION The delicate balance of the world's ecosystems is seriously threatened by hu­ manity's wasteful ways. It will not be possible for future generations to live from day to day without considering the consequences of their actions . economic losses due to traffic congestion and airport delays in the United States were . But does not personal freedom involve mainly mental rather than physical mobility? Is it "freedom" to continuously worry about the inconvenience and expense of im­ minent mechanical problems? Is it "freedom" to be caught up in an unending series of payments? Is it "freedom" to be subjected to the stress of noisy and unsightly motorways and traffic jams? Efficient mciss transportation systems would provide people with the necessary mobility without the problems of ownership. This is particularly important considering that by 2020 surface travel in the United States will double from the 1988 levels [1]. There are also economic reasons for developing efficient mass transportation systems. One symbol of the excessive consumerism of our time is the automobile. in 1990. Adver­ tisements have taught the public to associate it with personal freedom and mobility.

Mrs. trams. buses. The air cushion is due to the aerodynamic ground effect. 2 US$ 80 million and US$ 5 billion. and by the interest and dedication of the head of the Department . Dr. Delta-wing aircraft have wings with a triangular planform. Despite the fact that most transportation research is carried out in the automobile and aviation industries. including raceboats with water propulsion which skimmed just above the surface of the water and airscrew-driven flying boats that operated optimally in ground effect (see [2] through [17]). aircraft and ships. Gertrud Lippisch. a German aerodynamicist who had become widely known for his delta-wing designs before and during World War II. and in the years which followed. Afterward.e. he worked as a con­ sultant for companies in Germany and the United States. the flow of air past the forward-moving wing. In the 1960s and 1970s he designed a series of ground effect vehicles. gave some of his papers and photographs to the Parks Library at Iowa State University.a train suspended under a lifting wing. Mrs. This dissertation centers on a new concept for a tracked transportation system . Iowa until the mid-1960s. The wings start not far behind the nose and extend all the way to the tail section. Alexander Lippisch. i. it seems advisable to improve surface transportation systems. reaching a maximum span there. The phenomenon of the aerodynamic ground effect has been known and studied for many years. One of the researchers in the field was Dr. his widow. Lippisch was encouraged by the care the material received there. Future generations will once again be forced to rely on trains. This wing operates just above a T-shaped concrete track and rides on a cushion of air. he came to the United States and worked for Collins Radio Company in Cedar Rapids.. respectively. Following the war. Lippisch died in 1976.

who is German and was at the time a graduate student of aerospace engineering.the Ram-Air Train (see Figure 1). as a patrol boat along rivers and coasts. and became interested in the aerodynamic ground effect as a result. For this reason. The fundamental question was whether the wings would generate sufficient lift. The . Iversen in turn contacted the author. Most of the Lippisch material was transported to the Parks Library at Iowa State University in early 1991 [18]. She decided to donate the rest of the materials in 1990 on the condition that someone would come to her home to review the materials under her supervision. he wished to determine whether the aerodynamic ground effect could be utilized in a high speed ground transportation system. the present work centers on the design of a train-like vehicle that has wings . Since the author has a strong interest in public transportation. In the year which followed. The following chapter is an explanation of the aerodynamic ground effect and associated flow regimes. Chapters 3 and 4 are surveys of vehicles that utilize the aerodynamic ground effect and of systems of high speed ground transportation. The author had the honor and pleasure to visit Mrs. This person would have to speak German and be familiar with the subject area. Yates contacted Dr. Dr. For this reason. as well as for carrying freight across oceans. James Iversen. Dr. Professor in the Department of Aerospace Engineering. Stanley Yates. Dr. 3 of Special Collections. Lippisch several times before her death in late 1991. Lippisch had envisioned use of his ground effect aerofoil-boat on remote river systems for the transport of food and medical supplies. Iversen has a strong interest in history and has been very supportive of the effort to organize the Lippisch Collection. Dr. Dr. the author organized the material according to Lippisch's major research areas.

The experimental testing of that wing section is the subject of the Ninth Chapter. The final chapter includes conclusions. 4 various experimental techniques are identified and compared in Chapter 5. The Eighth Chapter is a description of the panel method used to find a wing section that would be useful near the ground. . In the Twelfth Chapter the proposed ram-air train is sized and its feasibility is discussed . open questions and plans for the future. In Chapter 10 an appropriate design tool for airfoils in ground proximity is chosen. resulting in an optimization of the airfoil in the Chapter 11.a sketch of the train is given in Figure 1 below. A motivation for studying two-dimensional wing sections in ground effect is given in Chapter 7. with a discussion of the aerodynamic theories that describe those regimes. This is followed by a chapter on the different flow regimes of ground effect.

or aerodynamic.1: Proposed ram-air train. Figure 1.cv) train . iWSSSKM®. levitation (Aerol.

which make use of the ground effect as they launch themselves out of the water with the aid of their wings to skim just above the waves for a brief . Examples of these birds are albatrosses and swans (see [19] and [20]). In fact. it also uses its feet to run on the water for additional thrust.1). During takeoff and landing these birds utilize the dynamic ground effect (dynamic because their wings are in motion to create even more lift). 6 CHAPTER 2. A summary of experimental findings will be given in each case. The presence of the ground only becomes significant when the wing-to-ground clearance is less than a semi-span of the wing (for an explanation of some of the aerospace engineering terms see Figure 2.2 and are described in the following sections. Another example from nature is flying fish. The various flow regimes are depicted in Figure 2. One usually finds that any efficient human mechanism already occurs somewhere in nature. THE AERODYNAMIC GROUND EFFECT AND ASSOCIATED FLOW REGIMES A wing is said to be operating in the aerodynamic ground effect (AGE) if the air flow around it is influenced by the presence of the ground. As the wing descends from that altitude to a height of just a few millimeters from the ground. This is also the case with wings in ground effect. When a large seabird slowly takes to the air its wings' trailing edges are just barely above the water surface during the downstroke. the nature of the flow changes quite dramatically.

L drag.1: Some cierospace nomenclature . b freestream. I <z € span. D thickness- X=0 ground A: = 1 leading edge trailing edge Figure 2. uoo ^ lift.

8 Aerodynamic Ground Effect (AGE) J Wings Near the Lubrication Small-Gap Regime Ground or Water Regime J L Strong or Extreme Classical Moderate or Close Ground Weak Intermediate Ground Effect: Ground Ground Effect: Boundary Effect Effect Ram-Wing Layer Regime Regime J Figure 2.2: The flow regimes of the aerodynamic ground effect .

the ground becomes a leading order influence on the flowfield. Nevertheless. Wings Near the Ground or Near Water When swallows shoot across a lake at dusk to feed on insects and when flying fish are airborne. especially by pilots who found that their craft seemed to ride on a cushion of air just above the runway. This is termed the weak ground effect regime. the effect of the ground on an airplane operating in its vicinity has been recognized for quite some time. The flowfield around an airplane operating at a height of approximately a semi- span above the ground is only affected by the presence of the ground at second order. This situation involves the intermediate ground effect regime. meaning that the ground effect changes the flowfield by more than ten percent. If the airplane is allowed to descend to within a chordlength of the ground. Pilots on long transoceanic voyages found that flight near the water was very stable and considerably extended the range.e. i. The conclusion was that a wing near the ground experiences more lift and less drag than an isolated wing. . 9 moment (see [21] through [38]). meaning that the ground effect changes the flow-field by less than ten percent. Airplanes certainly do not usually descend to within a chordlength of the ground.. In all of these examples the wings are at a distance of approximately a semi-span from the surface. The same is true during the takeoff and landing phases of airplanes. their wings are just inches away from the water surface. a wing which is far away from the ground.

The reduction in downwash causes the lift vector to tilt forward. airplanes experience a pronounced air cushion. The Do X was built in the 1920s. In the case of two-dimensional flow the wing is assumed to be infinite so that there are no wingtip vortices. This prolonged glide is not usually desirable in that it increases the landing distance. An alternative explanation is needed here because one is dealing with wing profiles. The cushioning effect is the reason for the often smooth touchdowns and takeoffs of aircraft. Al­ though the ground has only a second order impact upon the flowfleld. and the reduced drag results in a lower fuel consumption. thereby reducing the induced drag. In 1927. With a wingspan of 137 feet. The effect of the ground is to prevent higher pressure air from escaping from underneath the wing to the top side (see Figure 2. a length of 131 feet.3).000 pounds (54. The increased fuel economy was utilized by the Dornier Wal flying boats and by the Do X flying boats of the German postal service when they were operating in the south Atlantic in the 1930s (see [39]. The increased lift-to-drag ratio is due mainly to this reduced drag. it was the largest aircraft of its day. There is an increase of the lift-to-drag ratio. a flight weight of 123. [40] and [41]). as well as for the prolonged glide (floating) of low-wing airplanes on landing. Charles Lindbergh reported taking advantage of the ground effect to increase range. The strength of the wingtip vortex is therefore reduced. 10 Weak (classical) ground effect regime This flow regime entails ground clearance ratios (height-to-chord) that are greater than 3 and includes the case of aircraft flying above water or above runways.5 tons) and twelve engines in back-to-back pairs for thrust. The weak ground effect can be explained in a variety of ways. as is the associated downwash. The most elegant interpretation may be to use .

3: Wing showing the wingtip vortex . wingtip vortex induced drag Figure 2.

how­ ever. This effectively models the ground since there is no flow across the symmetry plane. As a consequence. the ground exerts a stronger influence upon the flowfield. WIG regime Once the ground clearance decreases to the order of a chordlength. meaning that following a nose-up disturbance the profile will of itself return to its original attitude.4). the maximum lift is reached at a lower incidence. The incidence for zero lift rises to become less negative or even positive. The profile drag remains constant. The aerodynamic center moves aft to increase the static longitudinal stability. it is the drag which is due to viscosity (skin friction drag) and to the vertical displacement of the air near the profile (pressure. This type of flow is known as the intermediate or moderate ground effect regime or as the wing-in-ground-effect (WIG) . or form drag). The image airfoil will induce small vertical and horizontal velocities at the airfoil. mean­ ing that there is a stronger increase in lift for a given increase in incidence. the lift-to-drag ratio increases and the induced drag decreases as the height decreases. One can thus interpret the ground effect as the image airfoil's influence on the flowfield around the airfoil. There is stability of height (heave stability). Experiments (see [42] through [74] and [5]) have shown that for wings at positive incidence in weak and moderate ground effect. meaning that following an upward displacement the profile will of itself descend to its original altitude. what happens to the maximum lift itself. It is not clear. The lift curve slope also rises. 12 the image method. the lift at zero incidence remains constant. One imagines the ground to be replaced by a mirror-image of the airfoil so that the midplane between the two airfoils forms a symmetry plane (see Figure 2. Moderate (intermediate) ground effect regime. Finally.

4: Profile in ground effect: image method . 13 suction lift low pressure main wmg high pressure ram-lift midplane symmetry plane image wing Figure 2.

viscous effects become significant. it not only uses its feet for additional thrust but also utilizes the strong ground effect. vehicles have been specifically designed to operate in strong ground effect. The other 30 to 40% are due to the higher-than-normal pressure underneath the wing. This low pressure on the top surface effectively sucks the wing upward.5 and 3. the inertia effects become negligible in comparison to the viscous effects. Since the air underneath the wing slows down. This is the realm of lubrication theory. the mirror-airfoil is inducing higher velocities near the airfoil than in the case of weak ground effect. ram-wing regime When a swan or a duck struggles to become airborne. If the gap becomes smaller still. These terms are no longer negligible. For this reason. and the extreme ground effect regime is entered. From a two-dimensional perspective. 14 regime. Of the total lifting force experienced by a wing-in-ground-effect. This means that more air flows over the top of the main airfoil. Strong (close) ground effect regime. This is because the trailing . one must include higher-order terms in the solution when the ground clearance-to-chord ratio is between 0. But experience has shown that the benefits of the ground effect continue growing with decreasing ground clearance. the pressure rises. causing an increase in the upward force. lowering the pressure there. Theoretically. a phenomenon known as suction-lift. between 60 and 70% is the suction-lift produced by the airflow over the top of the airfoil. Small-Gap Regime Due to geometric and safety constraints most aircraft do not operate at heights less than a chordlength. If the clearance is further reduced.

The result of these phenomena is a lift increase. the gap under the trailing edge is still larger in magnitude than the thickness of the boundary layer at that point.1. Reduc­ tions in ground clearance revealed a decreas drag for equal lift as well as an increase of lift at equal (positive) incidence. The air beneath its wings is nearly stagnant. For the case of the ram-wing regime. approximately. Of course. resulting in near-stagnation pressure on the wing-underside. But the point remains that when the trailing edge is close enough to the ground it interferes with the outflow of air underneath it. This ram-pressure causes a great increcise in the lifting force. [68] through [74]). As the air underneath the wing moves toward stagnation.05 to 0. the problem is complicated by bird beating its wings at the same time.3. the remainder being the suction-lift produced by the accelerating flow over the top surface. This allows the theoretical analysis to be of the inviscid type. ranging from 0. [54]. [52]. Any aerodynamic theory must therefore be inviscid and nonlinear. The resulting lift force is called ram-lift. 15 edge of its wing is just millimeters above the water surface during the downstroke. The trailing edge ground clearance-to-chordlength ratio is of order 0. it forms a ram-air cushion and the pressure rises. between 60 and 70% of the total lifting force is ram-lift. For normal airfoil thicknesses (about 10 percent) and moderate incidences (about 5 degrees). [5]. the geometry of the gap varies significantly so that the problem is nonlinear in that region. The ram-wing regime has been investigated experimentally by several researchers (see [48]. In fact. causing a highly unsteady flow. [61] through [65]. Bagley [58] showed that the boundary layers must . Nature is far more sophisticated than humankind's attempts at flight. causing the air to slow down. [57] through [60].

the flow was separated on the top surface. Miiller's results were also reported in [75]. Unfortunately. which validated the image technique and negated the need for a wake splitter plate. It is fortunate that Datwyler [48] and Muller [54] carried out flow visualization studies of the strong ground effect as this allows a validation of the flow characteris­ tics. free of separated regions and with negligibly thin boundary layers. he did not report the Reynolds number. One of Miiller's experiments involved a ten percent thick airfoil and its image. whereas it rose near the trailing edge. which is two and one-half times as high as for an isolated wing. 16 be considered if pressure distributions are to be found. Separation did not occur for airfoils with a thickness of ten percent and angles of attack below ten degrees. but he found that the wakes do interfere if the flow is separated. however. The maximum lift remains constant. The maximum lift-to-drag ratio was around 20. This is especially true on the bottom surface near the trailing edge. . confirming that the flow there is retarded as long as it is attached. Werle [61] confirmed the above results many years later. Serebrisky and Biachuev [52] found that the pressure dropped slightly near the leading edge of the upper surface. The photographs revealed a smooth flow through the gap. He noted that the two wakes did not interfere with each other. They also found the pressure on the lower surface to rise strongly. Due to the large incidence.17. This unfavorable pressure gradient causes separation to occur at a lower inci­ dence than for an isolated wing. but the thickness probably should not be much smaller. The incidence was approximately ten degrees and the trailing edge to chordlength ratio was near 0.

The problem is then governed by lubrication theory. An example of this is when two sheets slide with respect to one another. 17 Extreme ground effect regime When the gap under the trailing edge is of the same order of magnitude as the boundary-layer thickness. The gap flow is then governed by boundary-layer equations. Lubrication Regime The ground clearance can be reduced to such extent that the inertia effects become negligible in comparison to the viscous forces. viscous effects can no longer be neglected. .

or aeroskimmer.. This is called a planing craft or a hydroskimmer. If one flies the same wing further away from the ground. This is the ram-wing. the present chapter begins with a summary of the different types of craft. 18 CHAPTER 3. it is possible to fly a wing in such a way that its trailing edge prevents air from flowing out underneath it. or an Ekranoplan (anglicized version of the Russian word). For this reason. A SURVEY OF VEHICLES FOR OPERATION IN VARIOUS GROUND EFFECT REGIMES The nomenclature describing ground effect vehicles is often used carelessly. i. If one directs engine exhaust under the wing while its trailing edge is nearly touching the surface it is called a power-augmented ram-wing (PAR). For example.e. providing both a history of its development and a description of the relevant aerodynamics. it is called a wing-in-ground-effect (WIG). the associated ground effect regime. One can also design vehicles that use a combination of the . One could also make the rear portion of the undersurface touch the water and make the craft ride on a film of air that provides air-lubrication. The Categorization Of Ground-Effect Vehicles The many categories of ground-effect vehicles have evolved as a consequence of the different ground-effect regimes. Each subsequent section has as its subject a different type of ground effect machine.

Wing-In-Ground-EfFect (WIG) Vehicles (Ekranoplans) The most common ground-effect phenomenon involves the take-off and landing of common aircraft. An aircraft operating this way is called a wing-in-ground-effect (WIG) airplane. as will be discussed below. In the 1960s. there were also many tests . only they will be described in the following sections. as shown in Figure 3. a number of review articles were written about ground-effect ma­ chines (see [76] through [88]). it is also called an Ekranoplan airplane. But even for low-wing aircraft the distance between the trailing edge and the ground is on the order of a chordlength.1. Of course. This is in contrast to an aerodynamic ground-effect (AGE) craft. leaving plenty of space for air to escape underneath it. and it is capable of hovering without forward motion. Generally these Soviet aircraft also operate in other ground-effect modes. This led to systematic studies of the effect of the ground on full-size aircraft in the 1920s and 1930s (see for example [89] through [92]). which relies on forward motion to enable its wings to produce lift. Since aerodynamic ground-effect vehicles are the focus of this work. The present work relies particularly on the article of Foshag [88]. 19 above-mentioned ground-effect regimes. Based on the Russian word. On or very near the ground the wings of all common aircraft are within a semi-span of the runway. Early aviators and aeronautical engineers were very preoccupied with the takeoff and landing phase. This is known as an aerostatic ground-effect vehicle. Or one can forget about the use of wings all together and simply use fans to force high-pressure air into a chamber sealed with rigid or flexible skirts.

between (use forward e. Power- Aerostatic Planing Aerofoil Peripheral Ground. hovercraft aerostatic and motion to to o aerodynamic GEM) produce lift) X" Wing-in.g. Jet Effect Machine. Ground Effect Vehicles (GEV) Ground Effect Machines (GEM) Air Cushion Vehicles (ACV) Captured Air Aerodynamic Bubble (CAB) GroundEffect Aerostatic GEM Vehicles (AGEV) Vehicle (hybrid (uses lifting fans).. mented Chamber Hydro. Machine Winged Machine (WIG) Aero. Ram-Wing Vehicle skimmers Hulls Machine skimmer ^ (I^R) ^ Figure 3. PARAVIG -Boats. Ram. Wing Aug­ Plenum Craft.1: Categories of ground effect vehicles .

This cushion of stagnated air produces a strong lifting force on the wing. i. It was a torpedo boat designed and built for the Austrian Navy in 1916 by von Thomanhul (see [97]). Vessels that employ both a ram-wing and sideplates are known as channel- flow vehicles or tunnel-hulls. His . between 60 and 70% is the suction-lift produced by the high-speed airflow over the top of the airfoil. the remainder being the suction-lift produced by the accelerating flow over the top surface. One can further reduce the efflux of high-pressure air from underneath the wing by installing sideplates. These endplates are connected to the wingtips and reach down to just above the ground. thereby trapping high-pressure air under the wing. such as NACA's tests involving delta-wing models (see [93] and [94]) in the 1950s and British studies of a Gothic-wing in 1960 (see [95] and [96]). Ram-wing vessels are not capable of free-flight because they usually do not generate sufficient lift away from the ground. the ram-air cushion. Between 60 and 70% of the total lifting force is ram- lift. The other 30 to 40% are due to the unusually high pressure underneath the wing. 21 on models. In addition. Following the Austrian effort of 1916. Ram-wings operate with their trailing edges just above the surface.e. preventing air from flowing around the wingtips. the next application of the ram-wing con­ cept was the sledges built by the Finnish engineer Kaario beginning in 1935. Ram-Wing Vehicles (Aeroskimmers) The first vehicle designed exclusively to make use of one of the ground-effect phenomena was of the ram-wing type.. they are typically only stable near the ground. Of the total lifting force experienced by wing-in-ground-effect craft.

In the late 1950s. This was not a pure channel-flow vehicle because it used a peripheral jet sidewall system for lift at low speeds. This ship. The MARAD experts envisioned large air-cushion vehicles (ACVs). The project was under the direction of Peter J. The ground clearance to wingspan ratio was between 0. called the "Columbia". Other efforts during this time period included the ram-wings of Crook [101] and the compression-plane of Warner (see [102] through [106]). All of these vehicles experi­ enced pitch instabilities and control problems. The 30-foot long VRC-1 was built and first tested at Edwards Air Force Base in 1964. Vehicle Research Corporation (VRC) designed a model of a channel-flow vehicle. Mantle. In 1960. also known as surface-effect ships (SESs). 22 propeller-driven channel-flow sledges flew just centimeters above frozen lakes of his native land (see [98] and [99]). Kaario added long skis to his designs to avoid pitch-up problems. Between 1932 and 1959 Kaario built several improved versions of his sledge. the VRC-1. VRC was awarded a contract with the MARAD to build the model and to design a 100- ton vehicle for up to 150 passengers or 40 short tons of freight. and when pitch stability problems were encountered the project was terminated (see [107] through [115].-American merchant fleet more competitive.S. A symposium at Princeton University in 1959 [100] triggered a flurry of activity. At higher speeds.05. later at General Dynamics. it transferred to WIG mode with the side jets operating to function as air seals. Only hovering tests were carried out. was to have a range of 800 kilometers and speeds between 185 and 220 km/h (115-140 mph). His work remained nearly unknown until a strong interest in ground effect vehicles developed in the late 1950s. the United States Marine Administration (MARAD) sought ways to make the U. In fact. [67] . Kaario gave a presentation [99] at that meeting.02 and 0.

23 and [116]. Improvements to this design led to the . its length 20. [118] and [119]) and that of Swietzer Brothers (see [120]).3 feet. calling it the "winged hull" (see [132] through [148]). Other ram-wing vehicles built in 1961 were that of Bertelsen (see [117]. a constant chord of 3. The KAG-3 had satisfactory pitch stability. Kawasaki was building a ram-wing vehicle. a span of 8 feet. The ram-wing lifted most of the hull out of the water and the vessel's speed increased by 50%.27 and achieved a lift-to-drag ratio of 14. The KAG-3 had a wing with an aspect ratio of 0. page 12). Engineers fitted it with hydroskis at the bow and added rectangular wings with endplates on both sides. The Swietzer craft was tailless. Nevertheless. it displayed severe porpoising (see [121] through [131]). It had an aspect ratio of 1. In Japan. Its floats served as end plates and provided stability since they were always in light contact with the water. The vehicle attained speeds of 80 mph over two-foot waves. Its gross weight was 1320 pounds.58 meters. an angle of incidence of six degrees and a NACA 6409 airfoil section. It achieved lift coefficients in excess of unity. It was propelled by an 80 HP outboard motor and had horizontal stabilizer not far from the water surface. and its width 17. had a low aspect ratio. Its wingtip floats served as end plates and provided stability by remaining in light contact with the water. and its cruise efficiency Wcis over twice that of a good planing boat. and a length of 19 feet. the project was halted because it was found not to be sufficiently seaworthy.7. Under some conditions.6 feet. This was essentially a speedboat driven by an outboard motor. The KAG-3 project began in 1961 and was supervised by Ashill and Ando (see [116]. page 47). Lockheed-California began design studies of a ram-wing vehicle in 1961.

which encountered pitch stability problems. Its wings operated close to the surface and it was powered by an outboard motor. As will be discussed in a later section this vehicle operated as a ram-wing during parts of its flight only. Lippisch speculated that this may have been caused by flow separation on the excessively thin wing section. Armstrong's vehicle was essentially a stepped-bottom raceboat with a delta-wing that had large flaps. [3]. such as the craft built by Armstrong in 1962 (see [151]) . The full-scale boats were tested between 1969 and 1973 and were designated as X-114 through X-117 (see [2]. 24 Dynamic Interface Vehicle (see [149] and [150]). [4] and [12]). Lippisch applied the same idea to his aeroskimmers. which was just an open-water test system. Lockheed continued its studies of ram-wing vehicles through the use of the "water wind-tunnel".as they got too high ofF the ground the nose tended to pitch upward. Lippisch mounted the horizontal stabilizer so high that it operated out of ground-effect and he made it large enough to provide pitch stability at all attitudes. Pitch stability problems continued to plague ram-wing vessels. He began designing these raceboats for KiekhaefFer Mercury in 1968 . . Models were simply pulled by a boom extending ten feet to one side of the bow of a speedboat. It should be noted that the designation X-114 appears twice in Lippisch's work: once for the aeroskimmer designed by Lippisch Research Corporation (LiReCo). After encountering some initial stability problems. sometimes causing the craft to flip over backward. The pitch stability problem was solved by Lippisch in 1963 when he added a large horizontal tail to his Collins X-112 (see [5] through [16]).his models were designated as AS-1 through AS-5. and once for the aerofoil boat built by Rhein-Flugzeugbau (RFB) beginning in 1974.

Additional theory was developed in 1974 by Miller and others (see [160] and [161]). In 1973. or by making the craft plane on a film of air. Gallington and his co-workers reported successful flights of their ram- wing models . To circumvent this problem the contact area between the vessel and the water must be minimized . whose name appears in conjunction with S.I. It derived part of its lift from the redirection of the propeller slipstream underneath the vehicle. [70]. . [153] and [154]). P. Air-Lubricated Vehicles (Hydroskimmers) It was probably in the era of the first world war that both boats and flying boats needed to achieve higher and higher speeds. Widnall and with high speed ground transportation projects carried out at M. Planing. •25 Theoretical work involving ram-wings was performed by Barlow at the David Taylor Model Basin in 1962 and by Timothy Barrows working for the DOT/FRA between 1970 and 1972 (see [152]. [158] and [159]).T. In the same year a research project directed by M. Work began on the ram-wings of Dickinson [155] and of Kumar [156] in 1963. under funding from the DOT/FRA in the early 1970s. But a vessel operating in the displacement mode is subject to tremendous hydrodynamic drag.they had used Lippisch's high tail configuration to avoid pitch problems (see [69].E. aerostatic air cushions. Knowlton at Princeton University [157] involved a 1/16 scale wind tunnel model of a ram-wing for high-speed ground transportation. Working as a student with the Princeton team was Timothy Barrows.this can be achieved by using hydrofoils. In 1967 Kumar again reported on his efforts.

Power-Augmented Ram-Wing (PAR) One disadvantage of the ram-wing is that it only works after the forward speed is sufficiently large to create an air cushion. The westernized name given to Soviet vehicles of this type is nizkolet craft. Alexander Lippisch began work on planing boats in the late 1950s (see [2]. In 1961. The results are very low speeds and very little loading during the takeoff and landing phases.a tundra . Aerophysics Co. began studies on its Hydroskimmer HS-2 (see [165]. This type of vehicle is called a power-augmented ram-wing and it seems to have been developed in the Soviet Union at the Odessa Engineering Institute of the Merchant Fleet (OIIMF) . Studies on planing surfaces include the 1948 work of Korvin-Kroukovsky (see [162] and [163]) and the 1954 work of Savitsky and Neidinger at the Stevens Institute of Technology [164].the 1970/71 issue (page 327) of Jane's [168] includes a desription of the OIIMF-2. [4] and [12]). Budnitskiy. To create this air cushion at an earlier stage one can mount the propulsion unit ahead of the wing and direct its exhaust over and/or under the wing. built by students under the direction of Y. Most recently came news of the newest Soviet PAR/WIG design [170] . A similar configuration also appeared in a 1967 patent of Bertelsen [169]. He designed some planing speedboads during the early 1960s - they were designated as X-101 through X-106 and served as testing platforms for Collins Radio's avionics equipment. 26 Air-lubrication is achieved by lifting the bow and most of the mid-section of the hull out of the water and by skimming the water surface with the stern. [166] and [167]). It was a tandem-wing channel-flow vessel whose two propellers forced air past the main wing. [3].

Panchenkov. The first trials of this 313- ton airplane took place in 1965. The front engines could be tilted to produce the PAR-effect.2 and 0. it is then a wing-in-ground-effect (WIG) vehicle. capable of free-flight. Capable of 300 knot speeds. over deserts and over snow-covered terrain at 300 mph at a height of 40 feet. instead. •27 triplane designed by A.8. It had a span of 40 meters (131 ft). allowing exhaust to be directed under or over the two rear wings. it had eight turbine engines mounted on a stub-wing near the nose and two engines on the tail. The ratio of ground clearance to chordlength was between 0. As the speed increases. a length of 91 meters (300 ft). The vehicle has a wingspan of 166 feet and a length of 408 feet and.N. all of the engine output is used for thrust and the machine becomes a pure ram-wing. the trailing edge leaves the immediate vicinity of the surface and high-pressure air is able to escape from underneath the wing. it will skim over tundra. a chordlength of 17 meters and was able to carry a payload of 94 tons. This corresponds . follower of the Italian-born Soviet designer Robert Oros di Bartini. could pose a threat to the delicate tundra ecosystem. As the craft ascends further. The first PAR/WIG craft was the so-called Caspian Sea Monster. During takeoff and landing engine exhaust is directed under the wing to produce the PAR- effect. pages 50-52). This 4.000-ton vehicle will not be capable of free-flight. approximately. PAR/WIG Vehicles This type of hybrid vehicle operates in several ground effect modes. Of its twelve turbojet engines. if built. six are mounted just over the front-most wing. designed by Rostislav Evgenyevitch Alekseev (see [171]. which corresponds to 900 troops.

19 and sideplates to prevent wingtip vortices.000 pounds were pay load (see [172] through [178]. [179]. The vehicle had a length of 73 meters (238 feet) and its 33-meter (108 foot) wings had an aspect ratio of 1.4 and a cruise ground clearance ratio of 0. pages 128-129). The next Soviet PAR/WIG aircraft was the "Orlyonok".4 million pounds.150 Ekranoplan (see [180]). was involved in this project as well. page 133. Forward thrust was provided by a single turboprop engine mounted on . or "Little Eagle". It had four turbofan engines on a stub wing near the nose. The Naval Air Development Center had been carrying out related wind-tunnel work in the early 1960s. The range was to be 4. and [171]. at a cruise Mach number of 0. 28 to the trailing edge being between 3. Captain Thomas Meeks of the Naval Air Development Center initiated the Advanced Naval Vehicles Concepts Evaluation in July of 1976. The reason given was that the cruise speeds were too low.15 or less (-^ = 2Tm^' These plans were shelved despite good fuel economy and good payload fractions. The Orlyonok did not have canards.5 and 14 meters above the ground. In the United States.000-pound aircraft was the Sukhoi-Nizhny Novgorod A.000 nautical miles. Lockheed-Georgia designed a PAR/WIG logistics transport with a takeoff weight of 1. of which 441. The designer of the Caspian Sea Monster of the early 1960s. Their exhaust was trapped in the space under the 103-foot wing. Similar aircraft were designed by the Italian-born Soviet designer Robert Oros di Bartini. Rostislav Evgenyevitch Alekseev. The official des­ ignation of this 275.90. It had two takeoff turbine engines mounted in the underside near the nose that only operated during takeoff and landing. de­ signed by the Ekolen design bureau and first flown around 1980. which was sealed by sideplates and special flaps. this canard could be rotated for power-augmentation or kept straight for pure forward thrust.

in some cases. a length of 8 meters (23 ft). It had an aspect ratio of 1. With a length of 190 feet. Powered by a two-cylinder. This vehicle demonstrated the possibility of operating in all modes. The built-in incidence was 5. Such a vehicle was first built by Lippisch in 1963.5 meters (14 ft) and an empty mass of 170 kilograms (400 pounds). at very low speeds. It can operate in many modes.8. 25-horsepower engine it achieved a lift-to-drag ratio of 25 and a transport efficiency of 20 ton-kilometer per kilogram of fuel (about 30 ton-mile per gallon of fuel). The X-112 had a large T-tail to provide pitch stability. the craft could carry 30 tons of supplies or 150 passengers. or winged hull. with the floats sealing off the wingtips similar to sideplates. ranging from a displacement boat to a free-flying aircraft.7. He called his wooden Lip- pisch/Collins X-112 an aerofoil boat. If it leaves the surface more it becomes a wing-in-ground-efFect vehicle and finally. As the speed increases it resembles a hydroskimmer as it planes on its wingtip floats. Its wings had the shape of a reverse-delta and had reverse-dihedral. Its gross weight including two occupants was 710 pounds. a span of 4. a normal airplane. on its wings' trailing edges and on its stepped hull. it displaces the water like a ship.5 degrees. Planing/Ram/WIG vehicles (Aerofoil Boats. 29 the T-tail. With a chord of 30 feet this corresponded to a ground clearance ratio of 0. Winged Hulls) Another hybrid ground-effect machine resembles the flying boats of the 1920s. During cruise the speed was 250 mph (216 knots) and the clearance between trailing edge and ground was between 25 and 40 feet. the span loading was 10 and the ratio of weight to engine power was 50 The required power at a . It then lifts itself out of the water to become a ram-wing aircraft.

This resulted in two working prototypes. With the help of the West German Defense Department and the consultations of Dr. Lippisch. Wisconsin in late 1973. ey a wing area of 13 square meters (140 f t ).89 meters (19. It was amphibious since it had a retractable landing gear that folded into its wing-tip floats.3 ft).240. This so-called Weiland-tandem had a length of 500 feet. It was a center- fuselage trimaran design and had a fiberglass-tube structure. at low speeds . two wings of 500 by 100 feet and a horizontal stabilizer of 500 by 80 feet.6 ft. With its 40-hp Nelson powerplant.000-pound tandem- wing transport. rescue and civilian commuting. it was flown at the estuary of the Weser-river to demonstrate seaworthiness. It was a low-wing catamaran and was evaluated for roles in marine patrol. and it first flew in 1970. the X-113Am achieved efficiencies of 50 ton-miles per gallon of fuel. Despite the tandem config­ uration and the horizontal stabilizer the craft encountered the pitch-heave instability and was destroyed in a crash. a design weight of 300 kp. In late 1972. The 6-seat Lippisch/RFB X-114 had its first trials in 1977 and had a wingspan of 22 feet. The aircraft was do­ nated to the EAA museum at Hale's Corners (Oshkosh) near Milwaukee. Lockheed-California designed a 2. which is now a Messerschmitt Bolkow-Blohm [MBB] subsidiary) began work on its cierofoil-boats in 1967. Its wings connected two slender seaplane-like hulls that also served as floats. Rhein-Flugzeugbau (RFB was a subsidiary of VFW-Fokker Gmbh. and the project Wcis cancelled (see [62] through [64] and [181] through [184]). 30 speed of 150 km/h was 0. The single-seat Lippisch/RFB X-113Am had a length of 27. at speeds of 90 to 180 knots.2 (s®® [5] through [16]). a wingspan of 5. In the same year (1963). a flight weight of 360 kp (800 pounds gross weight).

whereas it improved to 20 in ground-effect.S. Both the X-113Am and the X-114 used seven percent thick Clark-Y profiles. 31 the landing gear also functioned as a water-rudder. This entailed the further development of the "Airfish-1". an aerofoil boat designed by RFB's technical director Hanno Fischer and his colleague Klaus Matjasic. In the Soviet Union news of the aerofoil-boat initiated research at the Central Laboratory of Lifesaving Technology (CLST) in 1971. It was built by a subsidiary of Deutsche Aerospace and could fly 280 miles on ten gallons of fuel. the ESKA-1.5 feet. Flarecraft introduced two new aerofoil boats in 1991. boat market in 1991. Its reversible-pitch propeller was housed in a shroud and was powered by a 200-horsepower reciprocating engine. In free flight. A commercial version of the craft called "Flarecraft 370" had its debut on the U. To complete the history of Lippisch's aerofoil-boats we briefly jump forward in time. But now back to the 1970s. Their efforts resulted in the flight of their first aerofoil-boat. [186] and [187]). no pilot's license was required. The two-seater could carry a pay load of 550 pounds and was able to fly over low obstacles (see [185]. The wingspan was 20 feet and the length was 34. the "Airfish" cruising speed was 72 mph and its range was 300 miles. in August . After some years of relatively little activity. to the late 1980s. RFB later added hydrofoils to improve the takeoff characteristics of the X-114. RFB began to cooperate with Flarecraft Corporation (formerly GECC: Ground Effect Craft Cor­ poration) in the United States in 1987. The CLST was a subdivision of the Rescue Organization for Inland Waters (OSVOD). Since the craft could not fly higher than a few feet. the X-114 achieved a lift-to-drag ratio of 8. Powered by a 65-hp engine in conjunction with a shrouded propeller. the "Airfish-2" and the "Airfish-3".

I.5 meters. In 1980 he first flew his 4-6 seater which was made of fiberglass (see [171]. Ivanov. His first full-scale vehicle flew in 1974. This huge transporter (by comparison. Gribovsky. Jorg in Germany during the early 1960s. In 1974.B. The ratio of ground clearance to chordlength was between 0. It could carry 20 kg per horsepower of engine output. Evgeniy P. Kuzakov and V. working at Lockheed California. S. Chernyavsky. This corresponded to heights of 0. Lippisch's proposals for a 300-ton winged hull transport became public in 1973 (see [189]. In order to be able to recognize the names of Soviet engineers in this field in the literature. pages 38-41). This "airfoil-flying-boat" and its successor had two seats and were not capable of free flight. Y. In 1966 Comisarov and Brasseur.3 to 1. Such a vessel appears in a 1965 patent [147] by Alter and Koriagin. A. studied a WIG transport aircraft [188]. A. Gorbenko. V. A similar freight . the C-5 Galaxy weighs 382 tons) strongly resembles the aerofoil-boats (see [190] and [191]). The idea of the tandem aerofoil boat (TAB) was taken up by Gunther W. Balnycv. This aircraft flew at 140 km/h (86 mph) and generated 40 to 45% more lift when flying in ground-effect.43.09 and 0. whereas a regular aircraft can only carry 4 kg. Gremyatsky. There have been thoughts of building very large winged hulls for the purpose of trans-oceanic transport. page 98). ESKA stands for Ekranolyetny Spasatyelny Kater- Amphibya (surface-effect amphibious lifeboat). a list of those involved in the ESKA-1 project is given here: A. Shavrov. pages 369-371). 32 of 1973 (see [171]. researchers at the Water Research Company proposed a 500-ton WSEV (winged surface-effect vehicle).B. N. Starting in 1963 he built 28 radio-controlled models. making this a tremendous improvement. Its wings were mounted at 2 to 8 degrees of incidence. Grunin.W. working at the DTMB.5 meters and a chordlength of 3.T.

400 ton gross weight. 33 plane was proposed in France by Bertin-Cygne (see [192]. Most of these transports only have a limited free-flight capability. . between 550 and 900 tons were payload. page 27). It was to cruise at 200 knots.000-1. of its 1. they are designed to operate optimally as wings in ground-effect (WIGs).

Many terms specific to the subject of high-speed trains will be used in the sections that follow. The railroad provides a track for locomotives and passenger or cargo cars. and has been influenced by his father. The ties or sleepers are wooden or concrete transverse supports to which the rails are fastened. For this reason some of them are discussed at this point. The topmost layer is the ballast. who is a steam locomotive enthusiast. the gauge is a measure of the distance between the two rails. Since the track refers to the two parallel rails. A SURVEY OF UNCONVENTIONAL TRAINS Introduction The reason for this chapter is that this dissertation has as its long-range goal the design of an unconventional train. The track gauge is the same as the width of the track. which is a permanent path having a line of two parallel rails fixed to ties and laid on a roadbed. Further. The track is also referred to as the classical steel duorail. which are given in chronological order ([193] through [236]). the author is very interested in the subject of public transportation. Please refer to Figure 4. Both consist of gravel or broken stone. They are called continuous rails if they are welded together. The railroad bed is the mass or heap that comprises the supporting surface for the ties and the track. followed by the subballast. It shows a railroad. The rails are bars of rolled steel. The information in this chapter comes from a variety of sources. 34 CHAPTER 4. The author believes .1.

each bogie carries one end of a railroad car. The concept of an articulated trainset was first proposed in 1818 [217]. If the railroad is electrified. then it is a superelevated track. if it is an articulated trainset. one refers to at-grade operation. If the track is banked in curves. which is a car used by the train crew. or poles. then a pantograph. and they are said to involve wheel-rail technology. This strip of land occupied by the railroad is called the right-of-way (ROW). The network of overhead lines is referred to as the catenary. or panto. The railroad bed rests on a foundation of earth which is typically owned by the railroad company or the state. However. Trainsets consist of one or more locomotives as well as passenger cars and cargo cars. 35 there to be a third layer called the suhgrade. Turning next to a description of a conventional train. from which the electric lines are suspended. Normally. i. or truck.2. or petrol.. The other cars are not powered and are also known as trailers or rolling stock. The only parts of the trainset in contact with the rails are the steel wheels. The locomotive houses the propulsion units and is sometimes called the power car. please refer to Figure 4. Steam locomotives also have a tender to carry water and fuel such as wood. Electric locomotives are powered by electric motors and pick up current from the overhead catenary or from special wayside electrified rails. is mounted on top of their roof. and two axles are mounted on a metal frame to form a swiveling carriage called the bogie. Some trains still have a caboose. if one rail is higher than the other. If the track is not superelevated. Two wheels share the same axle. and it is often a collapsible . The bogie is attached to the car or locomotive body so that it can swivel (rotate about a vertical axis).e. A pantograph is a pick-up mechanism. there are periodic masts. then two adjacent cars share a single bogie. Conventional trains are also called wheel-on-rail trains. If they pick up current from the catenary. coal.

^t?:^%i:7^ earth foundation Figure 4.1. Railroad nomenclature . 36 pantograph overhead line ' catenary steel rail tie or sleeper ballastr railroad bed subballast su"bKrade.

or aerodynamic lifting forces. super-speed trains (SST). diesel engines. A low-speed ground transportation system operates at speeds up to 60 mph. many operate on concrete or steel guideways or are suspended from monorails. Systems operating at speeds above 250 mph are called super-speed ground transportation systems (SS/GTS). or they are suspended from an overhead monorail. 37 frame. . moving cable systems. In addition to a classification according to technological differences. train sys­ tems are further differentiated on the basis of their speed ranges. vacuum forces. Unconventional trains seldom feature steel wheels. and even gravita­ tional forces. They are instead suspended by means of magnetic forces. A high-speed ground transportation (HSGT) system is designed for speeds between 125 and 250 mph. or also ultra-high-speed ground transportation (UHSGT) systems. Instead. pressurized air. turbojets (aircraft engines). It begins with a brief summary of the social. gas turbines. A medium-speed system op­ erates between 60 and 125 mph. The chapter is subdivided according to the different technologies of unconven­ tional trains. Propulsion systems in­ clude electric motors. Diesel motors and turbojet engines may drive an aircraft-type propeller. which may or may not be shrouded (ducted). linear induction motors (LIMs). The summary is approximately chronological and illuminates the relationships between the systems. economic and technological circumstances leading to the development of the various systems. political. Unconventional trains oftentimes do not use steel rails.

2: Trainsets and bogie types . 38 I C I 3Er:aE steerable bogie normal bogie power car (locomotive) trailer. passenger car end trailer power bogies normal normal trailer articulated bogie trailer motor bogies bogie driving bogies bogie Figure 4.

tube flight and TACV in the US) 1970s oil crises. research into unconventional vehicles (French Aerotrain. a brief sketch of these changes will be provided. the development of a new transportation system follows a significant change in the social.1: Milestones in railroad history 1712 invention of the steam engine by Thomas Newcomen (UK) 1804 first steam locomotive by Richard Trevithick (UK) 1821 invention of the electric motor by Michael Faraday (UK) 1832 invention of the linear induction motor by William Sturgeon 1860 invention of an early internal combustion engine by Etienne Lenoir (France) 1863 first subway in London (UK) 1876 invention of the internal combustion engine by Otto (Germany) 1879 first electric locomotive by Siemens (Germany) 1881 first electric streetcar by Siemens (Germany) 1885 first motor-powered automobile by Karl Benz (Germany) 1893 invention of the diesel engine by Rudolf Diesel (Germany) 1899 first diesel locomotive in Wiirttemberg (Germany) 1930s rise of diesel-electric locomotives 1950s rise of jetliners and automobiles 1960s increasing pollution and traffic congestion problems. or in technology. Three technological breakthroughs Three important technological achievements lie at the root of most conventional transportation systems. 39 Table 4. German ICE. research into conventional high-speed rail systems (French TGV. Japanese Bullet train. or economic climate. Before describing the different systems in the sections that follow. British hovercraft. A tabular summary of the changes and events is provided in Table 4. The first was the construction of the first steam engine by .1. Swedish X-20G0) A Sketch Of Train History Typically. maglev in Germany and Japan. political.

40 Thomas Newcomen in the United Kingdom in 1712. Trevithick built the first steam locomotive in 1804. which was also the year in which Karl Benz built the first motor-powered automobile. by William Sturgeon. by Michael Faraday in the United Kingdom in 1821. electric streetcars. The diesel engine was invented by Rudolf Diesel in Germany in 1893. Lenoir's early engines ran very rough. When electricity became more widely available toward the end of the 1800s. Germany. The third technical breakthrough was the invention of the internal combustion engine by Etienne Lenoir and Nikolaus August Otto in 1860 and 1876. became very popular in large cities. The first subway was built in London in 1863 and used steam locomotives. which averaged 20 mph. or rotary induction motor. respectively. The first practical combustion engines for transportation were built in 1885. the famous "Tube" of 1890. The second technological breakthrough was the invention of the electric motor. The first tramway was horse-drawn and opened in New York in 1832. It featured the first multi-tubular boiler and reached a maximum speed of 29 mph. The boilers could not produce enough steam to sustain speedy travel. but early locomotives had several problems. Robert Stevenson's "Rocket" of 1828 was the first useful steam locomotive. The first electric locomotive was built by Siemens and Halske in Germany in 1879. Closely related to this was the invention in 1832 of the first linear induction motor. Steam engines were improved by James Watt (improved condenser) and Richard Trevithick. and . or LIM. or tramways. by Siemens in 1881. and the cast iron rails broke all the time. and Otto's four-stroke engine was a vast im­ provement. The first electric tramway was built in Berlin. Siemens was also first to introduce the overhead contact wire. London was also the first city to inaugurate an electric subway.

they soon powered factories. These powerful loco­ motives became very popular in the mid-1930s. which supplies the electric drive motors. which was intro­ duced in Sweden in 1913. It is sometimes effective to combine two of the above propulsion systems. and steam locomotives. especially in the United States. and it operated near Albany in the state of New York. Even . One extension of this technology is the gas-turbine locomotive. The second combined system was the steam-turbine locomotive. Steam- driven locomotives and ships played an important part in carrying raw materials. Here a boiler supplied a turbine which drove a generator. such as in the name of the German company "Daimler-Benz" and its "Mercedes-Benz" automobiles). laborers. Steam locomotives and vacuum technology Even though steam engines were first used mainly in pumping water out of min­ ing shafts. steamships. That honor belongs to Nicholas-Joseph Cugnot. 41 Gottlieb Daimler the first motorcycle (their names are still associated with cars. The first locomotive powered by an internal combustion engine was a diesel locomotive in the German state of Wiirttemberg in 1899. The first of these combined systems was the diesel-electric locomotive. and it was not the world's first automobile. introduced in 1925. Here a jet-engine drives the generator. The first train used exclusively to transport passengers was the "De Witt Clinton" of 1831. and finished goods during the Industrial Revolution. That system was introduced in Switzerland in 1941. Benz's car was a tricycle with a two-stroke engine. who built his steam-truck in 1769. The first exclusively steam-powered train used in revenue service was the Liverpool&Manchester Rail­ way of 1830. Diesel-electric locomotives use their diesel engines to power electric generators which in turn supply electric drive motors.

This tube was slotted and housed the piston. The airplane's nose wheel assembly is hooked to a link that is in turn attached to the piston. The slot was usually sealed by means of leather flaps. when gasoline and rubber rationing reduced the number of auto­ mobiles. many countries continue to use steam locomotives in passenger and freight transportation. Piston-like passenger cars would accelerate to 150 . This re­ sulted in the construction of four vacuum-powered subway systems in England and France in the mid-1800s. A link connected the piston to a conventional passenger car. and many speed records were set during this period. In the 1960s. The final bloom of the steam locomotive in the Western world occured in World War Two. whereas the tube behind the car was vented to the atmosphere. 42 today. The technology was also known as atmospheric propulsion and pneumatic propulsion^ and it was proposed by George Medhurst in London in 1810. the tube in front of the car was evacuated using steam-powered pumps. carrying raw materials. A very similar system is used today to launch airplanes from aircraft carriers. reinforcements. the idea of atmospheric propulsion resurfaced as part of gravity- vacuum transit (GVT) systems. food. A similar system is today used by hospitals for transporting patients' records throughout the building. Such catapults consist of a slotted tube and a steam-powered piston. Steam locomotives were the work horses of the war machines. parts. and civilians. The rise of the airplane in the 1920s and 1930s resulted in the transfer of stream­ lining techniques to steam locomotive design. prisoners. troops. One by-product of steam-engine technology was vacuum technology. Other vacuum-powered subway systems utilized a smaller tube next to a rail system. Some systems used a cylindrical passenger car in a subterranean tube.

One variation of this idea was to generate a swirling air mass within this tunnel to facihtate the vehicle's levitation. Another system involving closed tubes was conceived in the 1960s. Such vehicles are actually lifting bodies and propel themselves through the tube by means of on-board engines and propellers. which consists of a tube-train that flies through the closed tube. But due to overcapitalization and the rise of the automobile. 43 mph while descending into deep tunnels.V. Other designs featured streamlined vehicles with cross-sections ranging from a thin sickle- shape to a nearly semi-disk shape. which is the device that transmits the electric current from the overhead wires (catenary) to the vehicle. No trial systems were ever built. The car would gradually lose speed and ascend to its destination using its inertia and the vacuum force. more than 30% of the streetcar companies were broke by 1919. and city catenaries are also used by trolleybuses. . the cost of tunneling and maintenance is prohibitive. Foa at the Rensselaer Polytechnic Institute proposed a vehicle consisting of an axisymmetric body supported by a number of short wing pads. Researchers under J. It is the tube-flight concept.P. Electric propulsion The electric motor and electrification of the large cities made possible the rise of electric streetcars at the turn of the century. Trolleybuses are named after the trolley. Streamlined hfting bodies were studied by research groups at Tufts University and under M. Knowlton at Princeton University. Just as in the case of atmospheric propulsion systems. Many of the early streetcar systems are still in operation today. Both the vacuum ahead of the car and the gravitational force would cause this acceleration.

the decades following World War Two were marked by eco­ nomic growth and high birth and immigration rates. Even the highly efficient (but heavy) steam locomotive became threatened by its diesel. and that the railways suffered decreased ridership. The railways further suffered because the gov­ . Diesel locomotives became very popular in the late 1940s. with a top speed of 116 mph (187 kph). Diesel propulsion Following its invention in 1860. The "Zephyrs" averaged 90 mph (145 kph) in the United States. Diesel locomotives are very powerful and are able to operate on tracks that have not been electrified. By the 1960s. the internal combustion engine rapidly replaced the steam engine in most industries and ships. several western countries and Japan began building electric locomotives. diesel engines drive electric generators which in turn supply the electric driving motors with current. many people owned automobiles and used the rapidly evolving fleet of commercial jetliners. 44 Beginning in the 1950s. This method of propulsion was particularly popular in the mid-1930s. A German diesel-electric train set a speed record of 133. The direct effect of this was that many highways and airports were built. Their advantage over other locomotives is that they weigh less because they do not carry any fuel. Diesel-electric propulsion With diesel-electric locomotives.and electrically-powered relatives.5 mph (215 kph) in 1939. Post-war population pressure and energy crises In Western nations.

In the late 1950s and throughout the 1960s. the first operational hovercraft was built in the United Kingdom by Christopher Cockerell in 1955. A further effect of increasing automobile ownership and commercial air traffic was that by the 1960s. and it made a Channel crossing in 1959. This was true particularly in the United States. But most efforts in the 1960s focused on more daring technologies. and system costs and energy . As a historical note. these problems forced Western governments to reconsider railroads and related new technologies. Coupled with the energy crises of the early 1970s. and passenger traffic in the United States all but collapsed. It was first to offer high-speed passenger train service in 1964 by introducing its Shinkansen "bullet-trains". The primary consideration was to diffuse the congestion. by 1968 France opted for a high-speed conventional train in order to minimize cost. there were serious congestion and pollution problems. 45 ernment effectively subsidized the trucking industry by building and maintaining the roads. It had the additional geographical problem of having its relatively large population confined to its narrow habitable coastline. A significant percentage of railroads was closed and abandoned. during which OPEC (oil-producing and exporting countries) threatened to reduce oil exports in order to cause the oil prices to rise. tracked air cushion vehicles were built and tested. The oil crises had not yet struck. These Aerotrains hovered on a cushion of pressurized air. France was the first country to seriously consider improving its public trans­ portation system. and the West was probably quite optimistic and confident. The first full-scale hovercraft was the British SRN-1. Japan was particularly enthusiastic about high-speed rail. However.

In the United States. and magnetic levitation technology. Personal Rapid Transit systems consist of cars for between one and six passengers. The energy crises of the early 1970s forced the West to reduce expenditures and energy use. Both air cushion and magnetic levita­ tion technologies require a large amount of energy to be spent on levitation. the Tracked Air Cushion Research Vehicle (TACRV) and the LIM-powered (Linear Induction Motor) Tracked Levitated Research Vehicle were first tested in 1974. Another product of the optimism and energy abundance of the 1960s were mag­ netically levitated trains. Perhaps it was this mood that prompted research in the areas of air cushion technology. Both tube-flight vehicles and personal rapid transit (PRT) systems received much attention in the 1960s. The wheeled versions may operate on concrete U-channel guideways or hanging from overhead unibeam guiderails. some tube-flight vehicles were propelled by combinations of gravitational and vacuum forces. only maglev appears to have survived. Of all high-speed ground transportation technologies studied in the 1960s. tracked air cushion vehicles were investigated in the United Kingdom and in the United States. tube-flight technology. personal rapid transit systems. Others were lifting bodies that simply flew inside tubes. Following the lead of France. The British prototype of 1966 was a tracked hovercraft called "Hovertrain". Many personal rapid transit vehicles featured air-cushion suspensions (such as hoverpads) and linear induction motors. Both . and featured a linear induction motor for propulsion. The train was built by Tracked Hovercraft. As discussed before. Both Germany and Japan began their research in the late 1960s. Ltd. 46 efficiency were secondary.. with first operational models in 1969 and first prototypes in 1971.

By the late 1970s. The oil crises caused Western governments to slash research funding in the area of high-speed ground transportation. and power companies can sell large amounts of electric energy to the operators. the windings. They had decided that all of the exotic proposals were too expensive and too risky. The emphasis on conventional steel wheel-on-rail technology in the 1970s resulted in the French gas turbine-powered turbotrains. aerospace companies can build the vehicle and its control systems. research funding stopped completely in 1975. Nevertheless. the United Kingdom's diesel-electric High Speed Trains (HSTs). In the United States. and TGVs (Train a Grande Vitesse) were . and Germany's electric Class 103. Only research into conventional high-speed rail was funded. and this has killed every German and Japanese initiative up to this point. Research into tube-flight continued for a few years in an altered form. the cooling fans. vehicles flying in troughs were studied in the early 1970s. Advanced Passenger Trains (APTs) were introduced in the United Kingdom. It is possible that maglev survived because many types of companies are able to profit from its construction and opera­ tion.1 locomotives for its IC (Inter-City) network. There was also some federal funding of maglev research in Germany and Japan (about $1 billion between 1969 and 1993 in each of the two countries). and winged vehicles suspended from a monorail in the mid 1970s. and only at a low level. electrical component man­ ufacturers can provide the linear motor. and the electrification system. Construction companies can build the steel and concrete guideways. an initial investment needs to be made by someone. 47 tube-flight and personal rapid transit systems require large investments into civil engineering structures (expensive tunneling and guideways) and have a very limited passenger capacity.

which ceased to be funded in 1991. But so far. Italy. Spain. As a result. but they seem to have been carried out for propaganda reasons. There were many studies of potential corridors for high-speed ground transportation. In the United States. The government seems reluctant to invest in a technology that will not yield spectacular results by the time the next elections come around. Conservative governments were in place in most Western countries. The organization of the different . Discussion Of The Various Technologies Following the discussion of the chronological sequence of railroad developments. this has not been the case. The 1980s were a difficult time for railroad research. Since the dissolution of the Soviet Bloc there has been some hope that defense conversion will free some money and talent for high-speed ground transportation research. 48 nearing completion in France. including Intelligent Vehicle Highway Systems (IVHS). encouraging privatization and individualism. One example is the National Maglev Initiative (NMI) of 1990. and the United Kingdom. and not for the public good. France. The prospect of a united Europe gave rise to plans for a European high-speed rail (HSR) network that would connect Belgium. Germany. there was hardly any federal support for public transportation. Portugal. AMTRAK was created in 1971 to consolidate the declining passenger rail system. each type of technology will now be discussed. Most transportation money went into highway research. Perhaps the stronghold of the large automobile companies and their corporate allies is still too strong to allow alternative transportation systems to flourish. the Netherlands.

Conventional trains faster than 125 mph are called high speed rail (HSR) trains [231]. Some of the faster systems. and signaling systems. cruise near 100 mph. The land surrounding the mound is usually owned by the government or by private railroad companies and is called right-of-way (ROW).S. Conventional high-speed rail trains (steel wheel-on-rail. These rest on a mound of gravel and earth. communications. light-weight trains. Aerodynamic shaping of HSR is crucial because both the aerodynamic noise and the aerodynamic drag dominate at speeds above 125 mph (see [231]). as well as sophisticated sensors. such as the Amtrak system in the United States. Conventional trains. such as the European Intercity trains or the U. Examples of these special requirements include continuous- welded rail. full electrification.-American electric Metroliners. or duorail tech­ nologies) Here conventional trains are defined as those which operate with wheels on steel tracks. dedicated high-speed tracks. concrete ties.3. limited curvatures. 49 technologies is depicted in Figure 4. typically operate at speeds between 70 and 80 mph. The material that underlies the track and ties is called ballast and subballast. Usually the tracks consist of two parallel steel rails fastened to ties that are made of wood or concrete. Electrification is necessary because diesel engines weigh too much. The width of the ROW for a double track is typically 40 to 50 ft. special ballast and subballast. they are also called wheel-on-rail trains (see [235]). Research on high-speed rail began in Europe in the mid-1950s. HSR trains are in a special category because their high speed necessitates special technologies and high investments. France em- .

3: Categories of trains and tracked ACVs . Levitation Motor Motor Motor Trains Motor Motor Hoverirains Trains (LIM) (Maglevs) Figure 4. Monorails Trains Duorail Linear Magnetic Diesel. Levitating (Wheelless) Steel Wheel-on-Rail. Tracked Vehicles. Trains Conventional Trains. Gas. Tracked Aerodynamic Diesel Electric Induction Levitation Electric Turbine ACV.

or Shin-Kansen (New-Line. The French research resulted in the "Mistral" trains but did not have a widespread impact. it was capable of 200 km/h (125 mph) (see [223]). research had begun on tilting trains. The construction cost was US$ 13 million per mile (1991 dollars. and Hikari (lightning) electric trains on the Tokaido line linking Tokyo and Osaka operated at 210 km/h (130 mph) (see [221] and [203]). This project was started in 1964. see [231]). In 1964. The oil crises of the early 1970s spurred the development of new transportation systems. Curves of 4. which were trains that tilted as they negotiated bends. as could gradients of l-in-65 (1. The first commercially successful and most famous of the high-speed rail trains was the Japanese Bullet train. A great drawback of the system is that six hours of maintenance on the track and the catenary are required every night. Britain's High Speed Trains (HST) and Advanced Passenger Trains (APT) and personal rapid transit (PRT) systems (see [220]). In 1977 the British HST diesel-electric locomotive reached 232 km/h (143 mph). the German Federal Railway (DB) introduced its Class 103. and cruised at 200 km/h (125 mph). In 1970.000-meter radius could be negotiated. two gas turbines drove generators that powered electric motors. among them France's turbotrains. also spelled Shinkansen).1 electric locomotive into its IC (inter-city) network. These trains achieved speeds of 200 km/h (125 mph) in 1973 and were manufactured under license in the United States by Rohr Corporation (see [214]). The first functional tilting train was the British APT of . The production version of the HST was called the Inter-City 125. 51 phasized electric propulsion whereas Germany favored diesel engines.5%). An example of the turbotrain was the RTG (Rame a turbine a gaz) built by the French company ANF-Frangeco.

and con­ ference areas were available (see [233]). Meals were served at the passengers' seats. The first-generation TGV.50 Hz AC drives. The APT was mothballed in 1978.000 Volt . The problems with tilting were solved by the Swedish-Swiss engineering group Asea Brown-Boveri (ABB). Its cruising speed was later raised to 270 km/h (169 mph) and it reached a top speed of 380 km/h (238 mph) in 1981. However. Their X-2000 tilt-train was developed in cooperation with the Swedish State Railways (SJ). The train sets were built by GEC Alsthom of Paris . 52 1972. France's National Railways (SNCF) felt that for high-speed trains. passengers felt quite uncomfortable during this maneuver because they did not feel the train turning even though they saw it turning. [227]. In addition. The advantage of tilt-trains is that they can negotiate turns 30 to 40% faster than comparable trains without tilting mechanisms. The X-2000 had radial self-steering trucks. It is for this reason that in 1968 France opted for a conventional train of great speed (train a grande vitesse. too much energy would be wasted for lifting power in the case of levitating vehicles. The electric APT-P (prototype) was designed for running speeds of 257 km/h (160 mph). the drag due to wheel friction was quite negligible compared to the aerodynamic drag. It was capable of braking from maximum speed within 1. the TGV (see [222]. This reduced the forces exerted by truck on the rail and pre­ vented derailments. It tilted slightly less and did not eliminate the sideward force completely. work areas. and telephones. and [230]).940 meters. or train a tres grande vitesse). its average speed was 133 mph (212 kph) and its maximum speed 168 mph. the TGV PSE (Paris-Southeast) entered regular revenue service between Paris and Lyon in 1981. Powered by 25. The train had AC propulsion and provided airplane-like service. so that passengers could still sense the approaching curve. with the two axles pivoting independently. fax.

In December 1989 the TGVA reached a speed of 300 mph. The line cost USS 7 million per mile to build (1991 dollars [231]). The TGV may become a European standard .5 million passengers) above that for the same routes in 1980. The TGV is shown in Figure 4. In 1989 a new generation TGV.in 1987 Spain launched . breaking the world's record of 253 mph set by a German locomotive (Transrapid 06. 53 Figure 4. the TGVA (TGV Atlantique) was introduced. TGV service had increased ridership by 75% (or by 6. Most of the gain came from auto and air traffic.4. In 1991 there were 251 TGV trains per day in France. It links Paris to Le Mans and Tours and cruises at 186 mph.6 million each. By the late 1980s. French high-speed train (TGV) and cost USS 11. 412 km/h) in 1988. In 1991 a modified TGVA reached 322 mph (516 kph) [231].

The unsprung mass of each bogie is about two tons. conventional high-speed . It is built by a consortium led by Siemens Transportation Systems and has been in service throughout central Europe since 1991 [235]. Service between the three cities began in 1992. In the same year. However. the friction between wheels and the rails rises with increasing speed. The TGV is able to tolerate grades up to 1 in 29 and curves with a 4. The one-meter diameter wheels are subject to maximum dynamic loads of 140 kiloNewtons. The Spanish TGVAs will be built in Spain by GEC Alsthom and are called "Ave" (the Spanish word for bird. each logging about 300.43 meters. A further entry into the HSR market in Europe is the Italian ETR-500.000 miles per year. as does the wear on the rails. It can brake from maximum speed to a standstill in 3. whereas the sprung mass is five tons. The X2000 is built by ABB Traction of Vasteras. The ICE cruises near 190 mph and has been tested at 252 mph (404 kph) [231]. and also an abbreviation for Alta Velocidad Espanola. and electrohydraulically tilted in curves.000-meter radius. Most of the sprung mass consists of aluminum and plastics. There were 60 ICE train sets in operation in 1993.5 in) high-speed lines between Madrid and the cities of Sevilla and Barcelona. In 1993 the third TGV line was opened between Paris and Arras [235]. 54 a railroad improvement plan that included the purchase of 24 TGVA trainsets and the construction of European gauge (1. or 4 ft 8. The German entries into the European conventional high-speed train competition include the ET 403 electric train and the ICE (Inter-City Express) train. which is expected to reach 186 mph [231]. or Spanish high speed). Sweden. the ICE was undergoing trials in the Northeast of the United States. as was the Swedish X2000 train. and axles experience static loads of twenty tons.000 meters. For a given running gear.

It was electrically powered and is still in operation.these would be open to the bottom and would accomodate 10-wheeled trolleys from which the train's body hangs.000 workers check every meter of the Japanese Bullet train tracks every night. Each computer-controlled vehicle would carry up to 24 passengers and reach speeds of up to 150 mph. A recent variant of high-speed rail trains is the CyberTran. it was the "Airtrain" by the helicopter designer Frank Piasecki [228]. and the author does not know whether it was allowed to swing laterally as it negotiated turns.rather. In the 1920s George Bennie developed a suspended monorail that had propeller propulsion. The first suspended monorail began service in the German city of Wuppertal in 1901. Monorails and other unconventional non-levitating trains Several types of trains fit neither into the category of conventional railroads nor into that of levitated vehicles. His "Glasgow Sail Plane" remained in the experimental stage. a lightweight electric vehicle running on steel tracks. The first of these is the monorail system. 55 trains display very little friction in their running gear . operating on conventional parallel steel tracks. He envisioned guide tubes fixed 35 feet above the ground . The propeller could be driven by either electric or diesel engines. The author believes that a related project was undertaken in Germany around the same time. The vehicle . which consists of cars hanging from overhead unibeam guiderails. they encounter mostly aerodynamic drag. Their relatively high cost is due to the need for very precisely aligned tracks as well as extensive maintenance. The most recent monorail vehicle was proposed in 1991. it was a light-weight train with propeller propulsion. about 3. For example.

. The same type of vehicle was proposed by Lehl and Zumwalt [215] in 1974. It was also called "Airtrain". The Airtrain would carry 90 passengers and reach speeds of 300 mph. so perhaps the two are related.the wheeled versions may operate on concrete U-channel guideways or hanging from overhead unibeam guiderails (see [208]). One example of this was the Morgantown PRT project. The body's construction would be light and of the aircraft type. A second unconventional system of mass transit is the personal rapid transit (PRT) system.5: The Piasecki Airtrain would be nonderailable and would be able to swing to the side as it negotiates curves (see Figure 4). 56 Figure 4. except that their train had sizeable wings for lift. It consist of cars for between one and six passengers .

which was a prototype rail car. the 5. These companies joined with Rohr Corporation of the United States to form a U. It is possible to suspend the vehicle on one large cushion or on a series of hoverpads. [220] and [171]. Designed to cruise at 200 km/h (125 mph) this train achieved 346 km/h (215 mph). it is also known as hovertrain. 2 meter wide and 1. As this type of train is able to hover on its aerostatic cushion. The first major research in this area was initiated in France in 1957 by the Societe de I'Aerotrain in conjunction with Bertin et Cie under its chairman Jean Bertin (see [194].6 meter high body was suspended above the inverted-T concrete rail by means of an air cushion . pages 175 through 178). Tracked air-cushion vehicles (TACVs) All of the wheel friction can be avoided if the wheels are abandoned in favor of a static air cushion. Examples are the State-of-the-Art Car (SOAC). In 1965 the group tested its first half-scale prototype. jet engines or electric motors.500 pound Aerotrain 01.-American subsidiary. Aerotrain Systems Incorporated.S. and the Transbus^ which was a luxurious bus. or tracked hovercraft. These cushions are generated by lifting fans that are driven by piston engines. and was sparked by the confidence that resulted from the lunar landing of Apollo 11 in 1969. Federal funding for new technologies for mass transit in the United States sharply declined following the advent of Republican administrations in 1980. The 10 meter long. The two-car train featured a flywheel energy storage system and an aluminum frame. The last large project involved the Advanced Concept Train (ACT-1). [195]. which was a prototype of a futuristic standardized urban rapid rail vehicle. 57 Much mass transit research took place in the United States in the 1970s.

58 produced by lifting fans (see Figure 4).35 psi (see [197]). It had two axial fans to create lift and guidance air cushions.8 meters. This 80-seat train had a length of 25. including rocket boosters for the high-speed trials.2 mile section between the French cities of Gometz-la-Ville and Limours. The air pads were of the flexible skirt. The track was a 4. The second half-scale prototype was the Aerotrain 02. These beams rested on 4.9 meter (16 ft) high pylons and had a width of 1. and its grade did not exceed 1%. open plenum type. it was made up of 19 feet and 5 inch sections of precast concrete. To decelerate from 155 mph to zero. Various engine types were tried. Braking was accomplished by pitch reversal and by break pads that grabbed the vertical rail. The cushion pressure required to keep the train suspended was only 0. which achieved a top speed of 411 km/h (255 mph) in 1969. There were also air cushions on each side of the vertical rail to provide guidance. In the case of a power failure the train slid on wooden rollers. Occasional contact between the vehicle and the guideway occurred and led to some gradual wear. which is about one-fifth of the pressure of a human's bare feet on the ground. the Interurbain 1-80 "Orleans". Propulsive thrust was provided by a shrouded 7-bladed propeller that was driven by a silenced turbofan engine. with a third engine of 260 horsepower providing forward propulsion via a propeller of two-meter diameter. In the same year the company introduced its first full-scale medium-range intercity Aerotrain 250-80. Two internal engines of 50 horsepower each provided cushion pressure. The guideway consisted of 20-meter (66 ft) prefabricated concrete beams that had an inverted-T cross section.6 meters (84 ft) and operated on the Orleans to Paris line. the vehicle required twenty seconds and a distance of 800 yards. Its total width was five feet and ten inches. The center guide rail (stem) was one .

6: French Aerotrain . 59 Figure 4.

Another effort was the 44 seat "Suburban Aerotrain 180-44". In March of 1968 they introduced the 30-passenger URBA-4 prototype in Lyon. The cars were suction- suspended from a monorail by means of a sub-atmospheric air lift system. To produce a centrifugal acceleration of 0. Its linear .000 pounds. 60 meter high. and it was capable of carrying the two-person crew (330 lb).300 ft. The production model will have automotive engines driving wheels. the track radius needed to be 28. Another French effort in the area of tracked ACVs was undertaken by the Com- pagnie d'Energetique Lineaire under Maurice Barthalon. The latest design was the "Tridim" or "Tridem" of 1975. Designed to con­ nect stations that were 2-5 miles apart.100 lb) and baggage and freight (3. In 1974 this train attained a speed of 422 km/h (264 mph).20. Each car had eight circular hoverpads.000 pounds. The drag coefficient for the vehicle was 0. It was to connect stations that were less than two miles apart.400 lb) with a total payload weight of 20. a speed of 400 km/h (250 mph) was achieved.900 lb). 84 passengers (13. In this configuration the train reached speeds between 230 and 300 km/h (155 and 185 mph). fuel (2. both lift and thrust were generated electrically [218]. at speeds of up to 80 km/h (50 mph). so it was better to use grades than to build a winding course.000 ft.15g at 250 mph. It could negotiate a ten percent grade and could stop from 155 mph within 2. After new turbofan engines were installed in 1973. All of the Aerotrain systems required very little track maintenance. the prototype had a Merlin et Gerin linear induction motor and achieved speeds between 160 and 200 km/h (100 and 125 mph). an urban mass-transit hovercar. The system also featured hydraulically retractable wheels for low-speed and emergency operation. Its empty weight was 20.

Knowlton at Princeton University [157].g. of California. The hope was to provide better intercity service in the Northeast Corridor and in the California Cor­ ridor (in this context. Tan [210] mentions several publications. These included tracked ACVs. pneumatic vehicles and tube- flight vehicles.V. In the United States. the British "Hovercar" and the "Hovair" proposed by General Motors in the U. [201] and [207]). P. Department of Commerce. Tracked ACVs of the time were the French Aerotrain. or "Hovertrain". The Rensselaer vehicle was an axisymmetric body sup­ ported by a number of short wing pads. e. [193]. The project was cancelled in 1973. The vehicle traveled over concrete track made of box sections. 61 induction motors permitted speeds of up to 65 mph (see [198]. page 3 and [168].S. three-year national research program sponsored by the U.. The shovel-shaped front end of the Princeton design featured a slit through which part of the propeller slipstream was channeled to . serious work in the area of high-speed ground transporta­ tion (HSGT) started in 1965 as a result of a US$ 90 million. Inc. Tube-flight vehicles (tube-trains) were proposed by J. pages 4 and 149). Part of the money was spent on conventional trains. One example of a pneumatic vehicle was the gravity vacuum transit (GVT) proposed by Transit. (HDL) to develop a model of a tracked hov­ ercraft called the "Hovercar". In the United Kingdom. Foa at the Rensselaer Polytechnic Institute and by researchers under the direction of M.. This model was made public in 1966 and featured LIM-propulsion and a plenum chamber hull fed by six axial fans.S. but most was spent on unconventional trains. the National Research Development Corporation (NRDC) sponsored Hovercraft Development Ltd.

Four proposals were received. or TLRV (Tracked Levitated Research Vehicle). The final effort under this grant was the PTACV (Prototype Tracked Air Cushion Vehicle). blown into the closed tube at regular intervals. [209]). a division of LTV Aerospace Corporation. one from Transportation Systems. a 150 mph. a division of the General Electric Company. It was 52 feet long and had three externally mounted turbofan engines to create cushion air as well as forward thrust. Its TACRV (tracked air cush­ ion research vehicle) was first tested in 1974 at the DOT's High-Speed Ground Test Center in Pueblo. There were even ideas to make the craft ride on a spiral of continuously moving air. Funding for all of these projects ended in the mid-1970s (see [214] and [168]. It hovered one foot above the concrete guideway near 300 mph (483 km/h). one from Grumman Aerospace Corporation.S. In March of 1974 that vehicle reached 234 mph. allowing a speed of 450 km/h (280 mph). the Garrett LIM research vehicle.5 million contract in 1971. it had wheels and was accelerated with jet engines before the LIM took over. . 62 pressurize the air under the vehicle. In 1969 the U. This motor was also tested on another vehicle. Department of Transportation (DOT) requested proposals for a 300 mph (483 km/h) tracked air cushion vehicle (TACV) to be used for inter­ city distances of about 300 miles (see [214]). Colorado.S. flying inside a tube (Tan [210] mentions a good review article on tube flight. and a late en­ try from Vought Aeronautics. one involved a 350 mph bullet- shaped vehicle with retractable wheels and two turboprop engines. Grumman was awarded the U. This vehicle was also built for the Depart­ ment of Transportation.$ 3. 60 passenger vehicle built for DOT by Rohr Corporation [211]. A linear induction motor built by Garrett was later added. pages 154-155). Similar designs were pursued at Tufts University. one from Aeroglide Incorporated.

S.S. But by 1972. projects were administered by the Federal Railroad Administration under the Electric Power and Propulsion for High-Speed Tracked Ve­ hicles program. and related research was also carried out at Rensse­ laer Polytechnic Institute and Tufts University. efficiency consider­ ations and the cost of constructing tunnels had caused emphasis to be shifted to vehicles operating in an open guideway (see [154]. the high-pressure air under the vehicle was prevented from escaping the guideway by means of flexible winglets mounted on the vehicle (winglet seal).A.). 63 A related project was the UTACV (urban tracked air-cushion vehicle). Between 1968 and 1978 research into a High Speed Ground Transportation sys­ tem with aerodynamic lift received low-level funding from the U. which appears to be an ACV that resembles a car and travels in a simple trough made of either compacted . Initially. De­ partment of Transportation) under the Transportation Advanced Research Projects (TARP) program [72].S. Princeton University. [71] and [211] for example). contracted out the project to Rohr Industries. Massachusetts. a branch of the Federal Railroad Ad­ ministration. The Urban Mass Transit Administration (UMTA). The author speculates that the UTACV project was absorbed by the TACV project. which was to operate on an elevated guideway with a vertical metal guide at the center. Mitre Corporation (McLean. Virginia. These vehicles were wingless lifting bodies that flew in a guideway. U. Early research centered on tube-flight vehicles. In 1974 Bertelsen Incorporated introduced its aeromobile. Department of Transportation (DOT). as described above (see [157] and [72]).A.). U. and later by the Office of the Secretary of Transportation (OST. These winglets sometimes featured large dihedral angles for reasons of roll and sideslip stability. Contractors included the Transportation Systems Center (TSC at Cambridge.

The author did not learn whether this craft ever reached the operational stage. One interesting feature of this train is its telescoping mechanism. page 203). the TALAV. a tracked hovercraft with a LIM was scheduled for first trials in 1971 near Ely (see [205] and [168]. page 4). Technical introduction to maglev and linear induction motors Magnetic levitation (maglev) refers to any object that hovers freely above the ground by means of a magnetic cushion. parts of the train move with respect to one another to create access for passengers or cargo (see [171]. 64 soil or concrete (see [219]. However. It is reported to have reached 322 km/h (200 mph). The magnetic cushion can be generated by either permanent magnets or by . the British project was cancelled in early 1973 [211]. One disadvantage of trains operating on guideways is the difficulty of switching tracks and of turning the train around. page 175). Magnetic Levitation (Maglev) Trains and Linear Induction Motors (LIMs) This section comprises a technical and historical introduction to magnetic levi­ tation trains with linear induction motor propulsion. In the United Kingdom. Brazil introduced its 20-seat tracked air cushion vehicle. It is thus also known as magnetic suspen­ sion. Many of the designs featured air-cushion suspensions (such as hoverpads) and linear induction motors (see [208]). In 1981. In the early 1970s much research was done on PRT (personal rapid transit) vehicles.

Electromagnets are electric conductors that only develop a magnetic field when they are supplied with electricity. a magnetic suspension can either be attractive or repulsive. jet engines. while the stationary track with its coils/windings is a flattened version of the electric motor's stationary exterior magnets or windings (called the stator). There are six major types of magnetic suspensions. Maglev trains obviously also need propulsion systems to move them along the track. which describes the mutual interaction of electric and magnetic fields. the vehicle wraps around the guideway and its bottommost part is magnetically lifted up toward the guideway. Possible propulsion systems include propellers driven by combustion engines. the vehicle underside is a flat version of the electric motor's rotating central winding (called the rotor). most commonly. In an attractive mag­ netic suspension. Permanent Magnetic Suspension (PMS). and. Maglev trains differ in the type of magnetic suspension and in the type of propul­ sion system they feature. The first. Linear Induction Motors (LIMs) are often visualized as electric motors that have been unrolled and flattened. . there are three further cate­ gories of magnetic suspensions. There is a great variety of linear motors. First. The advantage of attractive suspen­ sions is that the vehicle cannot derail since it grasps the guideway. The advantage of repulsive suspensions is that the vehicle is inherently sta­ ble (the magnetic repulsive force is proportional to the gap between the vehicle and the guideway). In a repulsive magnetic suspension. the vehicle's underside and the guideway's top side repel one another. 65 electromagnets of various designs. linear induction motors. Beyond being described as attractive or repulsive. Electromagnets owe their magnetism to the phe­ nomenon of electromagnetism.

this suspension only functions when the vehicle is in motion. Returning to the linear motor's description as an "unrolled" electric motor. or forward propulsive . the guideway is the active part of the motor. only the different types of linear induction motors will be described here.2 degrees Celsius. On the negative side. Superconducting magnets are cooled to temperatures that are within thirteen degrees of absolute zero (0 degrees Kelvin. or -459. Electrodynamic Suspension (EDS). which is an advantage of EMS. Permanent magnets are metallic objects that have a permanent magnetic charge without having to be supplied with electricity. The third. The second. It contains the electrified magnets that produce the thrust. Turning now to propulsion systems for maglev trains. Electromagnetic Suspension (EMS). and that it allows for larger gap spaces. -273. Ad­ vantages of EDS are that it requires less energy than EMS. requires superconducting magnets (SCMs) on the vehicle and ferromagnetic (conducting) strips or coils on the guideway.7 degrees Fahrenheit) in order to maximize their capacity to conduct electricity. uses com­ binations of electromagnets and ferromagnetic (conducting) strips or coils to produce interacting magnetic fields between the guideway and the vehicle. one differentiates between long-stator and short-stator linear motors. 66 uses permanent magnets on both the track and the vehicle. The lifting forces keep the vehicle afloat even when it is not moving. As the name electrodynamic implies. and magnetic radiation levels in EDS vehicles are unacceptably high. superconductivity is difficult to achieve. One alternative under study is to use high-temperature superconducting magnets that are cooled with liquid nitrogen instead of the expensive liquid Helium used in superconducting magnets. With the long-stator linear motor.

the vehicle is the active part whereas the guide- way is passive. Cross sections of three of these maglev trains are shown in Figures 4. who in 1880 experimented with permanent magnets in repulsion. 67 force. i. The electricity supplies the vehicle-mounted magnets that pro­ duce the thrust and that make the vehicle comparatively heavy. and 4.8. With short-stator lineax motors.7. A great advantage of the long-stator system is that the propulsion coils can be made larger on sloping segments of the guideway to provide more thrust where it is needed. The guideway generates a traveling electromagnetic wave which reacts with vehicle-mounted field windings to propel the train.9. The passive vehicle acts as the rotor and weighs comparatively little. A brief history of maglev technology Based on the invention of the electric motor (or. In the early 1900s a system was proposed based on the attraction of permanent magnets on a vehicle to ferromagnetic plates on the guideway. to keep the vehicle at a constant distance from . the electric rotary induction motor) in 1821. 4. more precisely.. But it proved difficult to control the magnetic fields to achieve height stability. William Sturgeon built the first linear induction motor (LIM) in 1832. The idea of magnetically suspending a vehicle over a track is due to Earnshaw.10. Reversing the polarity of the magnetic field provides contactless braking.e. Construction of the powered guideway requires great precision and is expensive. The vehicle picks up electricity from an overhead catenary or from wayside power rails. The various combinations of magnetic levitations and linear motors are depicted in Figure 4.

7: Combinations of magnetic levitation and linear motors . Maglev Trains with LIM Propulsion EMS. EDS. Attractive Repulsive Suspension Suspension crs CO Long-Stator Short-Staior Long-Stator Short-Stalor LIM Propulsion LIM Propulsion LIM Propulsion LIM Propulsion (Active Guideway) (Active Vehicle) (Active (luideway) (Active Vehicle) Figure 4.

wayside power rail levitation electromagnet (current collector) armature \\ (ferromagnetic plate.8: Short-stator EMS train with LIM propulsion . (passive) guideway passive rotor) AV ground Figure 4. 69 passenger cabin vehicle body upper frame sprung mass linear traction (short stator) secondary pneumatic motor suspension guidance &: levitation undercarriage ferromagnetic maglev bogie strip unsprung mass guidance &.

. lifting electromagnet) ground Figure 4.9: Long-stator EMS with LIM propulsion. . 70 passenger cabin vehicle body vehicle upper frame sprung mass secondary pneumatic mechanical suspension composite support skid idance electromagnet guidance ferromagnetic strip undercarriage lifting ferromagnetic strip maglev bogie pro'pulsion electromagnet unsprung mass . . similar to German Transrapid TR-07 . (long stator) generator windmg/coil levitation frame vehicle underframe magnetic primary suspension (active) guideway (rotor.

71 / •• \ /\ / \ /\ superconducting magnet for traction &.10: Short-stator EDS with LIM propulsion. guidance pneumatic wheel J ) (short stator) superconducting lifting magnet lifting ferromagnetic strip armature (ferromagnetic plate. similar to Japanese MLU-001 . passive rotor) Figure 4.

They proposed and designed the first EDS system in 1966. In the United States. Between 1969 and 1993. An early result of that investment was the work of Gor­ don Danby and James Powell at Brookhaven National Laboratory. 72 the track. When population pressure and automobile pollution were recognized as serious problems in the 1960s. which meant that the strength of the magnetic field could be adjusted by changing the current supplied to the electromagnets. In 1936 he proposed an evacuated tube maglev system. He used electromagnets instead of permanent magnets. The 1960s brought a change in attitude and a renewed interest in public transportation systems. He took out a patent in 1934. Kemper became known as "father of electromagnetic levitation". World War Two and the optimistic decades that followed it prevented further progress in the area of magnetic levitation. Motivated by mounting traffic problems. As is often the case.8 biUion (US$ . Western gov­ ernments acted to find solutions to mass transportation. the BMFT spent DM 1. This included magnetic levitation and linear induction motor propulsion. What follows is an outline of the various accomplishments in maglev and LIM technology. BMFT) awarded contracts to a number of companies to research the various maglev and LIM technologies in 1969. the maglev and LIM technologies lay idle until the social and political climate was ripe for their utilization. and introduced a working model in 1935. the High-Speed Ground Transportation Act of 1965 resulted in funding for research in the areas of maglev and LIM. the West German Ministry of Re­ search and Technology (Bundesministerium fur Forschung und Technologic. The control problem was solved in 1933 by Hermann Kemper in Germany.

introduced the first short-stator EMS model. and both utilized attraction suspensions. The year 1974 was a milestone in Germany's research into maglev and LIM systems. who introduced their Tran­ srapid TR-04 train.4 km long guideway at the Krauss-MafFei works near Munich. Krauss-MafFei and Messerschmitt Boelkow- Blohm (MBB). To test the stability of an EMS at high speeds. It was capable of a speed of 157 mph (250 kph). Krauss-MafFei Corporation of Munich. The rights for that system were later acquired by Boeing Company and by Carnegie- Mellon University. By 1971. Japan began funding re­ search into maglev and LIM systems. These were the world's first full-scale maglev vehicles. research began on an EDS-type system. . weighed 18. In the wake of the 1965 High-Speed Ground Transportation Act. 73 1 billion) on the two technologies. In 1970. MBB had mounted a rocket motor on its unmanned "Komet" vehicle. Driven by population and pollution pressures as well. In 1969. and operated on the elevated 2. It had a length of 15 m. The first full-scale long-stator EMS system was introduced by Thyssen- Henschel.5 tons. Rohr Industries advanced an urban maglev system called ROM AG. two German companies. had built full-scale short-stator EMS vehicles that achieved 45 niph. Research into EMS systems began in 1975. It reached 250 mph (401 kph). Germany. At the same time. Its "HMB2" vehicle achieved 20 mph on its 100-meter track. the Federal Railroad Administration of the United States awarded contracts to Ford Motor Com­ pany and the Stanford Research Institute for research on EMS and EDS systems. Meanwhile work had continued on short-stator EMS systems at Krauss-MafFei.

with a length of 26 m (85 ft). The result was the Transrapid TR-05.000 kg). and low magnetic fields of that system as factors in its decision. resulting in repulsive forces. where the Federal Railroad Administration (FRA) had built a test track near Pueblo. low environmental effects. International Traffic and Transporta­ tion Exhibition) in Hamburg. It was a two-section vehicle for 75 passengers. the Tracked Air-Cushion Research Vehicle (TACRV) was tested at this site at the same time. the German BMFT selected attractive long-stator EMS technology as the most effective technology for high-speed public ground transportation. The vehicle's superconductor was a hollow tube with coolant at the center. Their prototype "LIM Research Vehicle" was powered by a LIM built by Garrett Turbines and achieved 234 mph. In 1977. The guideway was an aluminum trough in which eddy currents were induced by vehicle-mounted superconducting magnets.). As a sidenote. 74 The same year also brought a breakthrough in the United States. It cited the comparatively low capital costs. It awarded Thyssen-Henschel a contract for constructing a public demonstration train for the 1979 Internationale Verkehrsausstellung (IVA. Some sources claim that it reached 255 mph (410 kph). Inc. and a weight of 40 tons (36. Other efforts in the United States in the mid-1970s included models of the Magneplane. which was a repulsive EDS system based on the 1970 patent of Henry Kolm of MIT (later the head of Electromagnetic Launch Research. The cylindrical vehicle was stabilized by a magnetic keel and rolled in curves. low operating costs. resulting in a very efficient cooling system. But all these efforts came to an abrupt end when the US-American administration under Gerald Ford stopped funding HSGT research in the 1975. Colorado. The Hamburg . which was the first maglev train for passenger service.

In 1980. five years of Japanese research into short-stator EMS systems resulted . ML-500. The EDS train was also known as the MLU. 75 track was a 900 meter long elevated guideway. It utilized a guideway-mounted LIM (long-stator propulsion). nine years of government-sponsored research into EDS systems culmi­ nated in the introduction of the unmanned MLU-001 test vehicle by the Japanese Railways Group (JRG) and the Japanese National Railways (JNR). There were tentative plans to build a line link­ ing Tokyo and Osaka by 2001. Krauss-Maffei. and Thyssen-Henschel. Siemens. including AEG. Strong magnetic fields in the passenger cabin were a problem for the system. These problems led to a change in the design of the guideway cross section from an inverted T-shape to a U-shape in 1979. The TR-05 was eventually moved to Thyssen-Henschel's plant near Kassel for further testing. In 1987. BBC. The gap space between the vehicle and the guideway was ten centimeters (4 in). Construction of a new test track was possibly begun in the Yamanashi prefecture. The TR-05 reached 96 kph (60 mph). A 40-km segment of the line would cost USS 2 billion. MBB. A consortium formed to undertake these tasks. so that the BMFT decided to fund further research. as were difficulties with the supercooling system and with negotiating turns. In 1979. transported 50. including a larger train and a test track in the Emsland region of Northern Germany.000 visitors. It involved sev­ eral German companies. The prototype weighed ten tons and was 44 ft long. It achieved a speed of 517 kph (321 mph) on its seven-kilometer test track in the Miyazaki prefecture. and was well-received. and MU 001. not least because 80% of the segment would involve tunneling. the project was taken over by the Japanese Railway Technical Research Institute (RTRI) and funded by federal and local governments as well as by the newly privatized railroads.

Battelle Memorial Institute. 76 in the introduction of the HSST (High-Speed Surface Transport) by the Development Department of Japan Airlines (JAL). In the United States. Thyssen Industrie. The train reached 125 mph on its test track. Some of these companies formed consortia including American MagLine Group. Hughes Aircraft. Ottawa. and possibly St. General Atomics. Portland. Bombardier. the feasibility studies . Philadelphia. Houston. Bechtel Group. Vancouver). so there were only numerous feasibility studies. Intermagnetics General. Worth. The single-car HSST carried passengers around the Expo '85 in Tsukuba. Dupont. San Francisco. Companies and institutions involved in these various studies include AEG Westing- house. and Texas TGV. Texas Instruments. Las Vegas). Carnegie-Mellon University. it is called the Empire corridor. the Florida Corridor (Miami. Los Angeles. the Northwest Corridor (Seattle. the California Corridor (San Diego. CSX Corporation. Baltimore. Parsons Brincker- holf Quade and Douglas (PBQD). Union Switch and Signal. There was also one Canadian corridor under consideration (Montreal. Corridors for considera­ tion for maglev and conventional high-speed trains were the Northeast Corridor (New York. Washington. and/or Cincinnati). Booz Allen and Hamilton. San Antonio). the Council on Superconductivity for American Compet­ itiveness. All American Magneplane Incorporated. Boston. and possibly Pittsburgh). Grumman. Columbus. Asea Brown Boveri (ABB). Argonne National Laboratory. Milwaukee. and the Midwest Corridor (Chicago. Maglev 2000. Lockheed. General Dynamics. Japan. General Electric. Maglev Transit Incorporated (MTI). Maglev USA. Minneapolis. USX. Austin. GEC Alsthom. Cleveland. Siemens. there was no real support for maglev and LIM research in the 1980s. Las Vegas-Anaheim Maglev. In terms of agencies. the Texas Triangle (Dallas/Ft. Foster-Miller. Louis. Orlando. Alcoa. Tampa). Toronto).

two loops.000 lb). the first short-stator EMS train entered public service in the United Kingdom. It was a low-speed system connecting the Birmingham airport to the Birmingham railroad station. In German it is called the Transrapid Versuchsanlage Emsland (TVE). 77 involved the United States Department of Transportation (DOT). The gap space between the 40-passenger vehicle and its guideway was 15 mm. MBB had pressed on with its EDS research. Army Corps of Engineers (USACOE). In 1984. a height of 3. A year later.11). The test facility for the Transrapid system in the Emsland region (near the river Ems) in the German state of Lower Saxony was finished in 1987. the Transrapid TR-06. and several state governments and Departments of Transportation. or 224. In 1981 it introduced its long-stator "EET 01" which reached 160 kph (lOOmph). or Transrapid Test Facility Emsland. In June of 1988 on the occasion of the Hamburg International Traffic and Transportation Exhibition.000 kg. and S-shaped switches on a rotating disk were utilized (see Figure 4. Each section had a length of 27.4 m (90 ft). in 1985. as had the newest long-stator EMS train. parts of the Emsland test facility in Germany had been completed. The TR-06 reached 257 mph (412 kph) at the Emsland facility in December of 1987. the TR-06 made 358 public demonstration runs.65 m (12 ft). two switches.8 m (10 ft).600 passengers. and a weight of 65 t (59. Testing of the TR-06 on the 13-mile guideway segment began in 1984. Department of Energy (DOE). It featured a length of 40 km (25 mi). The vehicle had two sections accomodating 200 passengers and was built by Krauss-MafFei. carrying 16. Despite the fact that the German BMFT had decided in favor of the short-stator EMS system. a straightaway between . a width of 3.

Segments of the guideway were . which was a partnership of the Deutsche Bundesbahn (DB.5 m (84. German Federal Railways). In 1989. a width of 3.5 tons). or about 100 passengers).11). were A-shaped or H-shaped concrete columns. It was electrified by a 110.7 m (ft). or support structures. The design of the other switch involved two wheels and an electromechanical actuator (see Figure 4. The shape of the vehicle had been optimized aerodynamically and aeroacoustically. The piers. Lufthansa (German Airlines).000 Volt line. implying that it had a guideway-mounted LIM and that it utilized elec­ tromagnets to achieve levitation even at zero speed. and could accomodate a payload of 16 metric tons (16. the first maglev train was certified and entered revenue service (as opposed to public demonstration service) in Japan. Both switches consisted of a steel track segment that was bent and allowed for speeds of 125 mph in the branched position. Each section had a length of 25. Direction of the project was now under the High-Speed Surface Transportation (HSST) Corporation instead of Japan Airlines (JAL).2 tons. The train was comprised of multiple coupled cars with nose sections at each end. The track itself was made of either welded steel or concrete spans. an elevated guideway consisting of segments ranging from 25 to 50 meters in length.12). with an elevation of 5 m (16. The TR-07 was a short-stator EMS system. In the same year (1989). and lABG (a federal vehicle testing organization).5 ft) and piers every 25 m (82 ft). a weight of 45 metric tons (45. The test facility was operated by MVP. For one switch.2 ft). the Transrapid Consortium introduced its TR-07 "Eu- ropa" (see Figure 4. glide bearings and hydraulic actuators were used. The short-stator EMS "HSST-05" reached 80 kph (50 mph) on its demonstration guideway at the Yokohama Exhibition Site. 78 the loops.

79 Rotating switch (top view) Hydraulic switch (top view) / / / / ' / hydraulic / actuators / Figure 4.11: Types of Switches .

80 supplied with a 26. In March of 1994 the German cabinet approved construction of a maglev line connecting Berlin and Hamburg. variable frequency (VVVF) current required to control the vehicle. which is a distance of 284 km (180 miles).000 Volt DC current when the train passed over them. By 1994 it had reached 450 kph (280 mph). However. and propulsion during emergencies. and the cooling fans. If . Propulsion and braking were controlled by varying the excitation voltage and frequency of the linear motor. The control system adjusted the current supplied to the electromagnets 30. and were recharged by lin­ ear generators on the undercarriage during normal operation and motion. the electrical systems. Between 1989 and 1994 the Transrapid has covered 160. The linear generators also supplied the air-conditioning. The TR-07 reached 432 kph (271 mph) in December of 1989. the economic difficulties following the German reunification and other factors prevented this from happening immediately. This cleared the way to the possible construction of a revenue line in Germany. The con­ trol system is crucial in that attractive levitation comprises an inherently unstable system. 440 Volt on-board batteries supplied the lifting and guidance magnets while the vehicle was not moving. technical supervision department) of the German region of Rheinland.000 times per'second to ensure a precise gap space of one centimeter (3/8 in). guidance. They also provided levitation. In 1991. the Transrapid TR-07 "Europa" achieved certification for revenue ap­ plication by the Deutsche Bahn (German Railways) and the Technischer Ueber- wachungsverein (TUV. There were wayside power substations that converted the three-phase utility power into the alter­ nating current (AC) variable voltage.000 km during test runs. But an appropriate bill still needs to pass both houses of parliament in late 1994.

12: German Transrapid TR-07 . 81 Figure 4.

The plan involves an unusual financing scheme. Lufthansa and Deutsche Bahn DM 300 million. One was the STARUM. including US$ 725 million for a maglev prototype. That money would be used to build the trains and to begin operations.3 billion.and medium-speed maglev train that utilized . 82 approved. and to rent it to the private operations consortium that operates the line and owns the trains.5 billion. The Intermodal Surface Transportation Efficiency Act (ISTEA) was passed in 1991. A further criticism is that the maglev train would not go to the city centers. The other was a low. Opponents of the plan claim that the system would not be profitable. which would pay approximately DM 310 million per year of guideway rent to the government. The entire project is expected to cost DM 8. The federal government expects to build and own the DM 5. promising over US$ 1 billion in support for HSGT. the operations consortium would be able to obtain loans amounting to a capitalization of DM 3. But the funding never materialized and in 1993 only US$ 13 million were allocated. There were at least two more maglev efforts for which the author does not know the dates. Starting with a private capital of DM 1. especially on a relatively short distance which could be covered by Germany's ICE train in just a bit more time.6 billion guideway. All profits from the line would go to the private operations consortium. Siemens. and several construction companies. These firms will furnish DM 500 million. The operations consortium is made of Thyssen-Henschel. Daimler- Benz/AEG. a Franco-German EMS maglev with short-stator LIM propulsion.9 billion (US$ 5.3 billion). the line will be comleted in 2004. banks a further DM 200 million. The early 1990s continued to be a poor time for high-speed ground transporta­ tion (HSGT) research in the United States. and other private investors DM 500 million.

At all speeds. and even more in urban areas. weaknesses. ten times that in suburban areas. 83 PMS technology. It was developed by Germany's Magnetbahn Gmbh. At a distance of 25 m. In particular. To give the reader an idea of noise levels. Maglev versus high-speed duorail trains In this section some of the strengths. and costs of various high-speed ground transportation systems are discussed. which is a subsidiary of AEG and is also known as M-Bahn. maglev trains are as noisy as a jetliner during takeoff at a distance of 200 m. [236]. some estimates from a 1990 study [1] follow. and [237]). [200].000 per mile in rural areas. The guideway makes up 70% of the . whereas maglev trains are only as noisy as an automobile during city-driving. The energy use and noise levels of several transportation systems are shown in Table 4. at a speed of 400 kph. high-speed duorail trains are as noisy as a truck. an additional US$30 million per mile is required. If tunneling is necessary. [204]. The cost of operating any such system includes energy costs and environmental costs. However. consider the different vehicles at 160 kph. To give the reader an idea of the cost involved in building a maglev or conven­ tional high-speed train line.2. [213]. Construction of the double-guideway would cost US$30 million per mile. maglev trains use between 30 and 40 percent of the total power for levitation. The data for this table were collected from a number of sources ([18]. maglev trains use less energy and produce less noise than high-speed conventional duorail trains. To purchase a 60 ft wide double track maglev corridor costs US$75. Opponents of maglev systems often cite the high development and construction costs and its use of electricity as its major drawbacks. [233].

84 Table 4.25-0. 80-86 dB/A U-11 dB/A 25 m distance energy use 10 Wh 30 seat—Km seat—km at 250 kph 2000 seat—mzle noise at 250 kph at 86-92 dB/A 82-83 dB/A 25 m distance energy use at 400 kph noise at 400 kph at 92-96 dB/A 25 m distance .2: Energy use and noise levels of transportation systems high-speed ground duorail maglev auto­ effect trains trains mobiles vehicles number of seats 480 200 4 average occupancy 70% 40% weight per passenger 1450 lb 500 lb payload qrossweiqht 0.5 24 gal/mile 40-50 HP/ton energy use orm JJCtSS TTXltl&S 75-200 HP/ton no HP gal juel ton—km at 100-160 kph 0.01—22' g g ton—miles seat—mile gal 03 Wh 19 seat—km 0 1 kWh o n ton—km seat—km seat—mile kg fuel noise at 100-160 kph.

from appendages. whereas the vehicles only account for about ten percent. and it became a foremost criterion in the design of the German ICE [224]. It is also called aeroacoustic noise. Both the designs of the Japanese Shinkansen and the French TGV took this fact into account. the skin friction drag (27-40 %). and from the separation region at the rear of the vehicle. The aerodynamic drag consists of the nose and tail pressure drag (8-13 %). A typical 80-seat vehicle costs US$3. including noise from the turbulent boundary layer. For maglev trains the interference drag due to the interaction of the maglev bogie and the guideway accounts for 50-65 % of the total drag [224]. the noise from the cooling fans dominates for speeds up to 200 kph. It is clear that proper aerodynamic design is of great importance . Beyond that. the interference drag due to the interaction of the underside and bogies with the ground (38-47 %). and is mainly in the frequency range of 250-500 Hertz. For maglev trains. it represents between 75 and 80 % of the total resistance to motion of conventional trains at speeds above 250 kph. The other contributors to the resistance are mechanical rolling resistance and transmission losses. 85 total cost of a typical line. from the interaction with passing poles. Aerodynamic noise dominates at speeds above that. the parasite drag due to protruding parts and inter-car gaps (? %). Aerodynamic noise is caused by many factors. For conventional trains the rolling noise dominates for speeds up to 300 kph. the aerodynamic noise dominates. from gaps and joints. the pressure drag due to the pantograph and roof equipment (8-20 %). Turning now to aerodynamic drag. from edges. Aerodynamics plays a vital part in any high-speed ground transportation system because it causes drag and noise. from the airflow underneath the vehicle. and the side wind drag.6 million.

It is for these reasons that an aerodynamically suspended train seems like a plausible solution. in the late 1960s. the width of the guideway is constrained to a few meters. Such a train could be suspended from a monorail and have wings that are operating out of ground effect. as suggested by Lehl and Zumwalt in 1974 [215]. of supercooling. which means that the wing area is quite limited. Trains using aerodynamic suspension It is clear from the preceding discussion that in order to arrive at an economical transportation system one must avoid the cost of close tolerances (such as precise guideways or high-speed tracks).I. 86 for high-speed transportation systenns to reduce both drag and noise. This in turn implies a constraint in the lifting force generated by the wing. This aerodynamic lift would be a by-product of the vehicle's forward motion and would not require separate lifting-fans or levitating magnets. it could have one or more wings operating in aerodynamic ground effect just above the guideway. Alternatively. The disadvantage of the system is that its size is quite limited. emphasis was shifted to ground effect trains operating in a trench-like guideway (see the work of Widnall and .T. thus eliminating mechanical contact. of electrification. much energy could be saved if levitation could be accom­ plished without lifting fans or magnets. The idea of a train flying in ground effect was apparently first proposed in con­ junction with tube-flight research at Princeton and M. In addition. and of machine maintenance. the gap between the trailing edge and the guideway should be at least ten centimeters. To allow for misalignments in the guideway. Due to high vehicle drag and the high cost of building tunnels. It is quite possible that the resulting vehicle will be too small to be competitive.

high lift-to-drag ratio (a measure of the aerodynamic efficiency). good range. protects the blades from foreign objects and reduces the danger of injury. A related idea was the "Hydrodyne". The shroud reduces the noise level. for example). Two drawbacks of this system were speed limitations due to cavitation and the high loading of the tube. Based on the works of Lippisch and the proposed design of Piasecki's Airtrain [228]. The train should be optimized for the following: ease of construction. or rectangular guideway. good fuel economy. which was mentioned by Lissaman in 1968 [196]. Lehl and Zumwalt [215] suggested that their "Airtrain" could also be flown in a trench. In this context. This vehicle was also studied by other researchers ([210]. The single fuselage was suspended from the centers of several straight rectangular wings. high payload. In 1974. The "Hydrodyne" was a vehicle that seemed to hover over two small tubular rails. [216]). easy access (load- . ducted fans may be the desirable propulsion system for an aerodynamically suspended train. the fuselage and each wing were connected by vertical empennage-like structures that moved through a gap in the middle of the hor­ izontal guideway (see Figure 4). providing enough lift to support the train. They produce more thrust per engine horsepower than an open propeller. mechan­ ical simplicity. hydrofoils operated inside these water-filled tubes. Davis. In 1970. 87 Barrows [199]. In reality. Ducted fans offer quite a few advantages compared to free propellers and jet engines. and Martin [202] conducted experiments on an airfoil operating between parallel walls. Harris and Seemann [206] proposed the "Ter- rawing" or "Terrafoil" which featured a wing flying just above an elevated guideway. ease of maintenance. so that one may effectively reduce the propeller diameter. Williams. to take advantage of the ground ef­ fect. In this overhead suspension.

88 Figure 4.13: Terrawing train .

engine failure. 89 ing/unloading. objects on the guideway. emergency exits) and safety (gust resistance. . emergency stopping and unloading).

that is. reflection principle. so that a plane of symmetry is expected to exist . In the following subsections. or image-model method. The aim of any experiment should be to model the actual flow as closely as possible. due to wind. To properly model the ground. The model and its exact mirror-image are placed into a wind tunnel. Kuchemann of 1978 [239]. One may also have to consider that in reality there usually exists an atmospheric boundary-layer that is due to the relative motion of the airstream and the ground. The test apparatus may therefore need to allow for relative motion between the ground and the airstream in the lateral direction. there should be relative motion between the vehicle and the ground. A SURVEY AND CRITIQUE OF TECHNIQUES USED IN GROUND EFFECT EXPERIMENTS There are six major experimental methods for studying the effect of ground vicinity upon the airflow about vehicles. reflection- model technique. and Howell and Everitt of 1981 [240]. the six experimental techniques will be summarized and critiqued in turn. General discussions and comparisons of the various experimental methods can be found in the works of Pistolesi of 1935 [238]. Mirror-Image Method This method is also known as image technique. 90 CHAPTER 5. with special attention to the testing of airfoils.

three problems begin to surface. [244]. [245] and [246]) biplane theory. The third and final problem with this method is that it requires two exactly similar models that are perfectly aligned. for small ground clearances. it was not utilized in an experimental setting until 1920. The theory involving the use of an image wing was an extension of Prandtl's (see [241]. One way to largely avoid this problem is to place a thin splitter plate at the symmetry plane near the trailing edge. He included pho- . and it was developed by Betz at Gottingen in 1912 [42] and 1914 [247] and. at about the same time. when a vehicle moves through still air the freestream velocity and the ground velocity both equal zero. destroying the symmetric flow. by Wieselsberger at Gottingen in 1921 [45]. First. the air between them is accelerated or decelerated. by Raymond [46] at the Mas­ sachusetts Institute of Technology. Two years later. the wakes behind the two models begin to interfere with each other. By contrast. more thoroughly. thereby preventing the wakes from interacting. The result is an unsteady wake flow involving turbulent mixing (see Miiller's paper of 1939 [54]). The image wing was first used experimentally by Cowley and Lock in the United Kingdom in 1920 [43] and. 91 between them. In 1939. The plane of symmetry acts similar to the ground for the case of large ground clearances. At the plane of symmetry the flow velocity is thus different from the freestream velocity.1). Even though the theory of this reflection principle was developed in the 1910s. [243]. Miiller [54] studied airfoils and cars in a wind tunnel. which is difficult to achieve (see Figure 5. [242]. However. This then eliminates the need for a physical representation of the ground. These tests involved large ground clearances. Wieselsberger performed similar experiments on a three-dimensional wing [45]. The second problem is that as the two wings approach one another.

92 mam wing symmetry plane optional splitter image wmg plate Figure 5.1: Mirror-image method .

In the early 1930s. Miiller's work of 1939 [54] also dealt with wings with small gaps. Some of the significant experimental works are those of Lay [248] in 1933. the fixed ground-board method and the moving-model method. there was a strong interest in automobile aerodynamics. but he focused on the visualization of the flow around two-dimensional profiles. Some decades later. Datwyler's model wings had an aspect ratio of four and trailing edge ground clearance ratios equaling 0. of Heald [249] in 1934.he therefore selected the reflection method for his second set of experiments at Zurich. The ground clearance ratios for cars are very small. Lay used the mirror-image method as well as a fixed ground-plate with and without suction to control the boundary-layer.200. near 0. In the same year. Gallington [70] used the image method to study the channel flow underneath his ram-wing vehicles.025.100 and 0. of Ono [250] in 1935 and of Muller [54] in 1939. . 0.05. This included the flow in the gap formed between the ground and the car's underside. they were much better than the previous set. Ono used the mirror-image method. His first set of experiments at Gottingen involved a ground-board and did not give satisfactory results . Most of these authors compared several experi­ mental techniques. Their report is very detailed and entails pressure distribution data. Similarly. the 1961 NACA report by Fink and Lastinger [60] is rich in detail. 0. Serebrisky and Biachuev used the image-wing technique in a wind tunnel of the Hydrodynamic Institute in Moscow [52].050. Not until the work of Datwyler at Gottingen and Zurich [48] in 1934 were smaller gaps were studied. Even though the results still did not agree very well with theory (he used conformal transformations to map a flat plate whose trailing edge was in contact with the ground into a simpler computational domain). 93 tographs in iiis publication.

[247]). a flap is often added to the ground-plate's trailing-edge to control the flow over its leading edge. Fukuchi. To reduce these error sources. especially if the boundary layer separates. whereas the tunnel walls. reflector plate method. Oftentimes the model is mounted at a distance which takes into account the boundary-layer thickness. or fixed ground technique. One may account for the presence of the boundary-layer by calculating its displacement thickness and by thereafter adjusting the value of the ground clearance. 94 The mirror-image method has been used for the study of train aerodynamics.01). and a blockage is created (see Figure 5. the model. and Yamamoto in 1968 [251]. The errors will be particularly noticeable for small ground clearances. the ground clearance is extremely small (near 0. Fixed Ground Method Tfe fixed ground method is the most popular method and is also known cis fixed plate method. reflection plate method. This causes inaccuracies since an actual vehicle moves with respect to the ground. In this case. ground board method. cambered plates. and the ground-board are stationary. only the air moves. His experiments involved finite. Ray- .2). In additionm. the model causes a circulation around the ground-board. Also. The fixed ground-board was first used experimentally by Betz at Gottingen in 1912 ([42]. Kawaquiti. For trains. This eliminates the circulation. The Japanese Tokaido-line trains were tested this way by Hara. In 1921. flat plate method. The flow over the fixed ground-board develops a boundary-layer which alters the flow field. the plate is kept very thin (to prevent blockage) and as short as possible (to minimize the boundary-layer thickness).

He found . 0. a flow velocity of 30 m/s. 0. Mod­ els of entire aircraft were tested by Tani. Datwyler's first set of experiments at Gottingen involved a ground- board. Datwyler sought to investigate finite wings in very close ground prox­ imity.200. This corresponds to trailing edge clearance ratios of 0.7 x 10®. In 1934. Itokawa and Taima in 1937 [254] and by Sears a year later [50]. with equaling 5. Inl939. Stalker [253]. and a Reynolds number of 3.025.2: Fixed ground-plate method mond [46] tested three-dimensional wings mounted over a fixed ground-board.20 and 40 millimeters. Ono [250] and Muller [54]. This included Heald [249]. In the 1930s several authors studied the aerodynamics of car models. Klemin [252]. 95 wing ground plate movable flap Figure 5. and with c = 200 mm and b = 800 mm.050. Lay [248]. Recant [53] used the fixed ground method to examine three-dimensional wings in ground effect.10.100 and 0.

The fixed-ground method is still used extensively because it is inexpensive and . These problems become worse as clearance tends to zero. This method was first implemented by Lay in 1933 [248] and by Stalker in 1934 [253]. Another problem is that the boundary-layer control has to be readjusted for every model and every test geometry. Datwyler improved his results when he switched over to the reflection method. In 1934. In 1970 Barrows and Widnall [255] investigated a stationary model of their tube flight vehicle in a circular wind tunnel test section. the fixed-ground method was improved through attempts to control the ground-board boundary-layer. the combined action of sources and sinks results in locally circulatory flow. A flow visualization study of airfoils in close ground effect was carried out by Miiller in 1939 [54]. He also provided allowance for the thicknesses of the boundary-layers on the airfoil and on the ground-board in order to arrive at a true inviscid flow geometry (also see Bagley's related work of 1960 [57]). Also. It can be controlled by local suction (removal) and tangential blowing (energization) or by simply bleeding it off (removal). The method is therefore also known as distributed-suction technique or bleeding-technique. In 1961. Removal of the boundary-layer decreases the danger of flow separation and the associated rise in drag. They were studying automobile aerodynamics. Bagley [58] added a thin leading edge and a controllable trailing-edge flap to the ground-board in order to avoid a circulation around it. 96 that the flow-field ceased changing very near the ground since the boundary-layer that developed on the ground board retarded the flow near it. Implementation of this method is very difficult because the normally smooth surface is interrupted by either the sources and sinks (in the case of distributed suction) or by the protrusion/slot (in the case of bleeding/skimming).

By contrast. It is also difficult to avoid belt deformations as a result of the pressure distribution. as it is stationary (see Figure 5. It is not possible to model a crosswind in a controllable fashion. The boundary-layer on the fairing in front of the belt can be skimmed off by means of a slot. 97 provides good qualitative information. Moving-Belt Technique The moving-belt technique is also known as the moving-ground method or the moving-ground-board method. Most of the time the belt and its fairing protrude into the tunnel. flapping and belt roughness. as well as gaps between the belt and the tunnel wall. disturbing the tunnel's geometry and causing some blockage. This is especially true for models of cars and trains (see for example Turner [256]. To control some of the above problems one can use suction to hold the belt flat to a guide plate. and Grunwald [259] for train models). The belt's speed is often measured with a stroboscope. [257]and Carr [258] for automobile studies. A conveyor-type belt is moved at the same rate as the freestream airspeed. all of which lead to disturbances. it is easy to measure the pressures and resultants on the model. There must always be a ground-board (fairing) near the belt's upstream end. It is very difficult to avoid vibra­ tions. causing unavoidable boundary-layers and unevennesses are. According to Stalker [253] and Klemin [252]. thus properly modeling flight near the ground in still air. this method was first proposed by .3). It is easy to see how the method is cumbersome and expensive. and to match the speeds of the freestream and belt exactly. A final drawback is the virtual impossibility of measuring the pressure distribution on the ground or the velocity profile in the gap region.

It also includes the testing of the actual vehicle. May and Pound in 1963 [260] and Poisson-Quinton in 1968 [261]. Moving-Model Technique The moving-model technique is also known as the launch-system method. Other work of the 1960s involving cars includes that of Carr and Roser [262].3: Moving-belt method Eiffel in 1914. or tow-tank technique. so it is simply a scaled replica of . of Turner [257] and of Beauvais. Klemin himself used it in his work on cars and trains in 1934 (with velocities in the range of 70 mph). tow- technique. as did Butler. either in free flight or in tow. Only the model moves. 98 wing wedge plate moving belt Figure 5. though they did not give a reference. Train models were studied by Grunwald in 1970 [259]. Tignor and Turner [263].

in 1932 [91]. and that it is comparatively inexpensive. Janssen and Schwartz [264]. In 1932. This led to systematic studies of the effect of the ground on full- size aircraft in the 1920s and 1930s.4). The aerofoil-boats of Alexander Lippisch were all extensively tested in free-flight (see [6]. This was done in 1975 by Carr [258] and by Hucho. a towing-vehicle. Witmore and Turner . speed-boat. One additional problem with over-water tow-tanks is that the water surface may deform as a result of the pressure fluctuations. Germany. a catapult. [11] and [17]). These included the work of Reid and Carrall ([89] and [90]) in the United States in 1927 and by Tonnies at Hannover. Early aviators and aeronautical engineers were very preoccupied with the takeoff and landing phases. that one can easily measure the ground-pressure distribution. 99 the physical (full-scale) circumstances. Mathematically. or under its own power (as in the case of actual flight-testing). an incline (utilizing gravity). Kaario flew his first ground-effect sledge inches above the ice cover of a Finnish lake (see [98] and [99]). automobiles can be tested by instrumenting them and simply driving them down the road. This technique is performed by launching the model toward the test-section by means of a pneumatic launch-system. The great advantages of this method are that it properly models the flow. the tangency condition is replaced with a stress-equilibrium condition (see Figure 5. Most of these launch systems involve very high accelerations and the strain gauges must be able to withstand these. The great disadvantage is that it is very difficult to measure the forces and pressures on the model during flight. Similarly. The measuring equipment has to be on-board (except for the case of the tow-tank). tracked ground vehicle. a propelled test-rig (as in a tow- tank). A closely related method is to tow the aircraft or model behind a tow-plane. In 1940. or automobile.

100 stationary air moving wing stationary ground Figure 5. Moving-model method towed a glider behind a car [92].4. For the case of trains this was done by Howell and Everitt in 1981 [240]. and by Kemmerly. Lippisch towed his X-112 aerofoil-boat behind a speed-boat in 1963 [6]. The aerodynamics of ground vehicles can also be studied by accelerating models toward the test section with a pneumatic launch system. The moving-model method was used by Ono in 1935 [250] to study two-dimensional car models. These researchers confined their models by mounting them to a constant-speed tracked test carriage. Lippisch also catapulted models into unconstrained flight just above the water surface. Paulson and Compton in 1988 [266]. Papenfuss and Kronast used the . by Baker in 1986 [265]. Tow tanks were used in the early 1960s by Lippisch and Colton [5] and by Carter in 1961 [59] to test finite wings in ground effect.

A Comparison of Experimental Results for Wings Near the Ground Le Sueur [271] claimed that the results of all of the experimental techniques for wings in weak ground effect agree reasonably well with the results of classical ground effect theory. or intermediate ground effect) the . Another method for bringing the model to a certain speed is to accelerate and decelerate it with the help of gravity. In terms of the ground effect. However. But when the trailing edge ground clearance-to-chordlength ratio decreases to within 0. Rheoelectric Analog Method The rheoelectric analog method is also known as the electric tank technique. who was particularly interested in properly modeling small ground clearances.5 and 3. especially if the track or road is elevated. streamfunction. From today's standpoint an obvious method for testing models as they move over the ground is to control them with remote-control (R/C) equipment. this only became practical in the 1950s. Miller and Smith in 1971 [69]. It was introduced in 1961 by Malavard (see [268] and [269]).0 (moderate. The method was also used by Chaplin and Masters in 1964 [270]. However. remotely controlled vehicles were tested by Lippisch in the late 1960s and by Gallington. the effects of wind and wind-gusts cannot be neglected. 101 same technique for automobiles in 1991 [267]. and logarithmic potential. No references have been found for this method. argument-function. Some of these authors point out the importance of simulating the atmospheric boundary-layer. He showed how it is pos­ sible to use an electric tank for simulating the following fields: velocity potential.

102 various experimental methods yield different results. or whether it could admit unsteady flow in the wake region. This was because the boundary-layer on the ground-board reached a considerable thickness ahead of the leading edge. causing an accelerated flow and a change in the angle of incidence. whereas it was from 10 to 15 by the flat-plate method (for weak ground effect). [46]. Pistolesi [238] discussed the fact that there is poor accord between the reflection and the flat-plate methods in that case. Experimenters also found that for positive incidences the lift increases as the ground is approached. Questions were raised whether the reflection method really resulted in a symmetric flow pattern. This accelerated flow was then confined between the wing and the ground and energized the boundary-layer on the ground. [91] and [89]). thintiing it. . For airfoils in moderate to strong ground effect he observed that the fixed ground- plate yielded poorer results than the mirror-image technique. He cited the criticisms made of both methods by Cowley and Lock in 1920 [43]. Miiller found the image technique to work quite well. Results obtained from the reflection method indicated that the maximum lift-to- drag ratio of a wing in moderate ground effect was increased from 10 to 13. as the wakes of the two airfoils tended not to interfere. The criticisms of the flat-plate method were that the combination of the model and the plate gave rise to an unrealistic circulation and that it was difficult to account for the boundary-layer over the ground plate. Miiller [54] carried out experiments on both airfoils and bluff bodies in 1939. In 1935. The conclusion that the increase in the lift-to-drag ratio is about twice as high by the flat-plate method as by the reflection-method has been drawn by numerous authors (see [43].

[274]. [276]. Most of the time the flow about bluff bodies is three-dimensional and unsteady. [240]. and that they are generally not designed to produce lift. Most of these authors compared several experimental techniques. Ono used the mirror-image method. 103 A Note on the Aerodynamics of Bluff Bodies (Cars) Automobile and train aerodynamic testing are described here to determine the effectiveness of the different experimental techniques. [281] and [267]). of Heald [249] in 1934. and buildings such as cooling towers and skyscrapers. [263]. Freestream conditions usually involve wind-induced atmospheric boundary- layers. the experimental findings will become important in the design of the ram-air train fuselages. of Ono [250] in 1935 and of Miiller [54] in 1939. [280]. [279]. with extensive viscous regions of attached and separated flow. the fixed ground-board method and the moving-model method. Vortices are usually generated and shed from the body. The salient features of cars are that they are bluff bodies whose length-to-width ratio is near unity. that they therefore display a strong turbulent wake. [272]. [258]. tractor- trailers. Bluff body aerodynamics are of interest in the study of automobiles. [273]. [278]. [277]. There has been a renewed interest in the aerodynamic design of automobiles since the energy crises of the 1970s (see [262]. there was a strong interest in automobile aerodynamics. Lay used the mirror- image method as well as a fixed ground-plate with and without suction to control the boundary-layer. Some of the significant experimental works of that period are those of Lay [248] in 1933. The strong wake causes a large percentage . Furthermore. In the early 1930s. [275]. and to study the instrumenta­ tion that was used. [264].

railroad vehicle design witnessed a trend to lower mass which would reduce the vehicle's traction energy consumption and thus overall cost. A Note on the Aerodynamics of Slender Bodies (Trains) Beginning in the 1970s. Interestingly. In terms of geometry. so that much attention is given to the design of an automobile's nose. especially if it is of an invasive type. which consists of areas that are moderately oblique or even parallel to the freestream. The atmospheric boundary-layer turbulent length scales are of an order of magnitude larger than the length scales of the car. whereas the tail is called the afterbody or base. and to instrumentation. since normal operation will involve such natural phenomena. to surface irregularities. 104 of the total drag. is refered to as the midbody. . the nose of the car is called the forebody. and their underbodies are parallel to the ground (zero incidence). The remainder of the car. Here the drag is dominated by skin friction. Therefore the scaling is just as important as the experimental technique. Production cars have ground clearance ratios (h/L) near 0. All of these perturbations may alter the size and nature of the boundary-layers.05. Both of these regions include mostly surfaces that are normal to the flow so that the drag is primarily due to pressure forces. It is important to consider the influence of crosswinds and gusts on the aerody­ namic performance of automobiles and other ground vehicles. most of the external drag acting on a car is due to the forebody flow. Small models are more sensitive to flow disturbances. the keys to the successful testing seem to be the use of large models and the avoidance of unrealistic unsteady wakes. It is crucial that the experiment results in boundary layers that are of the proper scale and type. For non-lifting bodies.

the long-range goal of designing a ram-air train must include a consideration of atmo­ spheric boundary-layers due to cross winds. and Yamamoto in 1968 [251] (also see [282]. 105 The drawback was an increased risk from wind-induced accidents such as overturning in gusts. Their underbodies are parallel to the ground. fatigue life.01. Such concerns have given rise to much wind tunnel testing. This topic was treated in 1983 by Howell and Everitt [283] and again in 1986 by Baker [265]. Kawaquiti. Ground distance-to-length ratios are typically near 0. Crosswinds and gusts can have a pro­ found influence on the train's habitability (ride quality). and control system design. [252]. sideward forces. . Even though airfoils with a uniform freestream are the focus of this study. overturning (rolling) motions. Their great length reduces the relative importance of the turbulent wake to the overall flow pattern. an example was the aerodynamic design of the Japanese Tokaido-line trains tested by means of the mirror-image method by Hara. Fukuchi. [259] and [240]). Trains usually have a very large length-to-width ratio and are not designed to produce aerodynamic lift.

e. Just as in history. and to reveal how the theory can be applied to the study of the ground effect. several analytical methods will be covered first. the discussion of the numerical techniques will begin with two- dimensional panel methods. A SURVEY OF THEORIES ASSOCIATED WITH THE DIFFERENT GROUND EFFECT REGIMES In this chapter. the nonlinear small-gap theory. and lifting-line theory. this includes the conformal mapping technique. this is followed by lifting surface theory and its associated numerical methods such as the vortex-lattice method. 106 CHAPTER 6. As the focus of this dissertation is on the two- dimensional problem. a boundary- layer type channel theory.. and lubrication theory (see Figure 6). . i. The goal of the chapter is to provide a brief review of the relevant theories. Last will be a discussion of asymptotic methods. Following the chronological order of appearance. These include the one-dimensional channel flow theory (linear small-gap theory). thin airfoil theory. when the ground clearance is less that ten percent of the chord length. As always the literature is cluttered with long names and slight variations on the same theme. which are valid only in the strong and extreme ground effect. the various theories and numerical approaches that have been used to study the ground effect will be surveyed.

Aerodynamic Ground Effect Tiieory Weak/Classical Intermediate/Moderate Small-Gap Theory.1: Theories associated with the Ground Effect Flow Regimes . Effect: Inviscid Lifting Singularity Inviscid Inviscid Viscous Mapping -Line Distribution Linear Nonlinear Boundary Theory Theory Theory Small-Gap Small-Gap -I -ayer Theory Theory Theory Figure 6. Lubrication Ground Effect Theory Ground Effect Theory Ram-Wing Theory Theory Strong Strong Extreme 3-D 2-D or 3-D Ground Ground Ground 2-D Inviscid Inviscid Effect: liffect.

101. The conclusion is that only very spe­ cial shapes can be studied with this elegant method. Even though it is theoretically possible to map complicated body shapes and effective body shapes (including boundary layers and separated flow regions) to arrive at realistic solutions. such as Joukowsky airfoils in a purely potential flow. 152. [57] and [128]). 120. 100. . With regard to ground effect. The flow around the simple transformed geometry is solved and the results are trans­ formed back into the physical plane. The special case where the trailing edge of the plate was in contact with the ground was successfully treated by Datwyler [48] in Switzerland in 1934 and by De Haller [287] in France in 1936. the solution was given in series form and did not converge for small ground clearances (see discussions in [238]. [285]. The flat plate problem was again treated by Havelock in the United Kingdom in 1938 [286] but the solutions were again complicated and impractical (see discussion in [57]). the transformations become prohibitively complicated and cumbersome. both studies included errors. The first successful application of the mapping technique to the ground effect was due to Tomotika et al [47] (Reports 97. 108 Two-Dimensional Mapping Techniques (Analytical) Two-dimensional potential flow problems can be solved exactly by mapping the actual geometry (physical plane) into a simpler geometry in the transformed plane. They analyzed a flat plate operating near the ground and achieved good agreement with experimental data (see discussion in [285]). the first related application of the mapping tech­ nique was the study of profiles in the vicinity of two parallel walls in 1931. and 182) at Tokyo University. Both Rosenhead [284] and Miiller [54] were interested in the effect of wind tunnel walls. However.

Since the Laplace equation is linear. The latter work may have contained errors that were corrected by Tomotika and Imai in 1938 [289]. working at Durham in the United Kingdom. The foil was disassembled into an isolated symmetrical airfoil at zero incidence. they also used superposition to model a thin airfoil in ground ef­ fect. These were first studied by Hudimoto [288] in . mapped a thin airfoil and its image into two adjacent circles with circulation. In 1965. Green. Two-Dimensional Thin Airfoil Theory (Analytical) The idea behind thin-airfoil theory is to place a singularity distribution on the chord line or on a parallel line to the oncoming flow and to enforce the flow tangency condition on the camber line.Japan in 1934 and then by De Haller in 1936 [287]. Hasimoto and Urano in 1950 [291]. a zero-lift camber line at small incidence at the design height. Since . The first practical solution to this problem was given by Tomotika. An approach for determining the flow around airfoils near the ground was pro­ posed by Green [285] in 1947. In conclusion. but he gave no results (see discussion in [290]). California. and a camber line due to lift at the design height. but it was restricted to Joukowsky-type airfoils. It is not usable in the design process involving airfoils in strong ground effect. Strand and Brainerd [67] at Air Vehicle Corporation in San Diego. also treated this problem [55] but found that his results compared poorly with experimental data. attention shifted to circular arc airfoils near the ground. so that velocities are not altered by more than ten percent. the mapping technique is really only useful for specific airfoils in a potential flow. The airfoil is assumed to represent a small perturba­ tion to the flow. 109 From flat plates.

This is the theoretical equivalent of the image wing experimental technique. an image wing is usually placed symmetrically underneath the actual wing. The expression for the tangency condi­ tion involves the unknown singularity distribution. Lastly. Furthermore. the slope of the camber line. except that the ground clearance ratio can now be as small as twenty percent of chord (see the work of Bagley [57]). Thus there are two singularity distribu­ tions parallel to each other at zero incidence and a symmetry line is formed between them. This method is also known as the Birnbaum-Ackermann theory. It is an integral equation known as the fundamental equation of thin airfoil theory. modelling the ground. and the incidence. and the involved integral is called the airfoil integral. This places restrictions on the airfoil thickness and on the amount of airfoil camber (neither dimension should be more than ten percent of the chord length). When this theory is used to study the ground effect. He used standard Birnbaum distributions . small thicknesses and small cambers. the ground clearance must be larger than the chord length in order for the flow to be altered by less than ten percent. since these singularities are placed onto a line parallel to the freestream one is restricted to small incidences (on the order of six degrees). If second-order terms are included the assumptions remain the same. Thin airfoil theory was first applied to the study of airfoils in ground effect by Tani [292] at Tokyo University in 1932. meaning that they are enforced on the camber line rather than at the actual airfoil surface. 110 singularities are used. All the same restrictions apply here and the tangency conditions are again linearized to the two lines representing the airfoils. the theory is restricted to potential flow. Note that for this first-order thin airfoil theory to be applicable. the boundary conditions are linearized.

the thin-airfoil method was once again utilized to study the case . Since all vertical dimensions were ten percent of chord at most. and the external flow region. He simulated airfoil thickness by a single doublet placed at mid-chord and the image position. At the time. but represented the thickness by a source distribution on the chord line (see discussion in [128]). [57] and [293]). In the 1980s. the usual approach was to split the domain into two distinct regions. This method was used by Strand. He included higher-order terms in order to study the moderate rather than the weak ground effect. with the "airfoil" being comprised of the wing and its boundary layers and wake as well as the underbody flow. the internal underbody region between the ground and the body streamline formed by the wing and its wake. In 1937. and verified his results using wind tunnel data. It was also proposed by Lippisch in 1964 [8]. Unlike Tani he ignored the effect of the image system on the longitudinal velocity. Tani extended the theory to three dimensions in collaboration with Tomika and Simudu [294] (see discussion in [18]). Bagley's theory was valid for ground clearances as small as twenty percent of chord. In the 1960s. Ill of vorticity to model the profile and its image (see the discussions in [23S]. In the case where the external flow encounters an effective body that hugs the ground plane this is called humf flow^ mound flow^ or half-body flow. Tani's two-dimensional theory was extended by Bagley in the United Kingdom in 1960 [57]. Royce and Fujita [111] working on Vehicle Research Corporation's VRC-1 and "Columbia" ram-wings in 1962. Bagley used the same vorticity distribution as Tani.1). interest developed in ram-wing vehicles so that theories were needed for studying the strong ground effect « 0. the external flow region was like a thin-airfoil problem.

112 of large ground clearances (see [295]. Three-Dimensional Finite Wing Theory: Lifting Lines And Lifting Surfaces (Analytical) To model a finite (three-dimensional) wing one can replace the wing with a suitable distribution of singularities. one can replace the entire wing with a system of horseshoe vortices whose bound parts are placed at the quarter-chord position. linear) lifting line theory (analytical) When dealing with a straight wing of high aspect ratio (greater than three). One models the airfoil shape (thickness and camber) by using the results from thin-airfoil theory. The vortex strength varies in the spanwise direction only. Since the flow under a ram-wing is changed by more than ten percent it is necessary to abandon linear thin airfoil theory in favor of a nonlinear theory in which the boundary conditions are enforced at the actual airfoil surface (see [57] and [293]). In the simplest approach. [290] and [296]). the circulation repre­ senting the wing is concentrated on a single line (the lifting line) at the quarter-chord position on the chord line. Three-dimensional simple (classical. A more sophisticated method entails placing a vortex distribution on the chord plane that varies both in the spanwise and chordwise directions. Another reason for this is that more realistic body shapes and incidences may be analyzed this way. This two-dimensional distribution causes a lifting surface to be formed. A large part of the information on finite wing theory contained herein is based on the discussions in the books of Schlichting and Truckenbrodt [297] and Katz and Plotkin [298]. so that the lift-curve slope is .

An inherent assumption is that both the bound vortex and the traihng free vortices He on a plane parallel to the freestream. so that one is dealing with a linear theory. 113 27r and the aerodynamic center is at the quarter-chord. one is restricted to the potential flow around thin airfoils at small incidences. thereby reducing the induced drag. but not pitching moment can be deduced. Germany in 1912. Wieselsberger [45] was first to actually carry out the work at Gottingen in 1921 and verified his results experimentally (see discussions in [48]. As in thin airfoil theory. The use of biplane theory in the study of wings in ground effect was first proposed by Betz [42] at Gottingen. More recently in the 1980s. so it can not be used for the moderate or strong ground effect. upon which this theory is based. According to [299] this theory will work for ground clearances as small as 90 percent of chord. The decrease in induced drag is due to the upward velocity component that is due to the presence of the image wing. the lifting line method is utilized in the form of Prandtl's biplane theory. In the study of ground effect. But once again. with thin profiles and low incidences. . Both the wing and its image are modeled as lifting lines. Classical lifting line theory allows the determination of spanwise lift distributions from which lift and induced drag. [238] and [18]). This reduced downwash causes the lift vector to rotate forward. The method was also used by Pistolesi [238] in Italy in 1935. the method is restricted to straight wings of moderate to high aspect ratio. the method was once again utilized to study the case of large ground clearances (see [300]). and his paper includes an interesting review (for a discussion of his work. A linear relation between lift and incidence is the necessary consequence. see [60] and [18]). Lifting line theory was developed just after World War One by Prandtl and his co-workers (see [241] and [242]).

The downwash is then calculated at the three-quarter chord position. Mutterperl and Thwaites. but again relies on thin airfoil theory for chordwise properties (see [297]). It allows for the calculation of the spanwise lift distribution. The method is due to Weissinger. and since the lift and the incidence are therefore linearly related. resulting in the numerical lifting line method. 114 Three-dimensional extended (linear) lifting line theory (analytical) The extended lifting line theory is also known as the three-quarter-point method and applies to wings of any planform and aspect ratio. Ackermann and Bock [302] were first to apply it to the study of ground effect at Darmstadt. To start the procedure. The method is also a linear one. The theory is once again linear. since the bound vortex and the trailing free vortices lie on a plane parallel to the chord. It was first applied to arrow. Three-dimensional double (linear) lifting line theory (analytical) Double lifting line theory is another second-order extension of the classical lifting line theory to planforms of arbitrary shape.and delta-wings in ground effect by Thomas [301] at Braunschweig. Both the lift distribution and the moment distribution can be calculated. many narrow horseshoe vortices are placed side-by-side on the arrow-shaped quarter-chord of the wing. and again in 1966 (see discussion in [299]). an elliptic spanwise distribution of circulation is . Germany in 1958. with all vortex components lying in the chord plane. It is a second-order extension of the classical lifting line theory to wings with taper or sweep. Three-dimensional numerical lifting line method Lifting line theory can be computerized. In this method. Germany in 1961.

If this problem is posed analytically. The procedure involves the placement of a number of lifting lines parallel to the pitch axis or the placing of individual horseshoe vortices all over the planform. The third step is to determine the new circulation distribution based on the lift coefficient. and by Binder [304] at Braunschweig (Brunswick). This process is repeated until error toler­ ances are achieved. The second step is to find the lift coefficient that corresponds to this incidence from experimental data. then only rectangular wings can be treated. An alternative method is due to Glauert and replaces the wing with a spanwise series of elementary wings. Lifting surface theory was developed by Burgers and Prandtl in the mid-1930s. Computational extensions of the lifting line theory to treat the ground effect were also used by Kohlman [303] at Boeing in the United States in 1958. reliance on thin airfoil theory is not necessary and information about the pitching moment can be obtained. this is designated as the multiple-points method. These elementary . If several lifting lines are arranged in series as proposed by Wieghardt. The first step is to calculate the incidence based on the circulation distribution. in 1977. Germany. and forms the subject of 3-D lifting surface theory. 115 assumed initially. thicknesses and cambers can be treated. Since the lift coefficient is based on experimental data rather than on thin-airfoil theory. This allows for the study of tapered and swept wings. Three-dimensional lifting surface theory (analytical) Lifting surface theory applies to wings of arbitrary planform and provides both the spanwise and the chordwise lift distributions. Thus. not just straight and narrow wings. large incidences. the wing must still be straight and narrow. the singularity distribution is continuous. Nevertheless.

and the camber surface shape. This procedure is known as nonlinear lifting surface theory. which in turn involves the unknown singularity distri­ bution. An equation for the velocity induced by the entire lifting surface at any point can be derived. is then applied on the plane rather than on the actual surface. this condition makes the solution unique. linear lifting surface theory is the three-dimensional extension of thin airfoil theory. and Gersten [305] was first to use that theory in the study of the ground effect while working at Braunschweig.. 116 wings are vortex ladders that consist of several narrow horseshoe vortices that are arranged one behind the next (see [297]). Germany. inaplying a linear relation between lift and incidence. The earliest procedures for treating cases in the nonlinear incidence range as­ sumed that the trailing vortex sheets were shed a t half the incidence. In effect. This means that both the bound and the free parts of the vortices lie in the plane. The Kutta condition is enforced by requiring the vortex strength at the trailing edge to vanish. the vortex distribution is placed onto a plane. in 1960. The tangency condition thus produces an integral equation that relates the singularity distribution. The result again involves the kernel function. This equation involves a kernel function. The entire wing is modeled as a flat plate at incidence.e. For linear lifting surface theory. Just as for thin airfoil theory this implies thin and lightly cambered profiles at small incidence. This tangency statement involves a double integral and is called the fundamental integral equation . or kinematic flow condition. and the tangency condition. The tangency condition is then expressed in terms of this equation for the induced velocity. the incidence. i. usually the chord plane. a t a f 2 above the plane of the wing (see [297] and the references to the the works of Bollay and Gersten therein).

in which case it is a first-order method. Munk and Glauert he assumed a Birnbaum distribution in the chordwise direction and a trigonometric interpolation polynomial for the spanwise distribution. The first practical lifting surface theory was developed by Multhopp [306] in 1950. Two.and Three-Dimensional Numerical Methods for Potential Flow: General Comments The usual approach for numerically investigating a potential flow is the panel method. In that case the numerical method is a linear method and the boundary conditions are linearized to the chord surface. so long as separation is properly modeled or does not . Usually these are flat panels and they are often placed onto the chord line or chord surface. This leads to a difficult integral equation that is normally solved numerically. or downwash integral. profile shape. the wing geometry. The method was later improved by Mangier and Spencer. For example. 117 of subsonic lifting surface theory. The advantage of the panel method is that complex geometries can be modeled and that fewer restrictions apply to the magnitudes of the physical dimensions. in comparison to thin airfoil theory. in which the singularity distribution is approximated as a polynomial involving a number of undetermined coefficientsand is often solved through Gauss quadrature. the singularity density is often assumed to be constant over a given panel. or method of aerodynamic influence coefficients. Following Birnbaum. and incidence are given and the singularity distribution is the unknown. One method is the kernel function procedure. The body is modeled by replacing it with many discrete panels to which singularity distributions are attached. Also. In the direct problem. numerical methods can be used for thick bodies and high incidences.

This point is called the control point or collocation point and is usually the midpoint of the panel. The flow tangency condition is enforced at one point on each panel. The results are the singularity strengths. [308]. Two-Dimensional Panel Methods (Numerical) Two-dimensional panel methods are used to numerically solve the Laplace equa­ tion of potential flow around a known airfoil shape (for more details. see [307]. 118 occur. These equations are written in matrix form and customarily solved through an iterative Gauss-Seidel procedure. [310] or [311]). particularly in the case of a first-order method. numerical methods do not break down in the regions of strong curvature near the leading and trailing edges of an airfoil. A panel method is a point singularity method or discrete method if the singularity . where each algebraic equation entails the singularity strengths as unknowns. and not throughout the domain as in finite difference methods. This is much easier to implement and results in an acceptable error. Also. Another powerful aspect of panel methods is that points are located only on the body surface. It is of course most realistic to place the control points onto the actual surface. The actual body is replaced by a finite number of straight-line segments that have singularity distributions placed upon them. but this location causes numerical difficulties. The overall computational procedure consists of writing an algebraic equation representing the boundary condition for each control point. Boundary points are thus placed onto the air­ foil surface. These segments are called panels and their edge points are called boundary points or nodes. [309]. greatly reducing the number of equations. which allow computation of the surface velocities and pressure distributions.

it is a second-order method. A panel method is said to be a first- order method or constant-strength method if the singularity strength is constant for a given panel. For lifting flows. For non-lifting flows the enforcement of the tangency condition at the N control points results in a system of N equations. with N unknown singularity strengths. as well as an additional equation from the trailing edge Kutta condition. This problem can be resolved in a variety of ways. Since there are only N unknowns. the system is overdetermined in that it consists of N+1 equations. A further difficult aspect of two- dimensional panel methods is that the Kutta condition cannot be enforced directly at the trailing edge. The system can be solved quite readily. not a control point. N equations are obtained from the tangency boundary conditions. namely the singularity strengths. as discussed below. . Two-dimensional source panel method (numerical) The two-dimensional source panel method has been used for the non-lifting po­ tential flow over arbitrary two-dimensional bodies since the late 1960s (see discussion in [311]). If the singularity strength varies linearly along a given panel. The reason is that the trailing edge proper coincides with a boundary point. Therefore the usual procedure is to express the Kutta condition in a form that involves the two nearest control points. every panel has a constant and different source strength and the tangency condition is enforced at the midpoint. Customarily. The resulting system of equations can be easily solved numerically. 119 strength is lumped at one point on the panel.

the Kutta condition serves to guarantee that the flow leaves the trailing edge smoothly. The method is usually used for very thin airfoils at incidence. The method can accomodate thick bodies at high incidences.e. In the first-order method. The panels are then placed onto the camber line of the airfoil. or to introduce a fictitious control point (the Kutta point) just downstream of the trailing edge. difficulties in deciding the number and distribution of panels. the surface is covered with panels of constant vorticity. Since the trailing edge proper is a node or boundary point of two panels (panel 1 on the bottom and panel n at the top). the Kutta condition cannot be enforced there. This is like wrapping a vortex sheet around the airfoil. The tangency condition is enforced at all midpoints and the Kutta condition is enforced at the trailing edge. that there is no velocity discontinuity there.. Some of the problems with the method are accuracy problems. one common procedure to take into account boundary layers is to represent them by adding a displacement thickness to the actual body. one uses a point vortex at the quarter-chord of the panel and one enforces the boundary condition at the three-quarter chord position (the method is based on the lumped-vortex element). For example. In the discrete vortex method. i. provided that bound­ ary layers and separated flow regions are either negligible or absent or have been accounted for. and the troublesome decision of which control point to . Since there exists no wake for the two-dimensional case. The ususal procedures are to require that the two vortex strengths nearest the trailing edge are equal and opposite. 120 Two-dimensional vortex panel method (numerical) Lifting potential flows over two-dimensional bodies have been solved with tiie two-dimensional vortex panel method since the early 1970s (see discussion in [311]).

so the system was determinate. Source panels are used to model the thickness whereas the uniform vortex sheet provides the circulation. Since the Laplace equation is linear. the two solutions may be superimposed. The N tangency conditions and the Kutta condition resulted in (N-f 1) equations. pages 103-110). Hess and Smith assigned a constant but different source strength to each straight- line panel. resulting in (N-f-l) unknowns. Thus there were N unknown source strengths and one unknown vortex strength. With regard to profiles in ground effect. Part of the name stems from the fact that the problem of solving Laplace's equation subject to homogeneous boundary conditions is known as the Neumann problem. Two-dimensional Douglas-Neumann panel method (numerical) This pioneering source-based panel method was developed by Hess and Smith [307] at Douglas Aircraft Company in the mid-1960s (see the discussion in [310]. this method was used by Giesing in Germany in 1964 [312] and 1968 [313] (see discussion in [295]). The two-dimensional Douglas-Neumann panel method is also known as the Douglas plane-potential-flow method and as the Douglas-Neumann source- distribution method with circulation. as well as the same vortex strength. . 121 choose to eliminate to reduce the number of equations to N. The Douglas-Neumann Method does not perform very well for cases where the airfoil is thin or where the trailing edge is either sharp or cusped.

it is difficult to arrive at a realistic trailing edge condition. with the tangency boundary condition enforced on the planar lifting surface rather than at the actual wing surface. The theory is linearized in the sense that the wing and its wake are treated as a flat surface. The planform is thus covered with two planar vortex sheets comprising a lifting surface while the wake is composed of trailing vortices that form the planar wake sheet. Three-dimensional panel methods (numerical) In the classical linearized surface singularity method a series of lifting lines is placed onto the planform (chord plane) at small chordwise intervals (see page 261 of Anderson's book [311]). which have the advantage of yielding the surface velocities without the need for integration. He mentions the constant-potential method of Morino and Kuo. This shortcoming is overcome by Moran's linear-potential method. They also do not fare well for cases with sharp or cusped trailing edges. since the trailing edge proper . As in the two-dimensional case. This results in a grid of rectangular panels on the planform. which is a constant doublet-strength method that assumes a constant but different potential on each panel. as well as a vortex sheet in the wake. It is not useful for sharp or cusped trailing edges. Moran [310] also discusses potential-based panel methods. 122 Other two-dimensional panel methods (numerical) It is possible to design panel methods based on doublet distributions. but dou­ blets are difficult to deal with and do not allow the Kutta condition to be expressed explicitly. Each discrete panel's quarter-chord coincides with the local sweep-back and the tangency boundary condition is applied at the spanwise midpoint of the panel's three-quarter chord (see pages 130 to 134 of Moran's book [310]).

or bisector. To each vortex ladder a so-called Kutta control point is assigned at the trailing edge where the Kutta condition is enforced. tangent to the circumference. the velocity is usually required to be tangent to the trailing edge axis. which is the first control point in the wake (see [309]). Usually.e. or thickness problem) and a lifting surface (circulation problem. A control point is placed at the center of each panel. the singularity strength is assumed constant on a given panel. i. which follows here in a slightly altered form. 123 coincides with a boundary point and not a control point. Usually. or lifting problem). and the wake is initially just an extension of the camber surface. The usual procedure is to express the trailing edge condition in a form that involves the so-called Kutta point. At that point. one can think of a lifting body as being comprised of anon-lifting displacement body (displacement problem. with the singularity strength on each given panel being constant.. where the tangency condition is enforced. Thus. These surface panels are flat and lie on the surface of the body. Therefore. a chordwise series of individual horseshoe vortices. Schlichting and Truckenbrodt [297] give a description of a general panel method. These inner panels and the associated wake vortex sheet comprise a lifting surface. there are as many tangency statements as there are surface panels. The chordwise circulation distribution within each of the vortex ladders is assumed to be a Birnbaum distribution. there are as many statements of . As discussed. The circulation is modeled by placing a number of vortex panels on the inner surface. The displacement body can thus be modeled by replacing it with a number of discrete panels that are covered with a source-sink distribution. Each elementary wing consists of a vortex ladder. which is usually the chord plane. the lifting surface is composed of a series of side-by-side elementary wings.

California. Three-dimensional vortex lattice method (VLM. A modified version was used by Gray [316] in 1971 in his study of a rectangular wing operating in a rectangular trough (see discussion in [317]). A simplified panel method addressing the ground effect was presented liy Ger- sten and von der Decken [314] in 1966. In the linear VLM the entire lattice lies on a plane that is either parallel to the freestream or parallel to the chord line. The simplification was due to the addition of a slender-body assumption. The tangency conditions were enforced at the actual surface as well. . the nonlinear VLM provides for a curved wake and the panels are on the mean camber line. and at small incidence. moderate thickness. They studied wings of low aspect ratio. By contrast. numerical) The three-dimensional vortex lattice method consists of placing horseshoe vor­ tices that have very short bound parts all over the planform. This is analogous to thin-airfoil theory where the tangency conditions are linearized to the plane. resulting in a lattice work composed of trapezoidal panels (see [315]). This method was first applied to the study of wings in ground effect by Licher [56] working for Douglas at Santa Monica. They assumed a linear relationship between lift and incidence and then used a correction factor to adjust the results for nonlinear effects.the Kutta condition as there are vortex ladders. The nonlin- earities were due to the ground effect and to the presence of leading edge separation for slender wings in general. which makes the wake planar as well. Singularities were distributed on the surfaces of the wing and its mirror image. in 1956 (see discussion in [57] and [299]).

this is called the vented half-body theory and was proposed. 125 Three-Dimensional Wings in Close Ground Effect: Early Methods (Analytical) There are several simpler ways of modeling the gap flow underneath the wing. the flow around an empty engine shroud is calculated with the bottom half discarded. Their common feature is that they all assume a perfect seal at the wingtips. Indeed. by Lippisch. this is called the solid half-body theory. Here. among others. Asymptotic Linear 1-D Channel Flow Theory is based on the assumption that the ground clearance is no larger than ten percent of chord. camber. Their purpose is to provide a pragmatic solution for application to the preliminary design process. Three-Dimensional Wings In Close Ground Effect: Asymptotic Linear 1-D Channel Flow Theory (Analytical) If a wing has end plates that greatly reduce leakage. the governing equation . It has also been proposed to treat the problem like a ring-wing. if the ground distance is much smaller (ten percent) than the chord length. this is known as the small-gap assumption. thickness) are much smaller than the ground distance. The simplest method is to neglect the gap flow altogether and to simply assume that the flat bottom of the half-body is acted upon by the freestream pressure. so that the induced drag is identically zero. and if all other geometric parame­ ters (incidence. Another method is to assume that the gap flow has stagnated fully so that the wing's underside is at stagnation pressure. then the flow within the channel formed between the wing and the ground can be approx­ imated as being one-dimensional.

The theory is called linear in the sense that the tangency boundary condition is not enforced at the actual airfoil surfaces.S. the external flow is assumed to remain unaffected at leading order.I. which is thin airfoil theory applied to the flow over a distribution of sources on the ground plane. Lastly. The first leakage model was incorporated into the theory by Ando [127] at Kawasaki in 1966. so that thin airfoil theory can be used to analyze it. They treated the leakage as a simple slit orifice problem. Gallington and his co-workers ([69] and [70]) tested ram-wings in a wind tunnel using the image technique and reported good agreement of pressure distributions and . It was developed by Royce and Rethorst [108] in 1961 and first utilized by Strand. in 1973. His model displayed infinite velocities under the sidewall. The same theory was used by Ashill [128] in Japan in 1963 in the analysis of Kawasaki's ram-wing boat. but on a line that is parallel to the ground. the mass flow rate under the leading edge equals the sum of the mass flow leaking out beneath the end plates and the mass flow exiting from underneath the trailing edge.T. One variation of this theory does not include a leakage model and is known as closed channel flow theory. They were designing the VRC-1. Royce and Fujita [111] at Vehicle Research Corporation in 1962 [69]. 126 for the flow is essentially just a statement of continuity. since the wing is thin and so close to the ground. Air Force Academy in 1971. which was a model of the full-scale ram-wing craft "Columbia". The external flow was modeled as a mound flow. Miller and Smith (see [69] and [70]) in their design of ram-wing vehicles at the U. The approach was also used by Tan [210] at the University of Massachusetts in 1972 and by Boccadoro [71] working at M. the KAG-1 [69]. The first practical leakage model was developed by Gallington. which he graciously admitted.

I. The two theories diverge strongly as the ground clearance increases. the theory was again linear since the tangency boundary condition was enforced on a line parallel to the ground. However. Three-Dimensional Wings in Close Ground Effect: Asymptotic Linear 2-D Channel Flow Theory A more rigorous asymptotic development of channel flow theory was pursued by Widnall and Barrows (see [199]. Barrows and Widnall verified the results of Gallington. between 1969 and 1972. Miller and Smith for cases where the effects of the airfoil edges are ignored. Indeed. [153]. the results did not agree well with the two-dimensional channel flow theory that will be discussed next. the rear stagnation streamline was assumed to proceed parallel to . but they both show the nonlinear dependence of lift on incidence and ground clearance. Also. but their design eventually evolved into a vehicle that flew in an open trough. the theory was still restricted to thin wings that were nearly parallel to the ground. They were initially interested in tube flight ram-wing vehicles. but since the channel flow was assumed to be one-dimensional. [255]. since the curvature of the streamlines is ob­ viously very strong there. as Tan [210] showed. 127 resultants with their theory. These local solutions are called edge flow solutions and the regions are called inner regions since they require inner expansions (asymptotic se­ ries) of the flow variables. [318] and [154]) at M. The inner solutions allowed the treatment of typical airfoil shapes.T. Following the Method of Matched Asymptotic Expansions (MMAE) the two-dimensional inner solutions were matched to the one-dimensional channel flow under the wing. In their asymp­ totic linear 2-D channel flow theory they allowed for two-dimensional flows in the vicinity of the leading and trailing edges.

No flow visualization experiments have been carried out to determine how the ground effect changes the nature of turbulent boundary layers on an airfoil in . the theory is nonlinear because the tangency boundary condition is enforced at the actual airfoil surfaces instead of being linearized. Miller at Brown University. A criticism offered by Boccadoro [71] was that the method was too complicated for design applications. U. Kida and Miyai in Osaka. with the assumption of zero leakage at the wing tips.S. The nonlinear theory was extended to three dimensions by Newman [323] in 1982. it was used by other researchers in the 1970s (for example. In this case.the ground. the flow is still assumed to be inviscid. Japan [319]. [160] and Curtiss and Putman at Princeton University [72]). However. The resulting asymptotic nonlinear 2-D channel flow theory can be used to analyze more realistic problems. for which the incidence is near six degrees and the thickness and camber are near ten percent of chord. Three-Dimensional Wings in Extreme Ground Effect: The Growing Influence of Viscosity It is not exactly clear at what ground clearance the viscosity becomes a significant factor.A. Nevertheless. irrotational and steady. Three-Dimensional Wings In Close Ground Effect: Asymptotic Nonlinear 2-D Channel Flow Theory The linear theory of Widnall and Barrows discussed above was extended by Tuck at the University of Adelaide in Australia in 1979 (see [320]. [321] and [322]) to accomodate cases where the geometric parameters are of the same order of magnitude as the trailing edge ground clearance.

it is not clear whether there are recirculating flow regions underneath the airfoil. Also. 129 comparison with the isolated case. In the unlikely event that the ground clearance becomes smaller still. This is particularly true when pressure distributions are desired and especially near the trailing edge. nor can the numerical panel methods when applied to the strong ground effect. the entire underbody flow is governed by boundary layer equations. Ohta and Kuwahara did for automotive aerodynamics (see [327] and the discussions in [290] and [326]). It is certain that the viscous terms and inertia terms come into play together once the ground clearance approaches the thickness of the boundary layers cis calculated for an isolated wing. then the inviscid channel-flow theories cannot be correct. The corresponding boundary layer theory was presented in a 1983 paper by Tuck and Bentwich [324]. In that case. the underbody flow is governed by lubrication theory. . If this is really the case. Bagley [57] claims that boundary layer calculations are needed when ground clearances become smaller than twenty percent of chord. or whether friction choking occurs as the boundary layers grow to effectively seal off the trailing edge gap. the inertia forces become negligible in comparison with the viscous forces. In that case. One can also resort to Euler calculations with boundary layer models (see [325] and the discussion in [326]) or even to full Navier- Stokes solvers as Hashiguchi.

It is therefore crucial to properly predict the performance of a wing in ground proximity. Studying a two-dimensional wing section is useful because three-dimensional effects are avoided and the relevant flow phenomena may be more readily identified. The aim of this work is to determine the feasibility of a train that uses the aerodynamic ground effect for levitation. The logical first step would seem to be an experimental study of a wing section near the ground. As it is expensive and laborious to construct a series of wings for the wind tunnel. flow separation needs to be avoided. it is more practical to develop a preliminary theoretical method for estimating the presssure distribution on the wing section. Since flow separation is related to the severity of the pressure minimum on the upper surface near the leading edge. It is for these reasons that a preliminary . In order to witness such flow phenomena. extreme pressures in that region need to be avoided. This indicates the need for an experimental investigation of the flow phenomena occurring in ground effect. THE STUDY OF WING SECTIONS IN GROUND EFFECT In this chapter an approach to designing a wing for use in ground effect is laid out. The first question that needs to be answered is whether the wing can produce sufficient lift to suspend the train. 130 CHAPTER 7. The review of the theoretical literature in the preceding chapter has shown that very little experimental data exist to verify those theories.

The resulting theoretical model can be used to optimize the wing section. in the hope that flow separation would be avoided. The most pressing goal of the design is therefore to maximize the lift. the wing geometry is severely limited to minimize the cost and unsightliness of the guideway. The comparison will clarify whether it is necessary to include boundary layer calculations in the theoretical model to account for viscous effects. A comparison of the theoretical and experimental data point the way to an appropriate theory for wing sections in ground effect (Chapter 11). The panel method was used to design a wing section that showed a mild pressure minimum. 131 inviscid panel method is developed in Chapter 8. In the case of the ram- air train. That wing section was then tested in a wind tunnel to provide insight into the flow phenomena and to produce a set of experimental data (Chapters 9 and 10). .

But for now the goal is to obtain preliminary insight.1 and an incidence of 0. . it is necessary to develop an analysis tool for the design of a section that avoids flow separation on the top surface near the leading edge. in terms of an actual terrafoil train. The goal is to design the camber line such that the steep pressure drop and rise at the top surface just downstream of the leading edge is minimized. Other fixed parameters of the analysis are a trailing edge ground clearance ratio of 0. A number of vortex panels will be distributed on the actual surfaces of the profile and of its mirror image. The wind tunnel test is expected to yield results that will point the way to the proper analytical approach. In this way the underbody flow can develop without being influenced by the separated flow region of the top surface. 132 CHAPTER 8. Also.1 radians. flow separation would lead to high drag levels. Both the thickness distribution and the shape of the camber line correspond to NACA's four-digit series of profiles. Each surface will be described by prescribing a standard ten to twelve percent thick symmetric profile and by allowing the designer to vary the shape of the camber line. A PANEL METHOD FOR PRELIMINARY WING SECTION DESIGN Motivation for an Analysis Using an Inviscid 2-D Panel Method Before setting out to test a profile in the wind tunnel. so the choice is a two-dimensional panel method.

Following established procedures set fourth in the literature.1: Dimensional geometry of a profile in ground effect Nondimensionalization The advantage of nondimensionalizing the problem parameters is that the re­ sults can be easily scaled to any actual geometry that is dynamically similar. the coordinates and lengths are nondimensionalized with respect to the physical chord length c*. 133 Figure 8. so that . The physical geometry of a profile near the ground is shown in Figure 8. with all of the dimensional parameters identified by a * superscript.1.

which is equal to poo*{voo*)^' * V = Voo POO = ^ (8-2) ^P = -^ qoo It is customary (see [311]) to define a pressure coefficient as follows. Uoo*5 and the pressures with respect to the freestream dynamic pressure. The velocity potential.* X — (8.^ " ~ c* Also. 134 X 2^ c* Z 7^ c* . Cp = -2 (8. the velocities are nondimensionalized with respect to the magnitude of the freestream velocity. which is defined as .Spoon^'oo*)^ or as Cp — p — Poo — 1 — (S-'J:) according to the Bernoulli equation for potential flow.1) zl " ~ c* h .3) ^ O. f/oo*.

9) ^ Coordinate Systems Let the origin of the global coordinate system be located halfway between the leading edges of the profile and its mirror image.5 can be written as ^ (8.7) Equation 8. 135 needs to be nondimensionalized as well.2).6) Uoo C* then Equation 8. with the x-axis pointing downstream along the chord hne and the z-axis pointing upward. the vortex strength and the sheet strength must be nondimensionalized: i — —T r* T (8. The Y- axis points along the span of the wings to complete the Cartesian system. If the nondimensional velocity potential is defined as $= ^ (8. . Let the origin of the local coordinate system coincide with the leading edge of the profile.7 may be divided by Uoo* and simplified to yield 5$ « 5$ - ^= (8-8> Since point vortices and vortex sheets will be used in the present panel methods. with the X-axis pointing downstream along the symmetry plane and the Z-axis pointing upwards (see Figure 8.

2: Coordinate systems The incidence. a. the temporary (a. 136 z mam wmg 5- flow direction image wmg Figure 8.'. 2') system is displaced from the global (X. This is most easily accomplished by a translation and a rotation.dz)- Thus. is defined as the angle between the horizontal and the chord line. The local coordinate system can be ex­ pressed in terms of the global coordinate system. which implies that the local coordinate system is inclined with respect to the global coordinate system by an angle a. or .Z) system through a distance described by {dx.10) z z' dz K . X ' x' ' dx > =z i ' +. (8. In the translation.

15) z sin a cos a Z —dx sin a — dz cos a < .z) is performed.13 are combined. .3. From Figure 8. 8. 8. .3 the following two matrix equations can be written: X cos a — sin Q J X * < (8.+ < (8.11) z' z -dz Next a clockwise rotation from the temporary (x'.14) Z — sin a cos a z dz or x cos a — sin a X —dx cos a + dz sin a ' — * < • + < (8. 137 XJ X —dx > = •( • + < (8. To achieve an overall coordinate transformation between the global (X. as shown in Figure 8. .12 and 8.10. Equations 8.12) sin a cos a J or J X cos a sino X > = * ' (8.z) system.11. resulting in: X cos a sin a X dx ^ _ * < .Z) system and the local (x. .13) z' — sino cos a z .z') system to the local airfoil system (x.

138 j zsina xcosa I xsina zcosa < X cos a z sina X sina cos a Figure 8.3: Rotation of the coordinate system .

(8.16) The flow is governed by Laplace's equation. = 0. irrotational.18) The geometric parameters of one of the profile's surface segments will be con­ sidered. . inviscid. The implied assumptions are that the flow is steady. As is the custom. and incompressible. As a first approach. The streamfunction can be defined as: (8. so the flow is also assumed to be unseparated and to generate boundary layers of negligible thickness. local panel inclination angles were used initially.17) The solution for $ of Laplace's equation is not unique (single-valued) until the value or its normal derivative is specified at every point on the boundary. For two- dimensional irrotational flow a velocity potential in global coordinates is defined as (8. But this procedure led to much confusion regarding the appropriate quadrants and signs especially when the image airfoil was introduced. separated flow regions or boundary layers are not modeled. It is based on the assumption of potential flow. 139 Derivation of a 2-D Vortex Panel Method In this section a two-dimensional vortex panel method will be derived. A more reliable procedure is to use vectors at all times.

A surface coordinate s is defined having its origin at the leading edge. boundary point 1 is at the trailing edge. the boundary points and control point are numbered in a counter-clockwise sense. the behavior . Since the profile will be replaced by many surface vortex panels. Panels i and j are shown in the context of the entire airfoil in Figure 8. The vector defines the position of panel i with respect to panel j. The unit vector is found in a similar fashion by using = 0. with boundary point 2 on the undersurface just upstream. with surface coordinates Sj and Indeed. They relate the midpoints of panels i and j. Therefore. Furthermore. beginning at the trailing edge (following [310]).5. the length of panel j is and its boundary points are j on the left and j+1 on the right. so • eQ^ = 0.19) Here was derived from the fact that ej. The following equations express relationships between the various parameters of the wing profile. 140 Consider two panels i and j as shown in Figure 8.j and are orthogonal vectors. which are at •^z+1/2^ ('^j+1/2' respectively: (8.4.

4: Geometry of panels i and j on the profile .j 'X- ^ j +1 / 2 -X.Y: ^"^^+1/2 Figure 8. 141 A panel j ^i+1/2- ^j+i I + 1/2 . ^ i +. trailing panel i edge wing bottom —h -Y. 11 .

Therefore. Re­ turning to the panel representation of the wing profile. panel j is assumed to have a distribution of vortices 7(3) upon it. The effect of a two-dimensional point vortex is the same as that of an infinite three-dimensional rectilinear vortex filament (see [311]. the panel's entire strength is: jj = f j{s)ds. 142 N- > X Figure 8. (8. ^ v p j e = 2?!^' where F is the vortex strength and where the subscript v stands for vortex. for example): r r r . ^vpj = .21) The sum of the strengths of the panels representing the profile is just the circu­ .5: Airfoil surface coordinates and panel numbering system of a single point vortex needs to be discussed.

which holds for linear partial differential equations such as the Laplace equation of potential flow.Therefore. Panel j induces the following potential. To find the effect of all N panels at the point { X p . streamfunction and velocity at a generic point P with coordinates {Xp. Zp).Zp): where the subscript 9 indicates that the induced velocity has no radial component (component along dpj). and let the corresponding angular position be Thus. Zp) = ^vpj. such as ^v{Xp.24) . The result is a summation. lis) Hdij{s))ds^ (8. the preceding equations are rewritten for the ith control point instead of the arbitrary point {Xp. one simply adds each panel's contribution according to the principle of superposition. Z p ) . '^vi = . 143 lation r. Since the tangency condition needs to be satisfied at the control points. Let the distance between points i and j be denoted by d^j.

The vortex strength can be factored out of the integral operation of Equation 8. Let the freestream component be identified by a subscript oo so that: Voo = VooL + OK = li. (8. so that 7(s) = 7j = constant (8.. To simplify the problem.25) on panel j. The flow tangency condition states that the flow cannot penetrate the airfoil surface. Vn = 0.24 to yield: The key to the panel method approach is to write expressions for the boundary conditions at the control points. with no restrictions on the shape of the panel or the distribution of vorticity. 144 The derivation to this point is general.e. Therefore. At this point the tangency condition for panel i may be written as: . the strength of the vortex distribution is assumed constant on a given panel. at the airfoil surface.27) where the number one indicates that all velocities are nondimensionalized with re­ spect to the freestream speed. The velocity at any point on or near the airfoil has contributions from the oncoming freestream and from all N panels. the normal component of the velocity must be zero. i.

.29) 2 where = arctan arctan (see Figure 8. This expression of the tangency condition at control point i is the key to the panel method. 145 vn —0 (8.180^ ^iV+1 ^bisector (8. To make the system unique.6).5£]^ ^bisectori^l ""0-5^1 bisector). This new panel (N + 1) is placed downstream of and adjacent to the trailing edge. Therefore. the Kutta condition needs to be applied at the trailing edge. and Vqq is the freestream velocity vector.. 7j. i.28) where is the velocity induced at panel i by the vortex distribution.e. It is applied at all control points to obtain a system of N equations in terms of the unknown N vortex strengths. Panel (N+1) has the same length as panel One (£]^) and is oriented such that it lies on the trailing edge bisector. control point is also called the Kutta point. the coordinates (^yv+l/2''^iV+l/2^ control point N+1/2 are (X]^ + 0. To this end. its surface angle is given by % + (gi . an additional panel is defined.

30) i=i The lift per unit span is according to the Kutta-Joukovsky Law. The problem is finally fully defined and the solution consists of solving a system of N tangency conditions of the form of Equation 8. Since there are only N unknown panel strengths the system of (N+1) equations is now overspecified. are determined the total circulation is computed according to N r = E ij^j. 146 trailing edge -i^V+1 iV+I Kutta point li Figure 8. This provides an extra equation. (8. This is achieved by simply dropping one of the tangency conditions on the profile and replacing it with the Kutta condition. It needs to be reduced to N independent equations.28 above.6: Definition of a wake panel The Kutta condition is approximated by enforcing the tangency condition at the Kutta point. Once the N vortex strengths. 7j. so there are now (N+1) equations. .

5poo*voo*^c*: By convention the nondimensionalized lift coefficient is designated as C£. at a given control point i is the sum of contributions from the freestream velocity. and from the velocity induced at the ith panel by itself: ( ^ \ ^si ~ ^oo.9. That surface pressure distribution is related to the tangential surface velocity distribution.31) where F* = c*Uoo*r from Equation 8. An expression for was derived in Equation 8.19.OO ' tti — 0-D ' Hi /ooi\ .st (8. which is in turn related to the panel vortex strengths.33) i=i Each of the three terms in Equation 8. and less commonly as i or L'. • . from the velocity induced at point i by all other panels. (8. The two-dimensional lift coefficient is obtained by dividing Equation 8. and will be needed to find the tangential components of the three contributions to the velocity at a given point. The tangential component of the freestream velocity is: ^oo. The tangential (surface) velocity.31 by q%Q = 0. 147 t = L'* = poo*voo*V*. To understand the flow around an airfoil it is very important to know the pressure distribution on the surface of the airfoil.sz ~ V. The symbol L' is used in many plots in this dissertation.33 is analyzed in turn.

38) .26. These limit processes are discussed at a later point in this dissertation. By using panel-fixed local coordinates (which will be designated as (^. Katz and Plotkin [298] (pages 84-86) show that the tangential velocity at any point on the panel is = = (8.. --12^ Next. r/) here) and a careful limit process. In a similar manner Katz and Plotkin [298] show that the normal velocity is zero. For the case of a constant singularity strength 7j on the ith panel.19 as well as to Equation 8. the tangential velocity at the top surface is therefore given by: = f• (8-37) It is now possible to find the surface pressure coefficient according to Equation 8. the tangential velocity induced by all N panels (except panel i) at panel i is: yN _ yN • • e+- (8 35) 1 (^N f^i+1 1 ^ ^ ^ ^ Ssf'J^f-Diisyki- . respec­ tively.4: (8.. They are not included at this point in order not to interrupt the flow of the derivation. the tangential velocity induced by the ith panel at its own control point needs to be found.36) where the plus and minus signs refer to the top and bottom side of the panel. 148 Again referring to Equation 8.

= /5oo*i^oo*r*. 149 * oo m'le Figure 8.. (8-«) 0 By dividing Equation 8.40 by = 0.7: Leading edge moment due to panel i Armed with the pressure distribution.*1).4). Referring to Figure 8.7. (8-39) where is the lift per unit span due to the ith panel. the ith panel contributes to the moment about the leading edge as follows: M'lEi = -L'i (A'f+i/a . Replacing L'^ with poo*Voo*T* yields: M'LEi = . i. the leading edge moment (per unit span) is nondimensionalized: .5/9oo*Voo* c* . the moment around the leading edge can be found.e.

The equations for the image wing will be developed in this section. the moment equation. Since the circulation Fj = 7^. the contributions from all panels are summed. here the velocity at the ground is allowed to take on the same value as the underbody flow.41.5/)oo*^^C3O* c*. A local right- handed coordinate system {x^z) is defined for the image wing as shown in Figure 8. (8-43) i=\ Extension of the Panel Method to Ground Effect The ground is modeled by the symmetry plane that is formed by the actual wing and its mirror image. This local system is rotated through an angle a = —a with respect to the global (X. giving N ^mLE = LE = LEi. .8.41) = -2ri(^. ft M '*I p^j M = — i-2 *•) 0. becomes. LEi = ~ ^l). This is not exactly correct since in the actual flow the velocity at the ground plane is that of the freestream.Z) coordinate system.—clz) from the global origin. (8. 150 .* for a given panel i.+l/2-A^l). (^•'^2) To find the two-dimensional moment coefficient. Equation 8. It is also displaced through a distance {dx.

151 Z a = —a Figure 8.8: Coordinate system for the image profile .

48) \X7-XrJ' Just as for all other control points. the tangency condition is applied at the Kutta point.9. Again by analogy to the actual profile. one obtains the coordinate transformations for the image wing. Assuming each .14 and 8.4 and Equation 8.44) z a z ih sin Q cos 1 »• X COS a sin a —dx COS a + dz sin a • = X .46) where 0^ = atan (8. The solution of the system consisting of the actual and the image profile consists of writing the tangency condition for 2N of the 2N+2 control points. X cos Q — sin a X dx ^ X < • + ^ (8.45) z — sin a COS a Z dx sin Oi + dz cos a where cos(—q) = cos a and sin(—a) = —sin a.dz. + < (8.15 with d = —a and the dr with . . The {N + l)th panel lies on the trailing edge bisector so that its inclination is jV+l bisector 2 (8.19).47) 0^ = atan (8. a Kutta point is established for the image profile as shown in Figure 8. One sets up the panel geometry just like before and obtains the same expressions for the associated position and unit vectors (see Figure 8. Also note the node numbering on the image wing. 152 If one replaces the a in Equations 8.

I iV. ^vi ^ni• + ' The resulting system will consist of 2N equations of the form given in Equation .26. One then apphes the tangency condition as before (see Equation 8.28) to obtain =0 — ILi §Lni • + ^^oo)-im (8-50) -vi ' ^ni • ^ni .9: Kutta point and numbering system for the image profile panel to have a unique and constant vortex strength one expresses the influence of all 2N panels upon the ith panel as ^vi = .1 Figure 8.V + 3 +1 2iV+2 N + NCHORD •IN . 153 .^ (8-49) which is based on Equation 8.

42. One then focuses on the actual profile to compute the lift and moment coefficients as discussed in the previous section.19 may be used directly.51) M'lEi = -^•fAiXn. two of these equations will be replaced by the two Kutta conditions. simplifies to .50. The system of 2N equations can be inverted using the Gauss Reduction method and provides the 2N unknown vortex strengths. Equation 8. 8.32. They can be used to solve the flow around a wing and its image.38. 8. From Equations 8. These strengths include the effect of the image profile upon the actual profile. Discrete Vortex Code The simplest way of implementing the preceding 2-D vortex panel method is to lump a given panel's singularity distribution at its midpoint. Since d^j{s) = the expression for the transverse velocity induced at panel i by all panels. Indeed. Equations 8. As explained before. and 8.43 one may compute the resultants and the surface pressure: CI = = ^pi ~ (8. This eliminates the need for all integration since all geometric parameters refer to the midpoints of the two panels and do not depend on the surface coordinate s. 8.30.i/2-Xl) CraLE = ^iLl^'lEi- At this point the equations for the two-dimensional constant-strength vortex panel method have been developed. 154 8.26.

50. A pair of the tangency equations on the airfoil are eliminated to make room for the two Kutta conditions.52 and 8. and one for each of the two Kutta points. ^ • f •\ / 7 • _ 7 • . Once the 2N vortex strengths are found.51 can be used to solve for the lift coefficient.52) The tangency condition. 155 (8.19: / . Equation 8. surface pressures.*i+l/2) (^i+1 - There are 2N + 2 of these tangency conditions.. Equation 8. can be rewritten with the help of Equa­ tions 8.\ (^i+1/2 . and leading edge moment coefficient: CI = 2r = Eyi7j<j Cpi = (8. one for each of the 2N control points. .54) M'lEi = CmLE = Eill'W'ifir . . .

These general expressions will then be simplified to treat the constant-strength vortex panel method in the next section. . 156 where is found from Equations 8. 8.33.36: ^si ~ The dot product of the two unit vectors can be carried out using Equation 8.55 becomes (8.34. the constant-strength vortex panel method will be described. In order to develop the constant-strength vortex panel method. and Equation 8. This method is the next level of difficulty above the discrete vortex method. In the next section. Vortex Panel Method In the preceding sections the discrete vortex method was discussed. and 8.56.54 in conjunction with Equation 8.56) +hi- The solution of the discrete vortex code consists of Equations 8. some very general expressions for vortex panel methods will be the focus of the present section.35.19. 8. This method is the simplest of the methods.

it can be written as: (8.58) The differential velocity induced at Q due to the differential vortex element is (refer to Equation 8.57) Since the transverse (circumferential) unit vector is perpendicular to the radial unit vector. Its differential vortex strength is (ITj — The radial position vector and its magnitude together with the unit position vector are: (8. the variable is a surface coordinate. Consider a differential length of the panel.ZpQ) on or exterior to the airfoil. Similar to x p . the a:p-axis coincides with the panel surface. The goal is to determine the induced effect of the panel on some point Q {xpQ. and it is used to describe a generic location along the panel. the following derivations are carried out using a panel-fixed coordinate system.10. 157 Following the procedure developed by Katz and Plotkin [298].20): . Refering to Figure 8. while the zp-axis is perpendicular to it. The panel extends from 0) to 0) and features some distribution l j { 0 vortices.

^0 / .Ul xet®' wte .

The ZpQ in front of the integral dominates the limit behavior.. from Xpj to yields: For the purpose of determining the surface pressure distribution. Integrating over the panel's length. the differential velocity potential induced at Q due to the differential vortex element is: = '' 27r It is now a simple matter of integration to determine the induced effect at Q due to the entire jth panel (using Equations 8. For now let ZpQ —> 0. 8.e. ^-6QI^^^^ 2irrQj = '^^xQl^h^^^zQIih- Also.62) .61.59 and 8. (8. so t:pQlj^^pQ'^pQ 0) .60). i. 159 dTj . the tangential velocity at the actual panel surface must be found.57. with (XpQ — ^) remaining finite in the expression for v^pQ jj in Equation 8.

then (8. if one chooses a limit in which both ZpQ and {xpQ — 0 are approaching zero at the same rate. 2K ^pj P+1 pQ-"^ (8.64) Now substitute the following variables (following [298]. pages 84-86 and 272-273). the integral goes to zero for all values of except when ^ ~ ^pQ' ~ pulled out of the integral.63 above. resulting in V . dX = — ( 8 . 160 which says that the tangential velocity at the panel surface is zero so long as x^q is not very near However. A= = finite 7 ^ 0 . JXpj /X 'pQ- (^pg-O-^0 ( 2pQ j + (8. 6 5 ) "pQ ^pQ so that the integral simplifies to ( ^pQ ^PJ+1 1 d\ Vg/i( ) .^pj+1 ^pQ (XpQ-O->0 ^ '^j^yo) lim r^pi+1 1 dj -TT.63) From Equation 8.66) •f • .

. very general expressions for vortex panel methods were derived. 161 In conclusion. (8. the vortex strength is taken as a constant on a given panel. the constant-strength vortex panel method will be described. so that IjiO — Ij — constant. the tangential velocity induced by that panel at a point x^q on its own surface is simply one-half of the vortex strength at that point: In this section.69 (see [298]. In the next section.68 into the integrals in Equation 8.61 yields: It is possible to simplify Equations 8. given any distribution of vortices on a flat panel. pages 272-273 and 573-574) by first integrating the expression for the velocity potential The velocity compo­ nents can then be obtained through differentiation of the resulting velocity potential equation with respect to x ^ q and respectively.68) Substituting Equation 8. Constant-Strength Vortex Panel Method In the constant-strength vortex panel method.

73) ~nO The limits of integration become A.74 can be integrated by parts. . Let the variables u and v be defined as follows: u = arctanX dv — x-'^dx (8. so J Xpj XpQ J-ri Xpj^l XpQ that I =—ZpQ X '^arctanXdX. (8.^ p Q) ^p Q from which = -^pQ^~^dX.71) so that 2 dX = —^9—^di = -—di (8..70) Jxr.= .-i i = -r:—. . (S.74) Equation 8.—^ ^vO and A.75) du _ 1 1+A'^ V = -A-1..72) ( ^ .^ -PJ \x„n-U ^ can be evaluated by making the substitution X = _1pQ_ (8.—. 162 The integral /= r^j'+lardanf (8..

163 After integrating by parts and some simplification.i+1 '=-2 (8. the integral in Equation 8.71.80). and if the resulting ? ^pQ expression is simplified (combining Equations 8.79) A.78) 2 JXj u 2 or A.77) a3- -du dX Then i+1 Uu = _i I n u ^i+i = _i /^i+1 (8.80) If A is replaced by — in accordance with Equation 8. one obtains: .76) arctanX ^j+l f 'pQ Xj ^pQ^- The new integral / is evaluated by defining the following variable.. ^ /i+i dx PQ ^ Xj -pQhj rjlA^A^ (8.69 and 8.74 becomes: I = z arctanX ^i+1 _ . ^pQ [K^)] -^i+i (8. so that Equation 8. Let u = 14-A2 du _9 (8.76 becomes \arctanX' 1= yQ[— .

XpQ )Oj + The velocity components can now be obtained by taking the partial derivative of the velocity potential: Oil ^xpQIj d x pQ 5$ (8.83) {Xpj . geometric relationships can be written that will sim­ plify the preceding equation: tanO' • I 1 = — — ^vQ ^"pQ~^'pj+l ^pi+l~^pQ ~^pQ _ pQ tanOj = X n~x • J -^pQ -^pj ^pj ^pQ (8.81 are tedious but relatively simple.Tpg)-'+2pg^ Returning to Figure 8.81 yields: ^ Q l j ^ ^ p Q ' ^ p Q ^ ~ 5ffK^pi+i ^PQ)^J+I . r? (8.10.82) ''j+I ~ i ^ p j + l ~ ^pQ)^ + ^pQ' Substituting Equations 8.82 into 8. and result in: .81) (a:pj-.84) Oil ^zpQIj ~cFi pQ The differentiations of Equation 8. 164 (8.

The geometry of Figure 8. 165 Oj Q j+ 1 top side — * ' — - bottom side 'VjT 7TT7J 0.11: Approaching the midpoint from the top i ^pj+l ^pQ/1 = S K . U .'i+1 (^pj ^PQ) +^p(5' (8. To show this.85 the following is obtained: . the self-induced effect is evaluated at a slight positive distance from the midpoint of the panel. ~ 1". + i Figure 8.83 and 8.85) At this point it is relatively easy to determine the effect of the jth panel upon itself.11 reveals that 6j ~ 27r. and rj_^i = rj ~ By inserting these findings into Equations 8. .

~ -^ .\ / bottom side 0.85 become: *5 = W .j ~ 0. 166 ^ top side J Kr T ~ I j" —r.12 in a similar manner.85 for the velocity components into . one finds 0.86 and 8. zpjj By analyzing Figure 8.87) "xpjj V • • = 0. zpjj Comparing Equations 8. . The next step is to recast the Expressions 8.87 observe that there are jump discontinuities in the velocity potential and in the tangential velocities between the top and the bottom sides of the panel. Also note that the self-induced normal velocity is zero at the panel's midpoint. Ij = —^ (8.83 and 8.12: Approaching the midpoint from the bottom JJ J xpjj XVI7 ^ •• = 0. ^ Figure 8.. ~ and L rj-^1 = rj ~ and Equations 8.

From the Figure it can be seen that: .13. 'j+1 = + . (j = ^J^X.+i-XJf + {ZJ^l-ZJf Using the definition of the dot product. 167 global variables.90) .a8. cosOj — ip • Eyj cosOjj^Y — ip'krj-\-\ and can be rewritten as. The geometric relationship between the local (panel-centered) and global coordinate systems is depicted in Figure 8. (8. relationships for angles 6 j and &j-^i can be obtained.a.

. 168 k. length .p Z / panel j .ZJ+l- .z.Q K -rj+l global •^J +1 4 X origin Figure 8.Z.13: Global and local coordinates .

how­ ever.91) Using the signs of the sine and cosine expressions of Equations 8. . instead of aligning the Kutta panel with the trailing edge .89 and 8. the Kutta panel inclination was made adjustable by the user. From a physical point of view. 1 \ /X.j / (M) \ -J J _\ / (^)li \ •] sinOjj^Y — ip ^ ^rj+l L-\ /Xr^-X: . Internal Boundary Conditions and Other Difficulties When the vortex panel method was first programmed (in FORTRAN 77) and executed._.91 the quadrant in which the angle lies can be determined. Thus. The incorrect results persisted. the resulting pressure distribution had an incorrect and oscillatory shape.Since the trailing edge Kutta condition was suspected of causing these problems. 169 The problem with these expressions is the uncertainty of what quadrant the angle is in. Mistakes in the scalar descriptions of the geometry were suspected and the problem was recast entirely in terms of vector relations. 1 -X:\ /Zn-Z. To resolve this problem the cross product is used to compute the sine of the angle: sinOj = ip X ij ._i N- J- (S. the wake is expected to be displaced upward by the pres­ ence of the ground. In efforts to remedy the situation the panel strengths of the top and image airfoils were forced to be antisymmetric (of opposing signs) and the Kutta condition described previously was at one time abandoned in favor of the classical Kutta con­ dition (7j = —7/v).

Furthermore the panel vortex strengths were antisymmetric and the code provided results for a wide range of cambers. The tangency condition for the jth panel can be written in matrix form: . These equations show that the normal velocity induced by a given panel on itself at its own midpoint is zero. Once this internal boundary condition was implemented for each panel the code performed as expected. The inclination of the Kutta panel and the type of Kutta condition were found to have no effect on the solution. and ground clearances. incidences. thicknesses. pages 240 and 327-332. resulting in a different circulation. internal no-slip boundary condition. In order to understand why the Neumann tangency boundary condition gives rise to problems with this method. The following discussion of the internal boundary condition is based on informa­ tion taken from Katz and Plotkin's book [298]. recall Equations 8. 170 bisector (at an angle ^hisector)'' panel was oriented so that it was closer to being parallel to the ground (at an angle This should have had the effect of moving the location of the front stagnation point.87. external tangency boundary condition is replaced with an indirect. The author then came across a remark in the book of Katz and Plotkin [298] that indicated that the constant-strength vortex panel method only works if the physical.86 and 8. Despite all of the above efforts the incorrect solutions remained.

which implies that the matrix of influence coefficients [A] has a zero diagonal. The only possible conclusion is that ^ = constant throughout the interior. It is obvious that no fluid should penetrate the physical body. The constant is equal to the value of the stagnation streamline. which is cus­ tomarily chosen to be zero. By the same token the velocity potential is constant . Provided there are no singularities in the interior.92) 72iV+2 . an internal no-slip condition can be used. no streamlines may traverse the boundaries.. Another inter­ pretation is that no streamlines can enter or leave the region occupied by the airfoil. This Dirichlet boundary condition follows from the fact that one may specify the internal velocity potential so that the physical tangency condition will be met indirectly. C'jj • ••^2N+2j\ ' = bj = 0 (8. Therefore there are no streamlines within the airfoil's interior. . and an explanation follows. This applies to both the region exterior and the region interior to the airfoil. i. Based on the above comments ajj = 0. and there is no flow at all. To eliminate this zero self-induced effect. no streamlines may begin or end within the airfoil's interior. For a proof of this fact Katz and Plotkin [298] refer to page 41 of Lamb's book [328]. or body streamline. This will cause difficulties with the inversion of that matrix. even if pivoting is used. . 171 71 72 [aij a2j .e.

iti = I J \ h J (8. Here the focus will be on the internal no-slip condition.85 for the local components of the velocity vector induced at point Q by the jth panel: . 1 lo-Z.27.96) Uoo ~ These expressions pertain to panels with constant distributions of vorticity.49 are used. . 8. 172 throughout the interior.. ^int ~ (constant. -0 dn ~ IT. Also recall expressions 8. They are restated here: . respectively. 1 /ON . (8-95) To rewrite this condition for the ith panel. which can be written as. ~ (8. 0 = "Z int = + ^^oo) • k. and 8. /X.93) This then implies that any derivative of equals zero: ^^int .94) ^= n i n t = ^- These are internal tangency and no-slip conditions. (8.19.y • . 1 lO-X. Equations 8.

"ZQ/j = "xpQIiC'i^ "J ^ -J To find the velocity v ^ q induced at Q by all 2N panels simply use a summation. (8. Then .91. Refering to Figure 8. 8.101) i=i To use this expression in the internal no-slip boundary condition. 2N ^vQ=T.13 one writes ip — *i±i_ii'i2 + ( 2i±j_^'i K Ti T" (8.100) /^j.88. 173 ^pQH = "^xpQIjk + '^zpQIji^V = SK'-^i+i] ^zpQIj^^pQ'^pQ^ ' where the radii and angles and their quadrants are known from Equations 8.99) where ^XQ/i .1 -X.^QIj.98) hp — which one can insert into Equation 8.90 and 8.''xpQ/j{-^^'%-^)+"zpQlj[-^ J. Then (8.95. Equation 8. (8. replace point Q with the midpoint of the ith panel..97. The local velocity components need to be transformed to the global coordinate system.

102) where 2N ^vi = L ^i/j^ (8-103) i=i and since the tangential unit vector is known from Equation 8. = cos ^i+l TJ W J\ (8.J+1 .100 and 8. the angles 0 and their quadrants. 8.. ' Z n .. Xj)' + (ZQ .106) . 1-Xn /Xn-X.96.105) ' Z..-. as well as the distances rj and are known from Equations 8.91: ''j xq .97.Z j y r.' _L 1 1 "zi/i = ¥ Again. + {Zq - = -1 ~z.A / Z n . 'Xi/j = ^[^-''j+l]( ')+^' 1 (8.4.-Z. 174 0= + Uoo) -Ih"' (8.104) (^) where.Z : COS [(^^)( ^ ) + ( ^ )(^)] J r/X.88.n /Z.Z . from Equations 8.. rewrite the internal boundary condition as: 0 = (8.90 and 8.

can be expressed in matrix form (see 8.Initial Results The preceding constant strength vortex panel method with internal boundary conditions was coded in FORTRAN77 and executed on a DEC 5000/200 workstation at Iowa State University.1.107. The trailing edge ground clearance ratio was fixed at 200 to approximate an isolated profile. 24 Megabytes of RAM memory. The present code was executed using 50 chordwise stations (98 control points on the top airfoil. Equations 8. and a SPECmark of 19.92). That computer has a TurboChannel architecture with a 100 MB/sec bus. The computer is part of Iowa State University's distributed computing client-server system called "Vincent".87) recall that the tangential velocity induced by the jth panel just at the interior of its own midpoint is = V j . Comparative graphs are depicted in Figures 8. implying that the troublesome zero diagonal has been successfully re­ moved. Constant Strength Vortex Panel Code . = -i- The internal no-slip conditions. Note that ajj = -i (8.15.14 and 8. 198 control points in total). . 25 Megahertz clockspeed. To gauge the effectiveness of the code it was executed for a NACA 4412 standard profile.104 and 8.108) in this case. 175 From the previous derivations (Equation 8. The results were compared to those from a code by Lockheed Corporation and to experimental results from page 89 of [297] (for a 4412 airfoil at 16 degrees incidence).

In addition. To test the validity of the Kutta condition the code was altered to accomodate .188. The only effect of choosing a different point at which to replace the tangency condition with a Kutta condition was that the spike moved from the midchord to that new point. causing irregularities near the trailing edge.14 and 8. The goal of the first case was to reproduce the lift coefficient of the Lockheed code. Grid studies were inconclusive but showed irregular solutions for fewer than 25 chordwise stations.73 degrees in the first case. The disagreement near the midpoint is due to the present method of replacing the tangency condition there with the Kutta condition. In the second case. In conclusion. the incidences were equal at 16 degrees. The difference between Figures 8. solutions near the trailing edge became irregular when 150 chordwise stations were used. it appears best to use between 50 and 100 chordwise stations. and the present code returned a lift coefficient of 2. 176 and with the Kutta condition replacing the tangency condition at the midpoint of the lower surfaces of the two airfoils. especially in light of the large incidence of 16 degrees. which was 2.291. and 16 degrees in the second. except near the trailing edge and near the midchord. The poor agreement between the panel codes and the experimental results must be due to viscous effects. Such proximity results in the strong local influence of nearby vortices. The two panel codes agree quite well. The behavior of the present code near the trailing edge probably has to do with the close proximity of other vortices to a given control point near the trailing edge. which is 14.15 is in the incidence used in the present code. this may be due to the small size of the panels and to the resulting close proximity of control points.

0 0.8 1.0 6.0 A '"~*- • Exp.12. x/c Figure 8.0 Chordwise Location.0 0.0 -2.4 0.6 0.0 A • Present Code 4.0 10.0 •\ • liOckheed Code X 2.14. Results.0 8.2 0. i from [ ] 0. Pressure distribution for a NACA 4412 .

0 Chordwise Location.8 1.12.0 0.0 A Exp.2 0.0 8. from ['2!)71 0.0 Lockheed Code 2.0 10. Results.15: Pressure distribution for a NACA 4412 .0 -2.4 0.0 • Present Code 4. x/c Figure 8.0 6.0 0.6 0.

Turning first to the effect of varying xcmaxi the chordwise location of the point . Preliminary Design of an Airfoil in Ground Effect The constant strength vortex panel code was extended to provide an interactive user interface so that it could be used for airfoil design. As the goal of the wind tunnel tests is to study airfoils in ground effect in the absense of flow separation. thicknesses.16 depicts a typical screen. As the trends are analyzed. The effect of varying the location of maximum camber. the amount of camber. The panel vortex strengths were antisymmetric and the code provided results for a wide range of cambers. 8. given a ground clearance ratio of 0. which is 7^ = —'JJV. so that the present method of using a Kutta point at the midpoint of a panel just downstream of the trailing edge is regarded as valid.18 and 8.17.5.1 and a lift coefficient of 1. incidences. respectively. The HOOPS graphics library was used to create a screen on which the user could change the geometric parameters of the problem. Certain trends can be identified based on these three figures.19. Buttons (rectangles in which the user could click the mouse) were added to allow the user to initiate the computation and display of the new airfoil shape and the corresponding pressure distribution. and ground clearances. and the thickness can be seen in Figures 8. Standard NACA airfoils were used for this purpose. the author set out to systematically study the effect of varying certain airfoil parameters. Figure 8. sharp pressure gradients must be avoided.The results were identical. one should therefore look for airfoils that display small pressure peaks near the leading edge. Once the code appeared to be performing satisfactorily. 179 the conventional Kutta condition.

Dftttnca (o v-vtcbf AOO REOftAW GAAPH COMPUTE PRESSURE CAPTURE T>HIS SCREEN Figure 8.16: Design code user interface . 180 TraMng Mg* a«trtrc« O.tZ 1 69477 DIf.

00 Chordwise Location.00 ' «W • 0) •$ O o 0.00 1 -0.00 m 0) M P4 -0.50 •H O •rl • ^' 1.60 0. x/c Figure 8.00 0.50 1 • NACA 5212 0) ''««A M 3 '6^ .00 -p c 0) 1.80 1.17: Effect of varying the niaxinuun camber location .50 22i22ii \ * NACA 5412 iiJ -1.20 0. 2.NACA 5312 m 0.20 0.40 0.

50 • NACA 6312 • • • ^ -1..00 01 CQ A NACA 5312 a> S -0.60 0.50 +) S 1.20 0.20 0.i •H ^ 0.50 f *5f 1 o 0 I • NACA 3312 1 (1) 'u.00 0.00 -0.18: EfTect of varying the camber . x/c Figure 8.00 Chordwise Location.40 0. ' NACA 4312 g 0.80 1. 1.00 * Aa •H •o A n o • OA L'n.

00 0.50 •p c: (U 1.40 0.00 -0.00 Chordwise Location.20 0.19: Effect of varying the thickness . 1.00 P (0 ' NACA 5312 0] OO CO (U A NACA 5314 M fU .60 0.0 .20 0.00 1 •H i AA • O• o O •H IH 0) 0. 5 0 ) k -1. x/c Figure 8.50 O O . • NACA 5310 0) U 0.80 1.

so that a twelve percent thickness was chosen. xcmax = 0.18 shows that increasing the camber has the combined advantages of generating a given lift at a smaller incidence and of reducing the pressure peak near the leading edge. twelve percent thickness.19).1 and at a lift coefficient of 1. It is for these reasons that a thick airfoil may be best for preventing flow separation. and the curve is smooth and displays no severe leading edge pressure peak. note that thicker airfoils generate a given lift at lower incidences and display a less severe pressure peak than do thinner airfoils. When the desirable airfoil parameters are brought together. .17). a NACA 5312 (five percent camber.5 is shown in Figure 8. Turning last to the pressure distributions for airfoils of differing thicknesses (Fig­ ure 8.It thus appears that a good chordwise location of the maximum camber might be thirty percent. if xcmax becomes too small. 184 of maximum camber (Figure 8. The corresponding incidence is 4. It was therefore chosen as a good candidate for wind tunnel testing. see Figure 8) results. On the other hand.21. A good choice for the amount of camber might therefore be five percent. note that the leading edge pressure peak grows as Xcmax is increased. Figure 8. The penalty for thick airfoils is that the form drag (pressure drag) increases. The pressure distribution for that airfoil operating at a ground clearance ratio of 0. a steep pressure gradient develops near xcmax.8 degrees.30.

60 0.20: NACA 5312 airfoil sliape .20 0.80 1.50 0. 00 x/c Figure 8.10 0.60 0.20 0.20 0.40 0.00 -0.40 0. 0.10 -0.00 0.30 o N 0.

50 • • •.1 . C£ = 1.40 0. 2.80 1.*• .60 0.50 • • 1 • • . O • • 0 0.21: Pressure distribution for NACA 5312.• • -1. .5. • • • • •3 ' • • 1. x/c Figure 8. ^ -0. m CO 0.00 -k 0.00 4J ^ •• •• •• • • • •• • 4 c u) 1. — 0.50 *-r-r d) * >^ 3 I ' ..20 0.00 0.00 —--T-: <w * • • Q) * i.00 Chordwise Location.00 ^— •.

Initially. Smoke injectors would be used to visualize the flow. For the record.1). and be equipped with pressure taps for pressure measurements. the author and his Major Professor decided against this method. the unevenness created by the railroad ties and gravel would be significant. the author's plan was to build a self- propelled rig that would operate on abandoned or low-usage railroads (see Figure 9. A wing would be mounted at some distance ahead of the front wheels. It would pass over the gravel. and testing of a free- flying model or water-tunnel model. Given the relatively small span of the wing (the inside edges of standard gauge rails are 56. grass. The trouble with the method is that it is difficult to find railroad segments that have a sufficiently smooth bed. A further disadvantage would be that the wing tips could not be fitted with effective winglets to suppress three-dimensional flow effects. apart). we did find some nice railroad . The three methods holding most promise for testing a wing near the ground are the moving-belt technique. or snow between the steel rails.5 inches. or 1435 millimeters. sand. Based on insufficient funding and on the difficulty of finding appropriate railroads nearby. each experimental approach has its advantages and disadvantages. the image method. CONSTRUCTION OF THE EXPERIMENTAL WIND TUNNEL MODEL Introduction and Initial Ideas As discussed in previous chapters. 187 CHAPTER 9.

Hopefully it will be possible to carry out these experiments in the future. Mike Weddell. Iowa. The segments belong to the Boone & Scenic Valley Railroad. The great advantage of this method would be that the ground effect is properly modeled. The symmetry plane models the presence of the ground. in the Spring of 1992. The rig was to consist of a four-wheeled chassis whose U-shape would be open toward the front to minimize interference with the oncoming flow. The wing would feature pressure taps for further data acquisition. and Pitot rakes for pressure measurements. Mounted to the parallel sides of the chassis would be seven-foot tall solid walls. which operates a steam locomotive on the tracks. In order to perform at least some experimental work it was decided (February. The design and construction of the wind tunnel model will be discussed in the following sections. 188 segments west of the city of Boone. Another of the author's plans (January. 1993) was to build a wheeled test rig that would be pushed over a level surface by an automobile or van (see Figure 9. Video cameras would be pointed at the plexiglass windows to capture the smoke filaments. As mentioned before. Mounted between the two parallel walls would be a prismatic wing. The author did not build the rig because of insufficient funding for experimental work.2). smoke injectors. Their chief mechanical officer. 1993) to build a wind tunnel model consisting of two symmetrically mounted pris­ matic wings. with the wing moving with respect to both the ground and the air. this is called the image method and produces a symmetry plane between the two wings. was most helpful. These would form a wind-tunnel­ like channel and would be interrupted by plexiglass observation windows. .

189 motor wing Figure 9.1: Self-propelled tracked test rig .

2: Test rig pushed by an automobile . 190 Figure 9.

Their diameter was 3/8 inch and they were located 0. 4. Each wing had a chordlength of nineteen inches and consisted of eleven spanwise pieces (see Figure 9. 12. 3/8 inch diameter drill rods with threaded ends were pushed through three of the holes in the eleven sections. which defined the span of the wing and its image. The fourth hole nearest the trailing edge was not used during assembly because the other three were sufficient to hold the wing sections together. Two of the pieces were made of aluminum 6061T6 and comprised the endpieces.5. the hole closest to the trailing edge was centered between the upper and lower airfoil surface. The four holes were used to align the pieces on the CNC (computer numerically controlled) mill and to fasten the pieces to each other during assembly.5 inches from the leading edge.3). They were 0.5 inches thick. All eleven sections had four reference holes at distances of 2. Eight of these were made of hardwood poplar and had a width of about two inches. and 14.5. respectively. . meaning that their cross sections had the same external shape at all spanwise locations.5. The eleventh and key part of every wing was a two-inch wide aluminum (6061T6) center section. It featured a very smooth exterior and tiny pressure taps. The wings were prismatic.5629 inches above the chordline. 191 Description of the Two Prismatic Wings The wind tunnel test section was 15 3/16 inches wide. During assembly. The reason for this odd dimension was that the holes needed to be colinear and parallel to the chordline to fit onto an existing fixture used for milling.

3: Top (main) wing during assembly . 192 spaghetti pressure tap tubing wooden sections aluminum center secondary spar aluminum end piece Figure 9.

The mill could operate manually or via data fed to it from a personal computer. and the location of the axis system is defined accordingly. which was here defined as the x-axis. Bridgeport is a division of Textron Incorporated. where the y-axis originated at the leading edge and pointed upwards.5629 inches in the positive y-direction. The mill was a three-axis movement machine with spindle speeds variable between 0 and 3. It used so-called G-codes that were part of the automatic programming tool (APT) language. Several sets of coordinate pairs were calculated to govern the numerous exterior and interior paths of the mill. The four holes formed a line parallel to the chordline. The CNC mill was used to perform the exterior cuts on this material to create the two endpieces. The line was 0.25 inches) of hardwood poplar that cost $60. a rectangular piece of aluminum (40 x 3. That fixture had four holes and was carefully aligned at the center of the mill table.25 x 2 inches) of 6061T6 aluminum at S30 each. The 20 x 2. and sixteen rectangular pieces (20 x 3 x 2. A typical milling session consisted of attaching a metal fixture to the moving mill table (see Figure 9). The CNC mill was a Bridgeport Series 1 machine. The half-inch thick aluminum block was cut in half and then to within one- quarter inch of the final exterior airfoil shape with a handsaw. Four holes were drilled into all wooden and metal pieces to create reference points for precision alignment.25 sides of the wooden pieces were evened on a planer to ensure two parallel sides for later alignment on the mill. A spacer .500 rpm. The airfoil was a NACA 5312 section. 193 Fabrication of the Two Wings The work began with two rectangular pieces (20 x 3.5 inches) at $15. All points were referenced to the origin at the leading edge.25 x 0. These data files were stored on a personal computer and downloaded to the mill when they were needed.

25 . The code's correctness was verified by executing it with the tool barely scratching the metal. Due to the thinness of the trailing edge.4 inch diameter bit was used to drill a 0. sanded once again. The next step was to mill out the interior of the top airfoil such that a 1/8 inch thick exterior wall remained around the 1/2 inch deep depression (see Figure 9. The tool then made a number of paths at increasing depths to produce the desired piece. and finally smoothened with steel wool. The tip of the tool was brought above the leading edge. since the strongest pressure variations occur there. stained. the exterior wall thickness there was further reduced there to make more space for pressure taps. and that point served as the origin of the coordinate system.2 inch to 0. The wooden pieces and the two center aluminum pieces were also cut to within 1/4 inch of their final shape using the bandsaw. Pressure tap locations were chosen to achieve the best resolution on the top side downstream of the leading edge. The wooden pieces were sanded. These cuts consisted of rough passes with a 3/4 inch diameter endmill and fine passes with a 1/8 inch endmill. The endpieces served as templates both during the cutting and sanding of the wooden pieces. followed by the piece that was to be milled. For each pressure tap.5). A series of three large holes was drilled at approximately the center of half of the wooden sections to accomodate the circular aluminum spars and the pressure tap tubing. The milling code was then downloaded from the personal computer. a 0. covered with sand­ ing sealer. 194 plate of aluminum was laid on top of the fixture. and the rest of the material was removed with a 3/4 inch side-cutting endmill on the CNC mill. The pieces were aligned and fastened by inserting at least two 3/8 inch diameter steel bolts into the four holes common to all pieces.

195 Figure 9.4: CNC mill with wooden wing section .

196 channels for the pressure tap tubing aluminum section for the image wing plastic pressure tap tube •f^steel tube connected to pressure tap aluminum section for the top wing Figure 9.5: The two aluminum center sections during construction .

Both spars were 12-inch long pieces of 1000-series aluminum conduit with internal and external diameters of 1.9 inches from the x-axis. respectively.25 inch hole at 6. with friction holding it in place. a 1/2 inch deep and 1/4 inch wide channel was milled from a point 1/4 inch beneath each tap to a common point near the middle of the airfoil.25 inches. The two center aluminum wing sections received a 1. The reason for locating pressure taps in the image airfoil was to guarantee the symmetric alignment of the two wings during the experiment.025 inch.35 inch long steel tube was then prepared for each pressure tap. Instead. That point was 6. The purpose of the channels and the center hole was to house the pressure tap tubing. A two-foot clear piece of plastic pressure tap tubing (0. To minimize effort. 197 inch deep hole in the spanwise direction.5 inches downstream of the leading edge and featured a 5/8 inch hole. Each tube Wcis tapped into a spanwise hole to a depth of 0. The tube had an external diameter of 0. Nine tap locations were chosen for this purpose.5 inches from the leading edge and 0. It was then threaded through the hollow center of the respective circular main spar and labeled. The nine pressure taps were drilled in the same way as described previously.1 inch.0 and 1. to connect to the previously drilled spanwise hole.2 inch diameter bit (drill number 75) was used to drill a hole through the pressure tap location into the airfoil. The resulting hole was 1/4 inch beneath its corresponding tap. the interior was not milled out as was done for the main airfoil. Both spars were then welded to their respective center alu­ minum wing section at Iowa State University's Metals Development Shop (at a cost .042 inch and an internal diameter of 0.065 inch external diam­ eter) was attached to the steel tube belonging to each pressure tap. A 0. A 0.

One end of each of these drill rods had been turned down and threaded on the lathe beforehand (a 5/16-inch tap with a pitch of 24 was used). and a uniform lift distribution. Also. a cantilevered Euler-Bernoulli beam. a 1/8 inch hole had been drilled at the other end of each drill rod to accommodate a 1. All pieces were finally bolted together using three 3/8-inch diameter drill rods for each wing. After assembly. a factor of safety of 1. standard atmospheric conditions. Each rod served to both measure and adjust the incidence of the corresponding wing. Tliese holes could later accomodate a steel rod that was exactly parallel to the chordline. four pieces of felt were cut in the shape of the NACA 5312 section and glued to the tips of the wings. 198 of $30).0 inch long steel roll pin. the entire wood surface of both wings was painted twice with a clear polyurethane lacquer to achieve a smooth surface finish. resulting in a spanwise wing loading of 170 lb/ft. and that the wings had to exactly fit the tunnel's width. The stresses at the root of the circular main spars (near the wing suspension mech­ anism) were calculated assuming a lift coefficient of 1. The predicted lift was 217 pounds. . Last. which is safe assuming a tensile elastic strength of 24. The reason was that the pressure taps had to be at the exact center of the wind tunnel. These served to seal the flow and to prevent scratching of the plexiglass wind tunnel wall. Holes were drilled into the spars at the end furthest from the airfoils.5. three of the wooden sections had to be narrowed using a bandsaw and a planer. Just prior to final assembly. a freestream ve­ locity of 200 ft/s. The predicted stress at the root was just under 11.000 psi. These roll pins fit into slots milled into the two aluminum endpiece sections.5.000 psi for 1000-series aluminum.

The threaded rod had a 3/4 inch diameter and a fine thread with a pitch of 16.7).6 and 9. This slider block was guided by a linear slider bearing running along a sturdy shaft. models are attached to a circular plug that fits into a hole in the tunnel wall.75 inch depth and a 0. They were then welded together. so that a clockwise turn of the crank caused the two wings to move further apart. The top half had left-hand threads. The nut was kept in place by a machine tap screw in the slider block. 199 Fabrication of the Wing Mounting Plug and Traversing Mechanism To be mounted in the wind tunnel.5 inches. Toward the front of the block was a 3/4 inch hole that housed a 3/4 inch nut at the bottom side. In the final design. and a metal pin was inserted into the holes to align the two halves. These had an interior diameter of 0. Two nuts were brought flush with the ends. while the bottom half had right-hand threads. The bottom end of the rod also had to accomodate one of the miter gears. a wing and its mirror image needed to be attached to the disk such that they could be moved toward or away from each other in a symmetric manner. In this case. Each of the halves was originally 3.3 inch diameter in the longitudinal direction of each. A threaded rod ran through the 3/4-inch hole and was constrained by the 3/4-inch nut. The two miter . The two halves of the turning rod were oppositely threaded so that the motion of the slider blocks was always opposite. The two halves were joined by first drilling a hole with a 0. so that the rod had to be turned down to that diameter on the lathe.0 feet long and cost $10. The two-piece threaded rod was supported by ball bearings on both ends. Turning the threaded rod by means of a crank resulted in vertical motion of the two slider blocks (see Figures 9. the circular main spar of each wing was bolted into a slider block. which also had a half-inch bore (internal diameter).

6: Outside of the model mounting plug . 200 fastener for holding the mounting plug^^: to the wind tunnel|i| wall mitre gears crank Figure 9.

201 rubber sheets (seals) covering slot threaded rod slider block holding main spar of the image wing the main wing slider block holding the image wing rectangular plug spaghetti tubing slider bearing rod Figure 9.7: Traversing mechanism with slider blocks .

respectively ($27 for the brushes). The larger disk was only 0. 202 gears had a twenty degree pressure angle. and cost $31. The side facing the wind tunnel was painted dull black. and painted with a clear lacquer on the exterior. the threaded rod. a hub (exterior) diameter of 1. the wing suspension system consisting of the two movable slider blocks. The assembled tunnel plug was painted with sanding sealer. The rod was heated using an acetylene torch (Civil Engineering Department) .5 inch diameter steel rod. and cost $13 each. The linear bearing shaft was made of induction-hardened alloy steel. Each had 24 teeth. Once the plug was painted.0 inch long flanges and kept the slider blocks at a distance of 0.6 inches and fit into the hole in the wind tunnel wall.3 inches thick and had a diameter of 34. sanded. thus producing an overlap on the outside of the tunnel wall. The 3.0 inches. The linear bearing consisted of recirculating linear ball bearings with a friction coefficient of 0. and the circular pillow block that was housed in the slider block.0 inch.75 inch. and the two end pieces.007. The smaller disk had a diameter of 30 inches and a thickness of 1. had a di­ ameter of 0. the linear bearing shaft. resulting in a pitch of 16 teeth per inch.5 inch-long angle irons had 1. with a grid of yellow reference lines at six-inch intervals. That slot was air-sealed and streamlined by means of two pairs of rubber strips and a pair of nylon brushes. The plug featured a 1 3/4 inch wide and 25 inch long slot through which the main spar traveled. The plug consisted of two circular disks. The crank was fabricated from a 15 inch long piece of 0. was bolted to the disk using two steel angle irons. The circular wind tunnel plug was made of 3/8 inch and 3/4 inch plywood sheets costing $37.5 inches from the disk surface.

25 inches). Furthermore. and support structure were approximately 15.25-inch diameter hole in the slider block and tightened by two set screws in the slider block. the existence of the symmetry plane between the airfoils had to be verified during the experiment.0 x 3. These each had a hole to accomodate the 1/2 inch crank and were fastened to the bottom of the wing support mechanism such that the miter gears made proper contact. Thus. the need arose to design an unobtrusive Pitot rake that could fit between the wings and that . This rod served both as an indicator of the wing incidence and to facilitate adjustment of the wing incidence. 203 and bent into the shape of a crank. One of the miter gears was slid onto the other end and fastened with a set screw. A one-foot long 3/8-inch diameter steel rod was pushed into the holes in the circular wing spar. bringing the total weight to 88 pounds. Just prior to final assembly.5 x 0. two handles were fastened to the outside of the circular wing support structure to facilitate handling of the structure. and 55 pounds. At this point the wings and the wing support were weighed. Fabrication of the Pitot Rake One desired result from this experiment was the velocity distribution near the airfoils. It was then inserted into the 1. Finally the top wing was attached to the wing support. 18. The crank was supported by two pieces of aluminum (2. The weights of the top wing. This procedure was repeated for the bottom wing. respectively. and especially between them. bottom wing. Its circular main spar was pushed through the two rubber seals ajid the set of brushes covering the slot in the wing support. A four-inch piece of aluminum tubing was slid over one end of the crank to serve as a handle.

The Pitot rake was connected to this water manometer board via plastic tubing.04-inch external diameter) arranged as shown in Figure 9. These Pitot tubes were aligned using scrap pieces of steel tubing and glued together with two-component epoxy. Fortunately there was already a rectangular opening about one chordlength downstream of the two airfoils. This support piece had two parallel rectangular appendages that pointed downward from the rake. The first design had the rake fixture penetrating the tunnel from the top wall.65 inch thick plywood frame had outside dimensions of 47. Its main part consisted of a 0.35 inch thick piece of acrylic and a 0. Flow visualization experiments were also videotaped and photographed from there. 204 could also serve as a wake rake. The two appendages were later sandwiched between two parts of the rake support structure that were fastened to each other with two setscrews.0 inches. Plastic tubing (0. An attachment structure was sawed out of a piece of 6061 aluminum and glued to the rake. This created a window behind which a manometer board was installed. The Pitot rake itself consisted of thirteen stainless steel tubes (0. It was possible to monitor the manometer board from the side of the tunnel that has the transparent wall. A suitable rectangular plug was built.0 by 25. A rectangular 14 1/4 inch by 19 1/4 inch (height) opening was cut into the lower half of the particle board.8. Its 0. This was the side from which the tunnel and the data acquisition equipment was operated.55 inch thick piece of high-density particle board. it was decided to enter the tunnel from the side. Its height was 43 1/8 inches and its width 20 3/4 inches. But since this requires a long support structure and great lengths of plastic tubing.065-inch external diameter) was attached to each Pitot tube and the body of the .

75 inch diameter hollow aluminum tube. filled.2- inch diameter holes drilled at one-inch intervals.4 inches long and had 0. As mentioned above.2-inch diameter holes were drilled at one-inch intervals and the 39-inch long tube was bent by 90 degrees 10 inches from its end. The main objective of this design was to present the oncoming flow with the least possible frontal area. A second aluminum tube was only 4.2 inch threaded holes at the center of both of its ends. The holes were perpendicular to the tunnel walls. The streamwise one to which the rake was attached was a 0. The geometry of the rake is such that it fits under a wing inclined at 14 degrees. The dried assembly was sanded. The tube had twenty-seven 0. The holes in the streamwise tube and in the solid steel tube allowed the connector to be moved. the rake was attached to its support structure by sandwich­ ing its two appendages between two pieces of that structure. The third tube was a 0. thus allowing movement of the rake itself. The support structure consisted of three metal tubes. 205 rake was covered with fiberglass matts soaked in structural adhesive. They exited the tunnel through a small hole in the . The plastic tubes from the Pitot wake rake were passed through the streamwise tube and taped to the outside of the other tubes. One-inch bolts were used to bolt the short tube to the first tube. The assembly consisting of the rake. and painted to achieve a smooth surface. Twenty-eight 0. A solid 5.75-inch diameter solid steel tube. and the short connector piece was next bolted to the vertical solid steel tube. Its ends were also milled such that it fit snugly against the side of the first tube. the streamwise tube.0-inch long aluminum nose cone was machined and friction-mounted to it. The short side pen­ etrated the rectangular wind tunnel plug and was fastened by means of a set screw when the long side was in a vertical position.

206 Pitot rake '' ^ f i Figure 9.8. Model and Pitot rake in the wind tunnel test section .

tools. The author spent countless additional hours searching for materials. But perhaps that is the nature of experimental work. funding. people. . and equipment. Construction of the two wings. the raising mechanism. the tunnel plug. and the Pitot rake required 56 hours of technician time and 223 hours of the author's time. 207 rectangular wind tunnel plug.

The tunnel's name derives from the fact that most experiments are carried out by faculty and graduate student researchers.1). accelerated as it traversed the nozzle. The wind tunnel was open-ended. EXPERIMENTAL PROCEDURE Description of the Wind Tunnel After the wings were mounted to the mounting disli. 208 CHAPTER 10. that tunnel was housed in room 130 of the George R. Town Engineering Building. the entire assembly was inserted into a hole in the sidewall of the test section of the wind tunnel. At the time of this experiment. encountered the propeller. decelerated as it passed through the diffuser. flowed through the test section at a constant velocity. The tunnel was capable of a maximum velocity of approximately 200 fps. The air moved through several layers of filters located in the settling chamber (reservoir).. The wind tunnel used for this experiment was the Iowa State University Department of Aerospace Engineering and Engineering Mechanics' Low-Turbulence Research Wind Tunnel. and was finally exhausted through the downstream end of the wind tunnel (see Figure 10. The tunnel was designed by Professor William James of ISU's Aerospace Engineering Department. which . A propeller driven by an electric motor was located downstream of the test section and causes air to be sucked in at the upstream end. meaning that it did not recirculate air.

1: Wind tunnel. and data acquisition system . smoke injector. •209 smoke supply hose smoke injection tube test section model computer scanivalve controllers & amplifiers . displays Figure 10.

The difFusor was 6 feet long and its downstream end was 38 inches wide and 32 inches high. presumably to compensate for boundary layer growth. This solid wall had several rectangular access holes and a 30. with steel walls toward the upstream and downstream ends. The top and bottom walls were acrylic as well. There was also a Pilot static probe penetrating the bottom wall 13. .0 inches into the tunnel and its streamwise dimension was 1. while the trailing edge was 50.2). The movable center section of the wind tunnel consisted of a steel frame. and with acrylic or wooden walls near the center. The left tunnel wall was made of wood. The compressor was ten feet long and its upstream end measured 36 inches in width and 30 inches in height.0 inches from the upstream end of the test section. The test section was 8 feet 2 inches long and its internal cross section at the center was 74 inches high and 15 3/8 inches wide.0 inch chord was mounted in this circular opening. its leading edge was 29. The section was 24 feet 2 inches long.0 inch. The probe penetrated 5. The test section's width actually varied linearly with the length. The tunnel speed was controlled by varying the blade pitch angle of the sixteen-bladed propeller. As one looked in a downstream direction at the test section. •210 corresponded to Reynolds numbers on the order of two milHon.0 inch diameter circular opening into which model mounts were inserted. but Ihey featured several access holes. with steel structural members. When a wing with a 19. the right side con­ sisted of an uninterrupted two-inch thick acrylic plate in a steel frame.0 inches from the downstream end (see Figure 10.0 inches from the test section's upstream end.

211 top (main) wing wind tunnel fan clear acrylic wall image (bottom) wing mounting plug Figure 10.2: Front view of model mounted inside the wind tunnel .

Odd-Even Decoder). . three digital inputs. the serial numbers for the scanivalve. For Scanivalve I. and 027207. serial number 3115A06896). and amplifier were 48S9-3011. 2310 Signal Conditioning Amplifier). The reason that the Pitot static tube was connected twice to each transducer was for verification purposes. 44 pressure tap tubes were connected to each transducer. One of the two scanivaives is shown in Figure 10. controller. and four digital outputs). During the experiment. display. as is the calibration transducer mentioned before. see Figure 10.3. respectively. and 028221. The controllers were in turn connected to a data acquisition board in the personal computer (Hewlett-Packard Vectra 386/25. The readings indicated whether the transducer was stepping properly and whether the results remained consistent in the course of the experiment.4. The output cables of each transducer were connected to an amplifier (Vishay Instruments. 717. Solenoid Controller CTLR10P/S2-S6) and to a channel display (Fx Scanner Position Display. 857BINY. the serial numbers were 48S2- 153. For a diagram of the data acquisition equipment. For Scanivalve II. Each transducer had 48 chan­ nels and was connected to a controller (Scanivalve Corporation. Model DT-A02. 997BINY. which was in turn plugged into the computer's data acquisition board (MetraByte DAS-8 board with eight ana­ log inputs. The other four channels were occupied by freestream static and stagnation pressure lines from the Pitot static tube at the bottom front of the test section. 797. 212 Data Acquisition Equipment and Instrumentation The plastic pressure tap tubes were connected to two Scanivalve pressure trans­ ducers (Scanivalve SS2 48).

the change in the water level of the manometer and the number of .3 psid. This transducer's output was amplified and fed into the computer's data acquisition board. serial number 029517) was set at a gain of 100 and an excitation of 7. PM5T+/-0. where'd' stands for differential) that served as the calibration transducer. accuracy of 0.3: Scanivalve and calibration transducer Calibration Procedure The Pitot static tube at the bottom front of the wind tunnel test section was connected to a differential pressure transducer (Statham 13739. and the filter was set at 10. 213 Figure 10. The amplifier (2310 Vishay. At every speed. The wind tunnel speed was then increased in increments. The system was zeroed with the wind tunnel switched off. The lines from the static probe were also attached to a crank-type water manometer.3-350.

4: Diagram of the data acquisition system .A02 board J SN 3115A06896 excitation = TV oils filter = 10 Figure 10. HP Vectra 386/25 gain = 100 sN 028221 SN 027207 ModelDT .3011 •• silver SN 857BINY cable: controller I SN 717 drive cable I iin'inii)! calibration amplifier amplifier amplifier slot 1 computer CPU SN 029517 II I DAQ . Statham 13739 display II Ploo PM5T ± 0. 214 Poo ^ ' calibration differential pressure JL transducer.3 .350 SN 997BINY . controller II : SN D S H ! Scanivalve II 0P4 OP3 4852 .153 drive cable II j Scanivalve I display I ' 4859 .

the crank-type water manometer was disconnected. Before each run. It was written in the Basic language and prompted the user for the atmospheric pressure. At any point.12745 (10. the corresponding pressure differential is readily obtained. inlet temperature. A computer program called "GR0UND2" had been written by the technician. 1994. The data ac­ quisition system was powered up and the two scanivalves were returned to the first (home) channel. geometric location of the pressure taps.1) where the pressure change Apoo is given in pounds per square foot and where the counts are normalized. which had been set at 100. The change in water level Ah was converted to a pressure change Apco using the hydrostatic equation. number of scanivalve channels.30336Ac^^^ + 0. A least square fit was performed on the 14 resulting data pairs to obtain a calibration equation. 215 counts displayed on the computer screen were recorded. given a number of counts from the calibration transducer. On completion of the calibration procedure. Mr. William Jensen.5833 ft).0 hours of wind tunnel operation. and required 13. and the counts from the calibration transducer were divided by the amplifier gain. length of the model (1. the barometric pressure was recorded from a mercury manome­ ter and the temperature near the wind tunnel inlet was measured. Experimental Procedure The experiments were carried out between April 8th and May 4th. That equation was: Apoo = 3. to serve as the interface between the experimenter and the data acquisition system. sampling .

it prompted the user to zero the two scanivalve outputs using the two respective Vishay amplifiers while the wind tunnel was still stopped and the garage doors shielding the inlet and outlet were still closed. and Reynolds number. If h is the height of the mercury column in inches and B ambient pressure in lb application of the hydrostatic equation vields: . As the propeller blade pitch of the tunnel was changed. then poo = ambient density q can be found with the help of the perfect gas law: PAMB = PAMB^'^AMB (10. density. Bernoulli's equation could then be used to relate the pressure differential to the flow velocity: (10. one simply adds 460 degrees to the temperature measured in degrees Fahrenheit. which was measuring the pressure difference sensed by the Pitot static tube. and name of the output file. the program computed the viscosity. To arrive at the temperature in degrees Rankine. The program then prompted the user to open the garage doors and to start the wind tunnel. flow velocity.2) If one assumes that the density is constant everywhere.1 allowed the counts to be interpreted as a pressure differ­ ential. The flow velocity was based on the counts produced by the cali­ bration transducer. 216 rate (200 was entered for all cases). The ambient pressure g was found with the help of a mercury barometer.3) ft-lbc where R is the universal gas ^AMB ambient temperature in degrees Rankine. Following that. Equation 10.

Reynolds number. and it was sometimes necessary to adjust the gain of the amplifier in order to generate maximum counts whose magnitude did not exceed 2048 and whose magnitude did not fall below 1000.515/i. the computer stored the barometric pressure. The plot allowed a first check of the data and also showed the nine points from the pressure taps on the image wing. the gain settings. the count from the calibration transducer (Ac in Equation 10. Whenever the gain was changed.4) Once the velocity was determined. the temperature. and model length have already been found. (10. the Reynolds number was computed according to „ tie — poovoo^ (10. the system had to be rezeroed and the run had to be repeated. the velocity.8 million in most cases). the incidence. and of course the counts that corresponded to each . 217 PAMB = PHg9^ = 70.1). and the air viscosity was computed based on the ambient temperature. The maximum magnitude of the count axis was 2048. and trailing edge ground clearance ratio were recorded on paper. Also. Once the proper Reynolds number had been attained (1. The resulting plot was proportional to the pressure distribution.5) Moo Here the density. the program instructed the data acquisition board of the computer to step through the scanivalves' channels and to record the data. After every step of the transducers the corresponding count was plotted at its streamwise location on the computer screen. velocity. For each run. Symmetry of the flow was assumed whenever the nine points fit in with the data from the top wing.

The reason that the two scanivalves were distinguished was that a given pressure differential did not result in the same count from both trans­ ducers. For both cases. Apj = PI — Poo 5 also appears in the definition of the pressure coefficient: Cpi = ^-^ = ^.1 that the calibration transducer count corresponded to the pressure differential Apoo sensed by the Pitot static tube. 218 of the pressure taps. The author wrote a Fortran computer program to convert the experimental data into pressure coefficients. Acoo. (10.7) ^ Qoo Qoo But Ap^ is known from Equation 10.8) qoo Acoo qoo The above equation allowed the count from a given scanivalve channel to be converted to the corresponding pressure coefficient. The pressure differential.At this point it was already known from the calibration Equation 10. Based on the assumption that the count rises linearly with increasing pressure one could thus write ^Pi = (10-6) ^Coo to process the count Ac^ for a given scanivalve channel i. . so that Api _ Apoo Acj c . The data acquisition system was zeroed before the tunnel was started.= = (10.6. together with their streamwise locations. The first step for each scanivalve was to average the two counts for the static freestream pressure and the two counts for the stagna­ tion freestream pressure. so that Acj = Cj. the average static pressure count was subtracted from the average stagnation pressure count to obtain an average pressure count differential.

5: Aerodynamic resultants acting on panel i . 219 voon ^mLEi Figure 10.

-)i where is defined as positive clockwise.9) where a is the incidence and 9^ is the surface orientation angle of panel i. To find the lift. simply sum the contributions from all n panels. and moment coefficients for the entire airfoil.10) The moment about the leading edge due to the pressure force is the trans­ verse component of times the moment arm r^: ^mLEi = +360-7. for example: = E C„. the angle between the x-axis and the tangent to the panel.5. In a similar manner. drag. that is. The magnitude of the resultant force is cjij = The lift acting on the panel is the negative of the component of c^^- along the perpendicular to the freestream velocity vector: ~ (10. 220 Once the pressure coefficients were known. so it is normal and pushes down onto the panel.7. A typical panel i of length /j is shown in Figure 10.12) i=l . the nondimensional drag acting on panel i is the component of along the freestream velocity vector: = —c^j5m(a + (10. according to the right-hand rule.-)) = . The nondimensional resultant force is due to the nondimensional pressure acting on the panel. (10. the aerodynamic resultants could be computed.

1 inches.3 inches. This behavior was not consistent and the model's alignment had to be checked during each run. Also. the top wing twisted by 1. respectively. those walls may have a significant impact on the pressure distributions. It should be mentioned that the digital water level used to measure the incidence was accurate to within 0. but did not complete that work at the time of the writing of this dissertation. Another source of error was the presence of the top and bottom walls of the wind tunnel test section. 221 Problems with the Experimental Procedure The biggest problem with the model was that it rotated and translated as the aerodynamic load increased during the start-up. This caused the author to take great care to monitor the model alignment thereafter. There . The model's incidence and height were accurate to within about 0. Upon inspection it was found that both slider blocks had play. Nevertheless. there were several sources of perturbances. During one of the first runs. The author began writing a panel code to account for the presence of those walls. particularly on the top surfaces of the wings.1 degrees. This had some influence on the results and may explain why the top surface pressures near the front of the main and image wings differed slightly in some cases.4 degrees and bent by 0.2 degrees and 0. the setscrews that fixed the wing's main spar to the sliding block of the traversing mechanism slipped on occasion. there may have been some torsional and bending displacement due to the elastic nature of the wing materials. Finally. Even though the wind tunnel was designed for low-turbulence operation. and that the optical characteristics of the acrylic wind tunnel wall made it difficult to make accurate measurements. The test section featured many grooves between adjacent wall segments and many patched holes from previous experiments. particularly in a pitching sense.

the two results were very close. Fog Machine 1500) that heated theatrical smoke (Selig Chemical Industries. The diameter of the tube cre­ ated a considerable obstacle to the flow. Wind gusts near the entrance and outlet of the wind tunnel were responsible for some error. since a primary aim of this dissertation was a full explanation of the ground effect flow phenomenon.. 222 was also a slight gap between the wing mounting disk and the wind tunnel wall. and air leaked into the tunnel through the brushes and double rubber seals covering the slot that allowed the wings to move up and down. . On one occasion.6-inch diameter copper tube. particularly when lower Reynolds numbers were investigated. It is difficult to assess the reliability of the data acquisition equipment. Fog Juice fogging compound) and sent it through a four-inch diameter clothes dryer duct to a 0. Par­ ticularly on humid days. but it was needed in order to supply enough smoke for a meaningful visualization. The smoke injection equipment consisted of a theatrical smoke generator (Rosco Laboratories. However. As will be discussed later. Inc.5-foot long copper tube penetrated the wind tunnel through the top wall of the test section. the computer output was checked against simultaneous readings from a water manometer. 22 inches upstream of the model. the readings from the transducers seemed to drift and the system often had to be rezeroed. The 7. any smoke visualization was a welcome addition to the experimental methods. Smoke Visualization It was known before the experiment that smoke flow visualization in this wind tunnel was difficult given its relatively high speed and the simple equipment that was available.

which is where the author was employed throughout his doctoral program. The smoke was sucked into the test section since the pressure there was lower than the ambient pressure. These locations were achieved by pushing the injector through the rubber and brush seals of the wing mounting disk. it was difficult to see the smoke flow between the wings because that region was not well lit. The author rotated the tube slightly to minimize the effect of the vibrating vertical tube on the smoke filament. If on the other hand the injector penetrated into the test section. Since the upstream injection of smoke did not allow for a good visualization of the underbody flow and of any separated flow regions. The smoke thus entered the flow 14 inches upstream of the model. The smoke provided some very rough visualization of the events occurring in the flow. it immediately cooled and grew in diameter and thus became less visible. thus making visualization difficult. then part of the smoke was carried away by the wall boundary layer. The problem was that once the smoke filament left the injector. The path of the smoke was recorded with a Hi-8 camcorder borrowed from the ISU Courseware Development Studio. the final eight-inch piece of the tube was ahgned with the freestream and acted as an injector. A further problem was that the injection tube comprised a vibrating obstacle to the freestream. The problems with this method were that if the injector was flush with the tunnel wall. A final method of smoke injection was to drill holes through the wing mounting . then it exerted an influence upon the flow. Finally. In this way. the injector was moved to two other locations just above and below the main wing. This was probably due to vortices that were shed off the tube. 223 It could be moved vertically and was bent 90 degrees near its bottom end.

Above all. It was easy to align and did not appear to exert an excessive influence on the flow around the two wings. The hope was that in addition to yielding wake data. Instead. Pitot Rake Experiments The disadvantage of pressure taps at the airfoil surface is that they only yield surface pressures and velocities.180. But it was somewhat easier to see separated flow regions. a Pitot rake was built as described earlier.187. respectively. but no information about the rest of the flow field. single wing with rake. and 1. The Hft coefficients for the three cases were 1. In all three cases. and to avoid smoke being carried away in the wall boundary layer.185. it would provide velocity profiles at other locations throughout the flow field. 224 disk just above the trailing edge of the top wing. Both laser-Doppler flow velocimetry and hot-wire anemometry were considered to obtain such data. It was still difficult to keep the injector from influencing the flow. and single wing with rake and with streamlined cowling for the vertical steel tube were investigated. The rake seemed to function well particularly when used for the wake survey. Despite all of the shortcomings of the above smoke injection methods. and whether or not the vertical steel tube was streamlined. More specifically. they showed that there was no recirculating air underneath the front portion of a wing in ground effect. As an inexpensive alternative. the single wing was inclined at six degree with respect . the air behaved as if it were in a converging duct. but they were deemed too expensive and time-consuming. 1. the cases of single wing without rake. The pressure distributions were virtually identical whether or not the rake wais inserted. they did provide some insights into the ground effect phenomenon.

In addition. Also. It is for these reasons that the information from the rake was treated in a qualitative rather than quantitative manner. leaks became evident from at least one of the manometers. the water levels stayed well within the range of the water manometer board.netrated further upstream. alignment of the Pitot rake became increasingly difficult and it tended to vibrate more as it >. When the flow field near the top wing was investigated some time after the wake survey. A final problem occurred in the vicinity of the wing. only a few of the 13 parallel probes of the rake lined up with the direction of the flow. 225 to the freestream. .

Presentation of Nunierical Results Results from the inviscid panel code discussed in Chapter 8 are presented in this section. Emphasis is on the design case. the plots show Z/'. Emphasis is further placed on the generation of lift. As a note on nomenclature. Less emphasis is placed on pitching moment and drag data for now. which occurs when the ground clearance below the wing's trailing edge is ten percent of the chord and when the incidence is 0. Information on lifting forces is also crucial in predicting the height stability of the vehicle. Unless otherwise noted. the ground effect flow phenomenon is discussed based on the experimental results. and M' . These will become important when the pitch stability and thrust requirements are analyzed. since an aerodynamic-levitation train can only function if sufficient lift is available at design speeds. After the presentation of the results. RESULTS AND DISCUSSION Introduction Both computational and experimental findings for the NACA 5312 airfoil in ground effect are presented in this chapter.7 degrees). D'. 226 CHAPTER 11.1 radians (about 5. 75 chordwise stations (150 panels) were used and the tangency condition at the chordwise midpoint on the airfoil's underside was replaced with the Kutta Condition.

2 depicts the variation of the lift coefficient with incidence. the flow underneath the wing becomes increasingly choked as the wing descends. The solid line represents an isolated airfoil and is a straight line. The squares represent a wing at a ground clearance ratio of 0. the surface pressures become less negative. the variation of the lift coefficient with trailing edge ground clearance is plotted in Figure 11. These results were confirmed by experiment. and pitching moment coefficient about the leading edge. following convention. Continuing the description of the ground effect in the most general terms. the lift rises more slowly at higher inclinations. respectievly. This is so that the negative pressure coefficients of the wing's upper surface appear at the top of the figure. Given equal increments in inclination.1 radians (5. and cm for the lift coefficient. . Once the wing descends below one chordlength. The figure represents the design case. one must turn to the variation of the wing's surface pressure distribution with ground clearance. Also note that the panel code overpredicts the lift coefficient at all ground clearances. and hence the tangential velocities decrease.3 shows that variation for a NACA 5312 airfoil at 0. To describe the ground effect in the most general terms.1 and show a distinctly nonlinear behavior.1. drag coefficient. Figure 11.7 degrees) incidence.1 radians. with a = 0. As the wing approaches the ground. The influence of the ground does not become significant unless the wing is within one chordlength of the ground (ftJ-= 1). As might be suspected intuitively. 227 instead of c^. To understand the exact fluid mechanical mechanism by which the lift increases as the ground is approached. the lift rises significantly. Also note that. Fig­ ure 11. But the behavior of the flow over the top surface may come as a surprise. the graphs showing pressure distributions have vertical axes that are turned upside down. c^.

this is known as the ram-efFect since it involves the choking of the flow between the wing and the ground.5 shows the variation of the pressure distribution with incidence for an isolated airfoil. the lift generated by the top surface decreases while the lift generated by the bottom surface increases. As the wing's nose rises. To appreciate the influence of the ground. both between the wing and the ground and above the wing. 228 This means that the flow slows down all around the airfoil. Compared to the case with ground effect. a. increases. This demonstrates that as the wing approaches the ground. Figure 11. the preceding figure (Figure 11. the flow soon begins to choke and the pressure rises. more fluid needs to move over the top surface. This is the point where the tangency condition was replaced by the Kutta condition. In Figure 11.4 shows the variation of the wing surface pressure distribution with incidence at a constant design trailing edge ground clearance ratio {hrpj^ — 0. the code was repeatedly .4) is repeated. Since the phenomenon is more pronounced on the bottom surface. resulting in stronger suction and more lift there. resulting in less pressurization and less lift there.1). The velocities on the bottom surface are also faster for the isolated airfoil. the velocities on the top surface are faster. To study the effect of varying the number of panels. but for an isolated airfoil. Figure 11. The lift from the bottom surface becomes proportionately stronger as the ground is approached. But since the trailing edge gap is quite small.3 note the irregularity near the midpoint of the wing underside. the wing attempts to scoop up jnore fluid and to pass the fluid between itself and the ground. so the velocities there rise and the pressure coefficient becomes more negative. As the incidence. the net result is that the wing in ground effect produces a higher lift.

By contrast.6 shows results for 50. and 100 chordwise stations. On the top surface the velocities for the 100-panel Ccise are slightly slower than for the 150. .1 ground clearance ratio). Experimental results are also included for reference. Figure 11.9 %. 229 executed for the design case (5. These differences are reflected in the lift coefficients. 75.and 200-panel cases. 150. The discrepancy between the computational and experimental results will be discussed later. which differ by 1. Whereas the results for 150 and 200 panels are virtually identical. they are faster on the bottom surface.7 degrees and 0. which corresponds to 100. and 200 panels. and particularly at the top near the leading edge. those for 100 panels differ slightly over the front third of the wing.

5 • Panel Code Results Experinnental Results 0.01 0.1 1 10 100 1000 Trailing Edge Ground Clearance.1: NACA 5312.001 0. h/c Figure 11. a = lift versus ground clearance .

2. lift curve with and without ground effect .2: NACA 5312. Alpha (degrees) Figure 11.5 4J c (D •rl 1. panel code results.5 -5 0 5 10 15 Inclination.5 •ri -0.5 O •H With Ground Effect tH 0) o No Ground Effect O +J 0.

60 0. x/c Figure 11.50 o •> e i' 1•«•••••' •• » • • • • • • # • • • • • ••• •• A A-.00 = m h/c 0. -1.h/c 2. = ••••• . NACA 5312.05 (1) O = O • h/c 0.0 -0.00 au = • h/c 100.00 4i c 0) •H = O • h/c 0. pressure distributions for various ground clearances .50 (1) = U h/c 0.00 0.40 0.2 3 mmit m 0.5 Q) to CO u = to P4 I•Da ^ h/c 1. 2. a = G''.00 Chordwise Location.1 0.3.80 1.0 0.0 •H = h/c 0. panel code results.20 0.0 AAA AA 4 A A—A_A_A_ALA A A A A» A ^ A A A A A A .

'V -0.4 = 0. L' = 2. panel code results.4.1 = Alpha 4.00 • Alpha = 12.80 1.1 -1.00 ^ Alpha a.00 3.60 0.50 " Alpha 6.3 0.40 0. L' = 1.50 ^ Alpha = 10.4: NACA 5312.9 '^'AAAA A . L' = 1. ^ = 0. L' = 1. A AAAAAAA.50 = o to • Alpha II 1 = o o 11 •» Alpha = • Alpha 2.00 Chordwise Location. pressure distributions for various incidences . L' = 2. L' = 1.1. x/c Figure 11.7 ««ij = 0.00 0.20 0.

L = 2.0 A A.0 l^^nna- • Alpha = 4.0 ^ • Alpha = -4 .8 1.59 2. 'llgggR.4 0.13 o o = Alpha 0.50 CO = A Alpha 12.94 V ' Alpha = 16. pressure distributions for var­ ious incidences . L' = 1. 3.0 I I I I ~l i O Minimum data for 16 degrees are left out 4.37 -2.0 0.2 0. L' = 1.0 0. L" = 1.5: NACA 5312.27 1<J g. panel code.04 = Alpha 6.0 Chordwise Location. isolated airfoil. n Alpha = 8. L = 0. x/c Figure 11. L' = 0.5. L = 1.6 0.!!!!! .

effect of the iiuinlxn* of panels on paii<!l code results . x/c Figure 11. 5 .8 1.1.5 +i v a 1.4 0.0 0.0 .5 1 •H 0) O 1 ^-^.0 L'=1.5 0.24 -1. 1.* CU A Experiment.0 -1.A U A • Code.664 0) U %* 3 CQ .0 Chordwise Location. 100 panels t 0. L'=1.695 Ci O MAA AAA •< 1 -l » ! '• f M r.2 0. 150 or 200 panels.6 0. L'=1. a = ^ = 0.0 Q) •H { o 0.6: NACA 5312.L to 1 A A A A t AAA* * 1 ' Code.

the test Reynolds number was 1. The highest possible Reynolds number was desired to most closely model the conditions a high-speed train would encounter. most aerody- namicists are familiar with the flow around an isolated airfoil and have a feel for it. leaving a "deadwater" region of slowly moving or circulating air toward the trailing edge of the wing's top surface.7. = is 0. This was the highest Reynolds number that could be achieved and maintained on a consistent basis. isolated airfoil data is useful for comparison with data obtained from airfoils in ground effect.8 x 10®. Second. Since panel methods are based on the assumption of . The most traditional method of depicting an airfoil's lifting performance is the lift curve. This means that the flow separates from the top surface. The reason is that inviscid theory represents a limit with the Reynolds number tending to infinity {Re —>• CXJ). 236 Presentation of Experimental Results The test procedure for obtaining experimental results was described in Chapter 10. The lift curve slope (c^q. After that the buffeting region is entered and a full stall occurs just after 14 degrees.e. a plot of lift coefficient versus incidence. Another benefit of a high Reynolds number is that a comparison with inviscid theory becomes more meaningful. There were several reasons to begin the experimental investigation with an iso­ lated NACA 5312 airfoil. First. i. The panel code results are also shown and consistently overpredict the lift coefficient.. data for the isolated airfoil can be used to assess the reliability of the wind tunnel and the data acquisistion equipment. The lift curve for the NACA 5312 airfoil is shown in Figure 11. The curve is linear up to twelve degrees of incidence. Third.11 per degree and the incidence at zero lift (a©) is —4. Unless otherwise noted. without the image wing nearby.31''.

0 degrees.02 produces considerably more lift for incidences between 1. The data from the inviscid theory are shown for reference. but this time results are included for wings in ground effect. and what effect that would have on the results? To attempt to answer these questions. and the discrepancy will be discussed below. and the wing in ground effect shows a less pronounced separation at higher incidences.5 degrees.9). the data look quite scattered and irregular. For the case of a trailing edge ground clearance ratio of 0. The reason is that flow velocities were very low.0 and 11. the variation of wing surface pressure distribu­ tions with Reynolds number was plotted (see Figure 11. Two questions in particular needed to be answered.8. The NACA 5312 airfoil was inclined at six degrees with respect to the oncoming flow. The wing at ground clearance 0. Maximum amplifier . would there be any indication of a transition from laminar to turbulent flow. the flow does not appear to separate by 16 degrees incidence.2 million). would the tendency of the data be toward the inviscid theory as the velocity was increased? Would the corre­ lation between the inviscid theory and the highest possible experimental Reynolds number be sufficiently close? Second. so that any disturbances such as wind gusts strongly impacted results. Lift coefficients are slightly higher between zero and 7. In addition. First. and the lift curve just continues in a straight line. 237 inviscid and attached flow. The lift curve becomes more nonlinear as the wing descends. Beginning with the lowest Reynolds number (Re = 0. The next question of interest was what effect the flow velocity had on the results. they cannot treat separated flow cases.1 the curve is close to that for the isolated wing. Both ground effect cases display a nonlinear behavior. The lift curve for the isolated airfoil is repeated in Figure 11.

. the flow encounters a thinner and more cambered effective body. A decrease in the boundary layer thickness leads to a thinner effective body. so that a thinner wing results in a somewhat lower lifting force. then the boundary layer would only be half as thick as before. 0.0 million show a trend toward higher lift coefficients. which is the sum of the physical body and its boundary layer thicknesses. A decrease in body thickness has a more complicated effect. and 1. Assuming laminar boundary layers and attached flow. The main reason appears to be the growth in magnitude of the pressures on the wing's top surface.5. velocities tend to fall very slightly as the body is thinned. Over the front half of the top surface there is a pronounced drop in velocities as the body is thinned. This statement is based on Prandtl's boundary layer theory. The reason is that thick boundary layers tend to fill in subtle variations in the body geometry.2. any increase in speed leads to a decrease of the boundary layer thickness. •238 settings had to be used to elevate voltages to useful levels. Both effects cause increases in the lift. velocities rise very slightly as the body is thinned. Nevertheless. On the underside. If one raises the Reynolds number. These observations are based on the results of the preliminary design of the present ground effect wing (see Chapter 8). the data for Reynolds numbers of 0. An increase in camber results in flow accelerations over the top surface and in flow decelerations near the underside. This pronounced drop dominates the lifting characteristics of the wing. which shows the laminar boundary layer displacement thick­ ness to be proportional to This implies that if the velocity (or the Reynolds number) were to be increased fourfold. Over the rear half of the top surface. in addition to increasing the volume of the effective body.

239

Returning to Figure 11.9, the undersurface velocities decrease slightly as the

speed is increased, resulting in more lift from the bottom. On the top surface,

velocities increase everywhere as the speed is increased, especially over the front

third of the wing. Both the top and the bottom surfaces contribute to a larger lift

coefficient as the speed is increased.

But how can one explain the drop in the lift coefficient and in the magnitude

of the top surface pressure distribution as the Reynold number is further increased

from 1.0 to 1.8 million? Perhaps the answer has to do with the nature of turbulent

flow as opposed to laminar flow. There is no clearcut Reynolds number at which the

transition from laminar to turbulent flow occurs. That transition depends on diverse

factors, such as the surface roughness and the level of turbulence in the freestream.

Nevertheless, transition for an airfoil is expected at Reynolds numbers anywhere

between 0.5 million and 3 million.

One difference between laminar and turbulent flows is that the turbulent bound­

ary layer is much thicker. Therefore the effective body tends to be thicker and less

cambered for turbulent flow than for laminar flow. As a body experiences transition

from laminar to turbulent flow one thus expects a drop in the lift coefficient.

If one would increase the Reynolds number beyond 1.8 million, the turbulent

boundary layer would become thinner. One turbulent boundary layer model claims
An
that the turbulent boundary layer thickness is proportional to Re and hence
A
to Vqo . This implies that the reduction of the boundary layer thickness with

increasing speed is quite subtle. The upshot of this is that the experimental results

for a Reynolds number of 1.8 million should not differ significantly from those at

higher numbers, or indeed from the inviscid theory.

240

Figure 11.9 shows a good match between experimental results at i?e = 1.8 x 10®

and inviscid theory on the top surface. However, the agreement between results for

the underside are poor. Since the effective body should be thicker and less cambered

than the physical airfoil, the expectation is that the experimental results show higher

velocities (lower pressures) there. This is indeed the case, but the magnitude of the

difference still seems disturbing. One further cause of the discrepancy is the Kutta

condition used for the inviscid calculations. The Kutta condition forces the tangential

components of the trailing edge velocity vectors from the top and bottom surfaces

to be equal. The tendency of the trailing edge velocity for panel methods is toward

stagnation {cp —> 1). This trend causes the bottom pressure distribution to be pulled

down towards stagnation, so perhaps this is part of the problem. The author tried

to remedy this situation by making the orientation of the Kutta panel adjustable,

but the results remained nearly unchanged. This may have been because the length

of the Kutta panel was the same as the panel nearest the trailing edge. Perhaps a

longer Kutta panel would have had a more profound impact on the location of the

front stagnation point, and hence on the circulation.

Following the analysis of the isolated NACA 5312 wing, the image wing was

added to the wind tunnel configuration to begin the study of the ground effect. The

idea behind using a wing and its mirror image is that the symmetry line formed

between them models the presence of the ground plane. To verify that the symmetry

really exists, nine pressure taps were located on the image wing. Figure 11.10 shows

how the readings from those nine taps match those from the main wing. The slight

discrepancy on the top surface near the leading edge may be due to a difference in

the incidences of the top and bottom wings. Alignment of the wings before each test

241

was accurate to within 0.2 degrees. The geometry of the model was carefully checked

before and during each test, and it was noted that the forces acting on the wings

typically rotated them from zero to 0.2 degrees, and deflected them apart from each

other from zero to 0.1 inches. There were some cases where the top wing rotated

by 1.6 degrees and deflected upward 0.3 inches. The case shown in Figure 11.10 was

one of the worst ones, and represents the design condition (6 degrees, 0.1 ground

clearance ratio). The top wing was observed to have twisted considerably (by about

1.1 degrees).

Some very crude smoke visualization was carried out to verify the presence of

the symmetry plane. Smoke entered the tunnel 14 inches upstream of the model.

The injector could be moved vertically. Figure 11.11 shows a typical frame from one

of these tests. The smoke does not penetrate the symmetry plane, thus verifying

the validity of the testing method. This visualization test was carried out at several

incidences and ground clearances. It was determined that at large incidences involving

flow separation, smoke did cross the symmetry axis if the wings were sufficiently

close. Also, whenever the trailing edge ground clearance ratio was smaller than

approximately 0.05, the flow was essentially choked and the smoke moved randomly

within the space between the two wings. Figure 11.12 shows how the ram-air cushion

prevents the smoke from entering the space between the wings. Since the air between

the wings is choked, the pressure there is at stagnation value, and symmetry is thus

effectively assured. The conclusion is that the symmetry assumption of the image

method is valid for all ground clearances, and for all incidences for which the flow is

attached.

With confidence in the experimental method, pressure distributions were recorded

for incidences ranging from minus four to sixteen degrees, and for trailing edge ground

clearances ranging from zero to one-half. Figures 11.13 through 11.19 each show pres­

sure distributions for different ground clearances at a given incidence.

Considering the design case (a = 6®) first (Figure 11.16), the effect of approach­

ing the ground is a decrease in the velocities over both the top surface and the bottom

surface. Since this deceleration is more pronounced on the bottom, the result is a net

lift increase. Indeed, the dramatic increase in lift for very small ground clearances

is due mainly to the deceleration and eventual stagnation of the air underneath the

wing. This is appropriately called the ram-effect, since the air is rammed under the

wing and forced to come to a halt. Results for the isolated wing are also included

for reference, and verify the trend of the data. From each pressure distribution, the

lift coefficient was computed. It was plotted against trailing edge ground clearance

in Figure 11.1 of the preceding section. The figure shows how the lift coefficient rises

as the wing approaches the ground. The exception to this trend is the isolated wing.

At large ground distances, the lift actually rises. To complete the study of the design

case, pressure distributions for various incidences are shown in Figure 11.20. Refer­

ring again to the pressure distribution in Figure 11.16, note the very high velocities

over the top surface and the comparatively high velocities along the bottom. The

high velocities at the top cause strong low pressures and generate much positive lift.

The higher velocities at the bottom cause a decrease in the positive bottom pres­

sures, resulting in lift losses. For the case of six degrees incidence the lift gains due

to the top surface flow outweigh the losses at the bottom. As a result the isolated

wing generates more lift than the same wing at a ground clearance of 0.4 and 0.2

chordlengths.

243

The above comments on the design case apply to Figures 11.15 through 11.18.

which pertain to positive incidences and attached flow. The case of separated flow

is somewhat different. Figure 11.19 shows pressure distributions for the wings at 16

degrees. There is a definite pressure plateau, or region of constant pressure, beginning

at the quarter-chord position on the upper surface. Such a plateau is typical for

separated flow, and corresponds to the deadwater region underneath the separated

flow. Air is slowly recirculating within this region, resulting in a very even and small

pressure coefficient. The underbody flow is as before, with the air decelerating as

the ground approaches. The flow behavior upstream of the separation point on the

top surface also remains as before, with the fluid decelerating as the wing descends.

But within the separated flow region, there seems to be no trend as the ground

distance decreases. Both the isolated case and the zero ground distance case show

higher velocities than the intermediate cases. The differences in pressure distributions

within the separation region are quite small. The trend of the lift coefficient is the

same as for the unseparated cases. As the wing descends, the flow slows everywhere,

and particularly on the bottom surface. This results in a net lift increase as the

ground is approached.

The wing in ground effect at 12 degrees also involves separation (see Figure

11.18), but only for ground clearance ratios below 0.4. The implication is that the

presence of the ground encourages flow separation. This is not surprising since the

ground causes the flow to slow everywhere, thus reducing its kinetic energy and

making it more vulnerable to flow disturbances.

Finally the cases of zero and negative four degrees incidence are discussed. Fig­

ures 11.14 and 11.13 show pressure distributions that differ considerably from the

244

cases with positive incidences. The case of minus four degrees is discussed since the

flow behavior is more obvious for that case (see Figure 11.13). The flow over the top

surface begins near the leading edge with a pressure coefficient of one. which indicates

that the stagnation point is located on the top surface. The top flow slowly acceler­

ates at it streams over the gently curving upper surface. Once it reaches the crest

of the body curve near the one-third chord streamwise location, it begins to slowly

decelerate until it reaches the trailing edge. The top surface flow tends to be faster

everywhere when the wing flies closer to the ground, but these velocity differences

are quite subtle. This implies that the top surface produces slightly less lift as the

wing descends. Turning now to the bottom side of the wing, observe the very fast

velocities near the leading edge, particularly for small ground clearances. But the

air quickly decelerates to near-freestream levels by the time it reaches the one-third

chord streamwise location. Beyond that the flow decelerates very gently all the way

down to the trailing edge. The trend of the underbody flow beyond the one-third

chord point of the wing is to accelerate as the wing descends. The exception to this

trend is the case of the smallest ground clearance, for which the flow is slower than

for the larger clearances up to the 60% chord position. Beyond that point, the flow is

faster than for the other cases. Judging from the shapes of the curves in Figure 11.13

it might be suspected that the powerful low pressure near the leading edge of the

wing bottom surface generates negative lift as the wing descends. In an overall sense,

it might be concluded that ground proximity implies less lift. The lift coefficients in

Figures 11.14 and 11.13 verify this, except for the case of closest approach.

Several additional experimental results involving the design case are presented

to complete the study. Figure 11.21 shows how the pressure coefficient distribution

245

changes as the Reynolds number is reduced from 1.8 to 0.2 million. Trends are

difficult to discern, and the results for the slowest case are very scattered.

The reliability of the data acquisition equipment was tested by conducting one

experiment with the additional help of a water manometer board. Figure 11.22

shows the water manometer data superimposed on the data from the data acquisition

system. The agreement is convincing.

To delve deeper into the comparison of results from the experiment and the panel

code, comparative data from three extreme cases were plotted in Figure 11.23.

In an effort to achieve a better match between experimental and computational

data for the design case, the lift coefficients were matched. Figure 11.24 shows the

experimental result [C^ = L' = 1.24) and the panel code result for the same lift

coefficient. The agreement is very poor. Panel code results are also shown for a

six degree incidence, which corresponds to_a theoretical lift coefficient of 1.664. The

agreement between this theoretical curve and the experimental data is reasonably

good on the top surface but again poor on the bottom surface. The author feels

that this figure shows that it may be better to match the incidences than the lift

coefficients.

Experimental Results

Panel Code Results

Oi

-5 0 5 10 15 20
Inclination, Alpha (degrees)

Figure 11.7: NACA 5312, isolated airfoil, lift curve


11 n
1.5 - •• • .. — — -— — — - -- --

e; !!


• Isolated Airfoil

h/c = 0.1

0.5 • h/c = 0.02
'

• '


A

-5 0 5 10 15 20
Inclination, Alpha (degrees)

Figure 11.8: NACA 5312, experimental results, lift curve with and without ground
effect

• Re =
1.8E+6
-p L' =
1.11
a
(D
•H •
o Re 1. O E + 6
•H =
L' 1.18
IH
(1)
O A Re 0 .5E+6
O
W
L' == 0.98
U
3
CO • Re 0 .2E+6
(0 =
(U L' 0 .87
M
CM
Inviscid Panel
Code, L' = 1.27

0.0 0.2 0.4 0.6 0.8 1.0
Chordwise Location, x/c

Figure 11.9: NACA 5:}r2, isolated airfoil, a = 6^, elTecl of Reynolds iunnl)er on
pressure distribution

1.5
fivr^
+» "j •
c 1 ••
(1)
•H i '

0 4
•H
(M 0.5 •.

«W
(U ••
O •
O 0 •
••
(1) • Main (top) wing data
u
0
m -0.5 • 11 '' Image (bottom) wing data tsi
CO • •••^ 1
Q) CO
U
^ -1
I

-1.5
0.2 0.4 0.6 0.8 1
Cho r d w i s e L o c a t i o n , x / c

Figure 11.10: NACA 5312, a = G'', ^ = 0.1, match of experimental data from top
and bottom wings

•250 ss^iiiilil^ii^SllS Figure 11.11: Smoke visualization image showing flow symmetry .

251 Figure 11.12: Smoke visualization showing stagnated underbody flow .

80 1.5 1 50 L' = -0.50 ft -0 L' = -0.00 • ••11 0) a\ t ^ h/c =0.20 0.50 \ 0. experimental pressure distributions for various ground clearances .117 • i M • 3 (0 ff D) 0 .13: NACA 5312.50 L' = -0.050 2 00 *• • h/c = 0. 0 .070 I -1 .00 Chordwise Location.60 0. 3 50 A 3 00 A » ® h/c = large 2 50 X ' L' = 0.40 0.2 H«•.1 to A .00 0.00 01 • h/c = 0. a = —4".036 A 1 . x/c Figure 11.00 -1 .

14: NACA 5312.1 CJ 0 .50 O • h/c =0.00 0. x / c Figure 11.05 m it L' = 0. a = 0®.50 i ft * A.39 0) r 0 0 H •haa' mm1 o o 3 0 5 " 1 5 ^p unoQQd • m • h/c =0.44 o 1 • (V 0 ^ h/c =0.40 0.5 o o 4J f* 2 L' = 0.60 0.2 •rl ((-i o L' = 0. experimental pressure distributions for various ground clearances .0 .50 • 0) o •H 0 1. ° flfl L' = 0. 5 0 P4 h/c =0.02 o o 1 -1 y L' = 0.34 <u .00 ChordwisG L o c a t i o n .80 1.20 0.1. 2 . 5 0 0.38 . 5 0 • h/c = 0.

. ..01 -1.J'" ' h/c = 0.50 • • ••.84 •H O j! ! ! _ • • •H A h/c =0.20 0.50 L' = 1.1 0.50 5 fln L' = 0.00 i.88 4J • h/c =0. experimental pressure distributions for various ground cU^arances .50 • h/c = large * • L' = 0.00 L. x/c L' =1.19 0) 0) o 4:- U f>—6-6-6-1 K>OOC>C-:.40 h/c =0.5 Ci (U 1.2 0.95 • 0) O f S 8' 0 • U • h/c =0.48 0. 1.00 • W • L' = 0. L' = 1.J .= 1.00 0. q = -l'*.60 0.02 -1.00 • ••• L' = 1.• ' ' .06 *** 0) • • • • < )a••••• U A A A A' i m 3 c.-i< h/c =0.80 1..40 0.h/c = 0.15: NACA 5312.52 Figure 11.0 Chordwise Location.05 CO -0.

40 0.00 Chordwise Location. L'=1.20 0.24 0) 0.50 (1) • h/c=large. L'=1. experimental pressure distributions for various ground clearances .60 0.O .4. L'=l. x/c Figure 11.10 o y y • ^ • h/G =0.ll •H • o 1.* • A h/G =0.00 -p c 1.57 -1.00 0.312.07 U-l W o 11 i ' ® 9 8W8 S . i S B .36 m • !•»! m XSaX *** f ! 1 « i j " " (1) 1 • • • • . L'=1.05. L'=1..0.•• h/G =0. L'=1.80 1.2. L'=1.00 ^ ^Qni • •H *CS .48 u ••• a • • • ' o P4 * ^VMvV«Vl|rl cc-o-o tij ioi iiSSSSSs .00 • ••• .02. L'=1. 2.01.16: NACA 5..50 • h/c=0.1. a = 6^.59 0.03.52 Ll h/c=0. e h/G =0. • • • • • • • • • © 5 P3 «••• *** h/G =0.h/G =0. L'=1.

23 ^ 0.40 0.1.50 ° 00 0) 58s D • h/c=0. 2.65 -1. L'=1.00 ri—i'. L'=1. experimental pressure distributions for various ground clearances .19 1..00 + .fnM'Vth'.0.2.50 . L'=1.47 cn • !X ' • ?c . L'=1. a = 8®.00 rie« ' 0) 0 A h/c=0.4.i'i > a • (1) • •H 1. L'=1.i rfiit'HV-t-r'ri- h/c=0.17: NACA 5312.50 ! • •.•••••" to m « • • • • m 1 llpn h/c=0.o O o V V 1 -1.20 0.h/c=0.02. • •O? ' ••• • • • • • .00 0.37 o •H (M 11 • • A • h/c=0.58 • r.01.80 1. x/c Figure 11.1 h/c=0.50 4J 2 .HcH'n'Vi'j-. g -0. L'=1.00 Chordwise Location. L'=1. L'=1.60 0.50 B • • • h/c=large.05. 0 0 T.63 0.35 r a g 3 0.

20 0.1.79 0 3.05. 5. • •• h/c=0. L'=1.00 0) •H • h/c=large. L'=1.40 0.00 Chordwise Location.00 U * • • • . L'=1.00 -p 1 a 4.00 -I o n 4 • A h/c=0.2.60 0.44 tw • 01 2. x/c Figure 11.77 0.02. L'=1.00 iilil i—l-B-l-l —n—Ln-^ i— h/G =0.01.50 o : ^ > 0) 1.68 -2.66 ifi* C) -1.80 1.00 • • u 3 So§ \• 1 *• *** • • U h/c=0.00 •H m s» • h/G =0.00 iJ h/c=0. h/c=0.54 lO (Q cP I! • B 8g y y • e. L'=1. L'=1. a = 12^^.66 0) Q) 0.00.4. L'=1.00 0. L'=1. experimental pressure distributions for various ground clearances .18: NACA 5312.

7.83 CO 1 _ 1.60 0.56 3 ii 1 ca 1. experimental pressure distributions for various ground clearances .20 0.05.00 • h/c=0.02.1.56 •H 4. L'=1.00 1 •p 6 .00 o • h/c=large. a = 16".01. U L'=1.19: NACA 5312. L'=1.80 1. L'=l. L'=1. L'=1.00 1 • h/c=0.78 m i 8 0) u 0.2. L'=1.00 1 i h/G =0.44 O 0) 2.00 V In:- Or 0) j ^ 5 h/c=0.40 0) 0 3. x/c Figure 11. 0 0 fi c: 0) •rl 5.a5 2.00 Chordwise Location.00 A h/c=0.00 • * * It I iu V} h/c=0.00.93 00 0.40 0.00 W i-mjlillLi h/c=0.3. L'=1.

56 -2. ^ = 0. x/c Figure 11.0 = • Alpha -4. L = 1.24 to = i A| A Li Alpha 8.0 Chordwise Location. L' = 1. experimental pressure distributions for various incidences .54 = Alpha 16. L' = -0.1.= 1.5. L' = 0.0 = • Alpha 4.0 0.06 = Alpha 6.35 • Ct Q * = A Alpha 12.2 0.0 0.20.07 Bottom Pressures = -• Alpha 0. NACA 5312.8 1.39 2.4 0. L' = 1. L.6 0. L = 1.

80 1. x/c Figure 11.2E+6 04 L' = 1. efTect of Reynolds nuinber on the pressure distribatioii .OE+6 1.20 0. /£.24 g 0.5E+6 0) (i L' = 1.50 zs o ! Li U u A Re L' = l.00 Chordwise Location. a — 6".66 0.27 t1 11 S -0-50 i ._.00 0.^2 ^ a U Re 0.60 0.40 0.21: NACA 5312.1. 2. an +J I C d) •H o -H 1.00 Panel Code -1.8E+6 1. n n ^0 D O M .50 L' = 1.34 .26 3 Re 0.00 ^' 1 • Re L' = 1.00 •a •n. ^ = 0. -1.

x/c Figure 11. comparison of scanivalve and manome­ ter results .50 • p"" Manometer Results • • ^• •• 1. • 0.1. NACA 5312.40 0.00 1.80 1. a = 6®.50 I • .22.50 0.00 " •• • Scanivalve Results 0.20 0.00 0.50 • • 1. ^ = 0.1.00 •• • • 0.60 0.00 Chordwise Location.

5 0 AAAA^^mi •••ncoiroj • Experiment.62 0 1. 0 deg.1.0 . •H iTi L'=0.24 P 0) 0 • CO A Panel code.00 L'=2.60 0.66 to • OS U to Pi .40 0. x / c Figure 11. 0) 0 L'=1.23.20 0. 0 deg. 5 0 m£i-I O A Experiment.80 1.39 U 2 •r| «H 2 Panel code. o o ' -1 L'=1.54 • . 6 deg. 12 deg.1. — ~ 0. 0. comparison of experimental and numerical results for three inclinations . 6 deg.27 ChordwisG L o c a t i o n . o ni o o o o (1) 1 • L'=1.00 0. 5 0 Ij Panel code. 12 deg. 5 0 a o o o o ni o (U 3 Q • • Experiment. • 0) L'=0. o o 4 • -p 3. NACA 5312.

05 degrees.6 0.0 Chordwise Location. 5 3.2 0.24 -1. a = 6^^. ^ = 0.5 N* -p a 1. 0) 6 degrees. matdiing experimental and numerical results T .8 1.24: NACA 5312.2 0.24 lO en (U 00 u (U Experiment. L'=1. x/c Figure 11.1. 3 (Q .0 0.664 O 0.5 -0. -1. O L'=1. 1.4 0.0 6 degrees.0 Q) U Panel code.0 01 » •H 'O o t * •H 0.0 . m L'=1.5 •W Panel code.

the pressure acts normal to the surface. It is due to the streamwise component of the surface pressure distribution. the other component of drag is the skin friction drag. This implies that ground proximity has little effect on the pitch stability of the wing. Figure 11.27 shows how the pitching moment becomes more negative as the wing descends. but also the leading edge pitching moment coefficients. It is due to the friction between the viscous air and the uneven wing surface. This means that the wing has a nose-down tendency as it approaches the ground. Nearness of the ground does not appear to significantly impact the variation of pitching moment with incidence. The same cannot be said for the variation of the pitching moment with ground clearance. and the absence of three-dimensionality avoids both the lift-induced drag and the interference drag due to the flow near the junction of the wing and the fuselages. The two sets of data are quite similar and show no obvious trends. c^p. The pressure drag is only one component of the total drag acting on the wing. since both the slope ( and the y-intersect ( c ^ q ) curve are the two crucial stability criteria. The pitching moment is important in that it provides information on the pitch stability of the train. Results are shown for an isolated wing and for a wing in ground effect. 264 Experimental pitching moment and pressure drag results The experimental pressure distributions were not only used to calculate the lift coefficients. The reason is that as the wing descends. At any given point. Figure 11. Subsonic speeds avoid the wave drag associated with shock waves. and the pressure drag coefficients. alternatively. which means that the pressure force is inclined with respect to the freestream. the center of pressure is moving aft. the region of high- . For a wing profile at subsonic velocities.26 shows the variation of the leading edge moment coefficient with the incidence.

and 11. The aerodynamic center is the location at which the pitching moment does not vary with incidence ( = 0.3 -0. 11. For static longitudinal stability.33 -4. It can be found since it equals the moment (c^q) at any point when the lift is zero. Figure 11.26. The conclusion is that the ground has a stabilizing effect since it moves the aerodynamic center aft. Turning next to the pressure drag. the aerodynamic center must lie behind the center of gravity.8. the aerody­ namic center needs to be found.08 h/c = 0.1. To study the effect of the ground on pitch stability in more detail. causing more lift toward the rear.35 -3.24 0. a = 6". (11-1) These relations were used in conjunction with data from Figures 11.09 0.1: NACA 5312.25 one can write the following relationships: "^mLE = ^CPH ^mO = ^acc(^.18 0. or cm = constant). 265 Table 11.12 pressure air beneath it expands toward the trailing edge. That Table reveals that the aerodynamic center [xac] moves aft as the wings approaches the ground.438 1.1 -0.391 1. It remains to be seen whether it actually moves aft of the center of gravity to provide stability.27 to produce Table 11. From Figure 11.28 depicts the variation of the . experimental data Experiment ^mLE xcp OQ ^mO Xac isolated profile -0.0 -0. or whether a horizontal tail or second (tandem) wing will have to be added to the design.15 0.

the more the wing pitches.25: Three equivalent loadings pressure drag coefficient with incidence. The overall trend is that the ground effect reduces the lift-to-drag . Such a plot of the (pressure) drag coefficient versus the lift coefficient is shown in Figure 11. the pressure drag seems to rise as the wing approaches the ground. Once again there are no obvious trends as the wing descends toward the ground. Small positive incidences display both irregularities and the highest lift-to-drag ratios. The irregularities may be due to the division by the very small pressure drag coefficients for those cases. When the pressure drag coefficient is plotted against trailing edge ground clearance. the higher the pressure drag is.30 and looks quite normal.31 the lift-to-pressure drag ratio is plotted against the wing incidence. 266 Figure 11. The lift-to-drag ratio can be used to assess the efficiency of a wing. Figure 11.29 results. In Figure 11. The wing shows nearly the same behavior whether or not it is near the ground. Though the data are quite scattered. One traditional depiction of the drag characteristics of an airfoil is the drag polar.

A likely explanation is that the strong upward pressure on the bottom surface of the wing dominates the pressure drag. which is turn inclined at six degrees with respect to the freestream. These comments are verified in Figure 11. 267 ratio of a given airfoil. That ram-pressure is directed nearly normal to the chordline.32. This simply means that even though the wing in ground effect produces more lift than the isolated wing. Therefore the pressure vectors all have about a ten-percent downstream component. which shows how the efficiency (lift-to-pressure drag ratio) varies with ground clearance. Even though the data are again quite scattered. resulting in a rise in the pressure drag. that gain is accompanied by a drag penalty that reduces its overall efficiency. they do reveal a near- 50% decrease in efficiency as the wing approaches the ground. .

Alpha (degrees) Figure 11. h/c = 0.6 • Inclination.3 • l -o„ • • A 1 With Ground Effect.26: NACA 5312. experimental leading edge j)it<liing moment versus inci­ dence . -5 10 15 20 c -0.1 to 05 OO a -0-5 11 • A -0.1 -A - (V •rl • O -0-2 iM 0) • • No Ground Effect o -0.

4 a C75 CO 1 S -0.27: NACA 5312. 1 a (D d -0.6 -0.7 Trailing Edge Ground Clearance.2 •H •h _ _ Q) -0. a = 6".3 o 0 -0. experimental leading edge pitching moment versus ground clearance .001 0. 0. h/c Figure 11.01 0.1 10 0 •P .0 .

12 O No Ground Effect 0.1 Q 0. 0.14 0) O 0.2 4J a 0.06 M to 3 —J 0) 0) 0.16 o •H 0. h/c = 0.08 0) 0. Alpha (degrees) Figure 11. experimental pressure drag curve .1 (U With Ground Effect.18 d) •H 0.28: NACA 5312.04 o 0) M 0.02 f:ii 0 -5 0 5 10 15 20 Inclination.

experiiiKMilal resulls .01 0.025 en 0.02 n) U Q 0. a = 6^.015 0.1 1 10 Trailing Edge Ground Clearance.03 0. NACA 5312.001 0.29.01 2 0.04 0. 0.005 0.035 0. h/c Figure 11.

h/c = 0.06 lO lO S 0.04 Q) M 0.5 2 Lift Coefficient Figure 11.5 0 0.18 Separated Flow 0) Ej 0.14 (1) o 0.2 C 0.08 g 0.30: NACA 5312.5 1 1.1 Q 0.12 • No Ground Effect 0> 0.02 04 0 -0. experimental drag polar .16 s 0. 0.1 (d • With Ground Effect.

. • • • • • No Ground Effect 11 .1 A •_ ic A i "S •— A •5 0 5 10 15 20 Inclination.31: NACA 5312. h/c =0.With Ground Effect. Alpha (degrees) Figure 11. • • . experimental efficiency results ..

01 0. 100 o •rl •P 90 & 80 (0 70 O (1) 60 U 0 50 m m Q) 40 u to PI 30 0 +) 20 1 +J 10 •H 0 0. a = 6^. experimental efficiency results .1 10 Trailing Edge Ground Clearance.32: NACA 5312. h/c Figure 11.001 0.

the lift remains nearly the same as for the isolated wing. so the focus will be on the case of six degrees incidence. for small ground clearances the deceleration is particularly pronounced beneath the wing.33. their ratio was plotted against ground clearance ratio in Figure 11. To illustrate the growing importance of the lift from the bottom surface compared to that from the top surface. the velocities all around the same wing in ground effect are lower. By contrast. As a result. . Discussion of the Ground Effect Phenomenon One of the primary aims of this dissertation is to shed more light on the ground effect phenomenon. For larger ground clearances the deceleration is small and affects both the top and the bottom surfaces to a similar extent. The air then moves with the wing. Compared to the isolated wing at six degrees. resulting in a strong lift increase. The strong deceleration under the wing is called ram-effect because the wing acts like a scoop and rams the air beneath it. some general comments can be made about the flow phenomenon in two dimensions. resulting in high pressures and the associated ram lift. The focus will further be on the ten percent trailing edge ground clearance ratio. Based on the preceding discussions of computational and exper­ imental findings. since that is the design condition for the proposed aerodynamic levitation train. Most wings in ground effect operate at positive incidences.

1 1 10 Trailing Edge Ground Clearance. h/c Figure 11. lift contributions from the top and bottom wing surfaces . a = 6®. lO a> 0.001 0.33: NACA 5312.01 0.

there might be a boundary layer near the ground plane. Even the crude smoke flow visualization showed that there are no recirculating bub­ bles beneath the front of the wing. the author does not feel that the ground effect has been fully modeled. The reason is that for the physical ground effect the ground moves past the wing with the same speed as the freestream. In general. several changes would be made. Its most important function for the author was to clarify that the wing underbody flow is similar to a one-dimensional nozzle. The effect of that boundary layer is probably not significant. If the author were to repeat the construction and testing of this wind tunnel model. . For the physical ground effect. lighting units would be embedded in the wing surfaces that face each other to facilitate smoke flow visualization between the wings. so its consideration would be an intellectual exercise. Last. A tilting mechanism would be developed for easier adjustment and strict maintenance of the incidence of both wings. The second most impor­ tant function of the wind tunnel tests for the author was that they furnished pressure distributions for comparison to the computational results. thus reducing the rotational and translational displacements of the wings during testing. For purposes of designing a train that utilizes the ground effect the present experimental method probably suffices. The present image method allows for the "ground" to move at speeds different from the freestream. The wing raising mechanism and the two slider blocks would be reinforced to further eliminate play. 277 Suggested Improvements for Experimental Work The present wind tunnel method has fulfilled its goal of deepening the under­ standing of the ground effect phenomenon. The only time fluid comes to rest or reverses direction is when the wing nearly touches the ground plane.

The ground boundary layer was not considered at all in the panel method used in this dissertation. Even though inclusion of viscous effects makes the solution much more complex. An Euler method on an appropriate grid might be useful for this purpose. They show the proper trends. Not only the orientation. The present panel method could be improved by reintroducing the adjustable Kutta panel. but also for the moment data. several steps could be taken. It would also enforce a physical tangency condition on the ground plane. the new method could include a model for the turbulent boundary layers on the airfoil and on the ground. First. 278 Suggested Improvements for Computational Work For the purposes of understanding the ground effect phenomenon and for design­ ing an optimal wing to utilize it. allowing for a stability analysis. Changing the geometry of the Kutta panel should result in a shift of the front stagnation point. panel methods suffice. it does allow enforcement of no-slip and tangency conditions on all surfaces. Rather than solving the entire viscous flow field. Second. even if the magnitudes are erroneous. The absense of a boundary layer model accounted for at least part of the discrepancy between numerical and experimental results. In this way the boundary layer displacement thickness would be added to the physical body to arrive at the realistic effective body shape. it would be possible to couple the inviscid solver to a boundary layer model near the physical surfaces. This would make an optimization . This is not just true for the lift. but also the length of the panel should be adjustable. the new method should yield results not just on the airfoil surface but also in the flow field. The panel code could then be "tuned" to the experimental results. If a thorough analysis of the two-dimensional ground effect is desired.

so at the trailing edge = 0. so any experimental results should be applicable to the entire flight envelope. it is useful to know the displacement thickness of the boundary layers.011.64 feet (60 cm) above the ground at 200 mph (293 fps. It also represents about ten percent of the profile thickness.1. Assuming a wing with a chordlength of 16. The flow will also be turbulent during takeoff and landing. This represents just more than ten percent of the nondimensionalized trailing edge ground clearance of 0.W/^^)(293^/^)(16-4/0 ^ ^^ ^i ^) 3. the new method could treat the entire domain a^ a viscous flow. Its thickness is calculated according to the procedure given by Anderson [311]: 0. The conclusion is that it .37)(6 m)(1. 279 of the wing more meaningful. The turbulent boundary layer will be thickest at the trailing edge and will threaten to choke the flow there. (11-3) The boundary layer thickness is nondimensionalized with respect to the chord length (c = 6 meters).72 X slug/ft-s The Reynolds number is most certainly in the turbulent flow regime. two-dimensional Navier-Stokes equations.225 kg/m^){l02 m/s){Q = 0.37x(^)0-2 ~~ (o. Third.7894 x lO-^kg/m-s) {pvoox)^'^ ((1.4 ft (6 meters) flies 1. This implies a numerical solution of the steady. To make an assessment of whether one should really invest time and effort into the writing of such complex codes.2 0> ~ 0. or 320 kph). Re = = (0-002377.0664 meters. the Reynolds number is:.

Inclusion of the effect of the boundary layers makes a profile shape optimization meaningful. A detailed optimization will need to incorporate a turbulent boundary layer model. and minimize the pressure drag. The thickness of the boundary layer on the ground was calculated in the same manner as the boundary layer on the profile. 280 is significant in the context of selecting a profile that will achieve maximum lift in ground effect. .016 by the time it reaches the trailing edge.Further assuming the ground boundary layer to start under the leading edge. the ground boundary layer has a nondimensional thickness of 0. provide stability. Based on the experimental results. On the other hand. the average underbody flow velocity at the design condition is assumed to be about 75% of the freestream velocity. the relative velocity between the ground and the underbody flow is O. Since the ground moves at the freestream velocity. The negative leading edge pressure peak can be molded to limit the pressure drag while still avoiding separation. so it will not choke the flow. the turbulent boundary layer is not thick enough to choke the flow underneath the wing. Its thickness is of the same order as that of the profile boundary layer.Suoo. The profile can then be carefully designed to maximize lift.

the bow lifted out of the water and the boat nearly flipped over backwards. He envisioned "Riverboats" that would carry medication. aesthetic. to build a fast speedboat that could be used to test avionics equipment. 281 CHAPTER 12. so let it fly". financial and safety considerations. In the early 1960s. and some with water propellers. and people along rivers. mail. He began work on vehicles that flew just above the water surface. Lippisch was asked by his employer. human factor. taking into account constraints that are due to environmental. Origins of the Ram-Air Train As mentioned in the introduction (Chapter 1). When Collins and Lippisch tried out one of those speedboats. Arthur Collins. When the civilian sector showed no . the author became interested in the aerodynamic ground effect while working on the Alexander Lippisch Collection at the Iowa State University Library. some powered with air propellers. A RAM-AIR TRAIN PROTOTYPE: FEASIBILITY STUDY The foci of the present chapter are the justification of the concept of the ram-air train and the establishment of the basic dimensions of the ram-air train. At that point Lippisch decided that "the boat wanted to fly. He also envisioned large ocean-going aerofoil-boats that would compete with ocean liners and jetliners.

1990). The author thus arrived at the wrap-around design with twin fuselages. In addition. but it was considered important for the passengers to have good visibility. The aircraft was to have an engine mount and propeller above its fuselage. or "AeroLev Train". as discussed in the next section. Hovering vfa. transports. That project was the aerodynamically levitated mass transportation system called "Ram-Air Train".s to be achieved by using a system of flaps with or without a rearward tilting engine mount. lifting fans were to be embedded in the wings to provide the hovering ability. 282 interest in his proposals. The Concept of the Ram-Air Train Once again. and submarine hunters. The earliest versions of the train were envisioned as operating in a trough. which was considered a safety hazard. They could thus derail vertically. reconnaissance vehicles. the airstream behind the propeller was to be captured beneath the vehicle (see Figure 12. In another design. which means that it achieves levitation above a flat track . the proposed concept is a self-propelled train that utilizes the aero­ dynamic ground effect. The author first sought to extend the idea of the "Riverboat" by adding a hov­ ering capability to it. In this way. aerodynamic interference between the wing and the fuselages would reduce the efficiency of the wing. he turned to the military and proposed using aerofoil-boats as patrol boats. from August. Lastly. The versions that followed operated on an inverted U-shaped guideway and did not wrap completely around the guideway (see Figure 12). it would be difficult for passengers to access the fuselages. But the applications of such a vehicle to military and criminal activities convinced the author to pursue a different project beginning in January of 1992.

283 Figure 12.1: Hovering aerofoil-boat .

2: Early ram-air train . Figure 12. I/?/32.

the author built the wooden model in 1992). one might interpret the advantage of the shroud as a reduction in propeller diameter. As a consequence. The ducted propeller was chosen for several reasons.. The surfaces connecting the wings and the fuselages serve to prevent high-pressure air leakage around the edges of the flange. The train hugs the horizontal flange. The train's wings are flown in the ground effect of the flange. there is no rolling friction since the train is flying. off-the-shelf reciprocating engines can be used to drive the propellers. . 285 without the use of lifting fans or magnetic fields.e. thereby making the flow over the wings nearly two-dimensional. preventing people and objects from entering the propeller's path from the sides. It also serves as a safety shield. The guideway has a T-shaped cross-section consisting of a vertical web and a horizontal flange. i. The prominent features of the train are the twin fuselages that are suspended from the single or tandem wing and the ducted propellers that are attached to the stern of each fuselage (see Figures 1 and 12. Further fuel savings should be due to the fact that there are no moving parts. The shroud has the effect of reducing noise and of increasing the thrust in comparison with an unshrouded propeller of the same diameter. only a small fraction of its power is used for levitation resulting in a high fuel economy in compar­ ison to trains using magnetic or air-cushion levitation. Conversely. Furthermore.such collisions could be due to windgusts and curves. It has been shown that both the friction due to sliding and rolling members and the roughness of the ride become serious problems for high-speed trains operating at speeds above 200 km/h. making derailment virtually impossible. They may also create enough air cushion between themselves and the edges of the flange to prevent lateral collisions with the guideway .

and hopefully at a reasonable price. There is a speed limitation with ducted propeller propulsion systems. Obstacles on the guideway will be quite dangerous. 286 Figure 12. and require only an existing level of expertise. Perhaps hydrogen engines will be available soon.3: Wooden model of the AeroLev train This is a great advantage especially in poorer and remote regions of the world because the engines do not cost much in terms of original purchase and maintenance. There are a number of drawbacks associated with the ram-air train. especially when visibility is restricted by . both in terms of noise and pollution. The speed of the propeller tips may not approach the speed of sound because shock waves would form that would greatly reduce the efficiency of the propeller and cause high loads on the blades. Conventional reciprocating engines are also becoming increasingly environmentally friendly.

how can one connect two sets of guideways? One possibility is to use the switch designed at Birmingham in the United Kingdom. trains. The train's stability must be such that collisions with the guideway will be avoided as much as possible. thus competing with cars. Prototype design: sizing. which is a large circular concrete plate with two guideway segments on it . Otherwise. Expected performance and possible uses of the ram-air train As pointed out by Lippisch there is a great gap in efficient vehicles in the 160 to 400 km/h (100 to 280 mph) speed range. the S-shaped segment connects the guideway on which the train is approaching to the track parallel to the in-line track. 287 weather or at night. Another is that the guideway should blend smoothly into the existing infraistructure.one of these segements is straight. it will be quite difficult to compete with today's highly efficient jetliners on routes exceeding 500 statute miles. buses. One is the high cost involved in building a continuous solid structure. without requiring the reloca­ . Of great concern is the stability of the vehicle. mission The width of the train's guideway is limited by several factors. If the train is to continue on a straight line. the other has the shape of a S-curve. and aircraft. both from the standpoint of passenger comfort and ride quality and of pitch and heave stability of the vehicle. The ram-air train should be useful for distances between twenty and 500 statute miles. Another disadvantage is the difficulty of switching . In terms of range. the disk is rotated to that the straight guideway segments lines up the train with the in-line track.

The wing planform is rectangular to make its construction as simple and economical as possible. the prototype's wing is assumed to have that profile. The width of a typical single-lane. The design speed of the train should lie above that of automobiles and conven­ tional trains. two-way city street was used as a gauge for an aesthetically and dimensionally acceptable size. Thus the wing span was fixed at six meters (19. It is positive enough to result in sufficient lift and small enough to definitely avoid separation.7 ft) as well. The choice of incidence is also of theoretical interest since it is a case to which nonlinear channel theory would have to be applied. allowing both for obstacles on the guideway and for heaving motions of the train. The design incidence of six degrees was chosen for several reasons. It also represents a very efficient incidence in that it corresponds to the maximum lift-to-drag ratio. For this small prototype the chord length of the wing was expected to be similar to its span. it should not rise far above the compressibility limit. In the interest of safety and of minimizing noise and drag. so it was set at six meters (19. Therefore the design speed was chosen to be 200 mph (320 kph). which is at a Mach number of 0..24 during the .7 ft). Given the six-meter chord length the trailing edge would be 60 centimeters (2 ft) above the guideway.1 was also chosen for several reasons. above 100 mph (160 kph). Such a large clearance is desirable for safety reasons.3 and corresponds to about 230 mph (370 kph).e. At the design inci­ dence and ground clearance. The maximum speed of the train would be around 250 mph (400 kph). 288 tion of buildings and other structures. that profile generated a lift coefficient of 1. Based on the preceding numerical and experimental studies of the NACA 5312 airfoil. The design trailing edge ground clearance ratio of 0. i.

•289 experiment.342i.8 mph. Returning once again to the Beechcraft King Air. The fuel weight corresponds to about 120 U.34 m/s (145 kph. resulting in four hours of operation. the ram-air train . The King Air has a takeoff weight (gross weight) of nearly 10. particularly when compared to the takeoff speed of a small sports aircraft (70 mph for a Cessna 172 Skyhawk).700 pounds for twelve people.24)i(1. then the lift required to lift the train must equal 10.two engines of this size will consume about 30 gallons per hour at sea level.482 Newtons. Given the wingspan and chord length the prototype is expected to carry approx­ imately 12 people. This is a very reasonable takeoff speed. gallons of fuel . crew. If one assumes the same engine configuration for the ram-air train one arrives at a fixed weight (passengers.000 pounds. making the train useful on shorter lines. and weight are assumed to be similar to a twin-engine turboprop commuter aircraft.1) ^ s" = 27. From the definition of the lift coefiicient.225-^)u^(^^)(6m)(6m) (12. construction.500 pounds (based on the King Air's all-metal construction and engines) and a fuel weight of 800 pounds.S. From Equation 12. including the crew.000 pounds. such as the Beechcraft King Air (see [329]). or 132 fps). T 1 2 c = (1.1 the takeoff velocity is 40. The takeoff roll will therefore also be reasonably short. an empty weight of 6. If one assumes the same gross weight for the ram-air train. This translates into a range of about 800 kilometers (497 statute miles) without refueling. that aircraft is powered by two engines weighing 700 pounds each and rated at 550 horsepower each. As desired. food) of 2. or 44. cargo. Its size. 90.^ Newtons.

290 could thus compete with other transportation systems on all routes with a length of up to 500 miles. by corrosion and by . This allows for light-weight construction and thus for very low inertia. communication and natural gas to people living near the track. thereby providing connections to water. branches and rocks. Drawbacks of the elevated guideway include that it is unsightly and that it is subject to damage by storms. snow and ice. Another advantage of a concrete guideway in comparison with conventional and high-speed railroad tracks is that track imperfections and misalignments are quite tolerable since the train is not in contact with the surface. The great advantage of the aerodynamic cushion is that it features no concen­ trated loads. Since the guideway will be made of concrete or of a similar material it should be relatively easy to clear the surfaces of dirt. The guideway can also serve for other purposes . An elevated guideway greatly reduces the possibility of obstructions in the form of humans. water. by the weight of snow and ice. The absense of direct contact also decreases track vibrations and the associated noise and fatigue. The distributed loading prevents much wear and tear and the associated maintenance costs. an­ imals. one may expect excellent acceleration and breaking performance.perhaps one could embed solar panels or solar cells in its surfaces to generate hot water or even electricity. One could embed easily accessible cables and pipes. sewage. electricity. The elevated guideway The ram-air train is expected to operate at cruising speeds between 150 and 400 km/h (93 and 250 mph) so that obstacles must be avoided at all cost. but rather presents a mild distributed load on both the track and the vehicle.

certainly for roads and railroad tracks. The conclusion is that the ram-air train is a feasible transportation system provided its flying height and pitch attitude can be controlled in a comfortable manner.such impacts may be caused by windgusts and by curves. The next steps in the analysis of the train would involve both the pitch and height stability. 291 earthquakes. or for lower speeds during takeoff and landing. But this is true for most civil engineering structures . and provided lateral collisions with the guideway can be avoided or at least dampened. braking and landing The train will have conventional rubber tires on the surfaces facing the guideway. A second wing might be added to increase the . Furthermore. At cruising speeds the wheels serve to protect the vehicle from impacts with the guideway . especially in terms of maintenance. the train achieves a range of 500 miles. Braking at high speeds will be achieved by changing the pitch of the propellers. Takeoff. Trailing edge flaps may need to be used to control the flying height of the train. including in the case of an engine failure. As long as the tires are in contact with the guideway they can also be used for conventional braking. These will suspend and guide the vehicle at low speeds. Guideways are generally quite expensive to build but they may be cheap in comparison to high-speed rail stock. The flaps would also allow for a higher takeoff weight. The Future of the Ram-Air Train The preceding analysis has shown that the prototype's wing generates sufficient lift to support the vehicle at speeds above 90 mph.

and noise. The human factors involve the ride comfort. so the wing needs to be analyzed to reveal accelerations in the vertical. It should be determined whether the train has sufficient lift capacity and stability. 292 pitch stability while at the same time increasing the lift capacity of the train. perhaps by means of a panel method. The possibility of lateral collisions with the guideway should be investigated. forward. the full three-dimensional train and its guideway should be modeled. . vibrations. After that. A preliminary numerical design should result in the construction and testing of a physical model and its model track. The pitch and heave characteristics of the wing(s) should be studied with regard to safety and with regard to human factors. and lateral directions. The optimization function might be to maximize the lift while flying at the design incidence and height above the guideway. The airfoil shape could be further optimized using the present panel method.

It was found that the panel method correctly predicted the trends of the flow charac­ teristics as the wing descended toward the ground. The first were historical surveys of the application and theoretical and experimental study of the ground effect. Unconventional trains were also discussed. panel method results were not useful for quantitative purposes. including magnetic levitation trains. Numerical and Experimental results were presented and compared in Chapter 11. 293 CHAPTER 13. and of the development of unconventional trains. The method could be improved by reintroducing . The two-dimensional panel method used for the preliminary design of an airfoil in ground effect was presented in Chapters 7 and 8. The third was to study the feasibility of an aerodynamic levitation train based on the lift data from the wind tunnel tests. This was followed by a descrip­ tion of the construction and testing of the experimental model in Chapters 9 and 10. However. CONCLUSIONS AND FUTURE DIRECTIONS The aims of this dissertation were threefold. The first six chapters provided a thorough review of ground effect research and of vehicles that utilize it. The panel method overestimated the lift mainly because the velocities all along the bottom wing surface were underpredicted compared to experimental results. The second aim was to design and test a wind tunnel wing in order to better explain the aerodynamic ground effect.

294 a Kutta panel whose orientation and length can be adjusted. This conclusion suggested a wide array of future research in this area. A further reason for the differences between numerical and experimental results was that the top and bottom walls of the wind tunnel test section may have had a significant influence on the flow around the two airfoils. Third. the panel method could be executed in conjunction with a turbulent boundary layer model near all physical surfaces. Second. The author does not consider the above studies to be very crucial in regard to the proposed transportation system. The conclusion of the prelimi­ nary design was that it would be possible to utilize the ground effect to aerodynami- cally suspend a small prototype operating at 160-400 kph. The preliminary design of the aerodynamic levitation train in Chapter 12 was based on the lift data obtained from the experiment. the two-dimensional Navier-Stokes equations could be solved on an appropriate grid. the moving-model approach should be used to properly model the ground plane and its boundary layer. . If there is further interest in the discrepancy between numerical and experimental results. several numerical methods might be tried. One way of doing this would be to build the test rig that is pushed by an automobile. It could not account for the fact that both the ground and the freestream move at the same velocity with respect to the wing. Should there be interest in further illuminating the ground effect flow phe­ nomenon. the wind tunnel top and bottom walls could be modeled using distributions of sources and sinks. A further criticism of the results were that the image method used for both the numerical and the experimental study did not completely model the ground effect. The author would suggest the following sequence for future research. First.

A second wing or a horizontal stabilizer should be added to guarantee pitch stability. A three-dimensional panel method might be used to design an optimally proportioned train. 295 In terms of continuing the design of the aerodynamic levitation train. Once the wing and its stability are fully understood. Such a model could be remotely controlled on a closed track. A model of that train should be built and tested on a corresponding guideway. the next step should be to study its pitch and heave stability. for example by using a panel method. The data from this dissertation should be used to establish the stability of a wing in ground effect. . or catapulted along a short segment of straight track. Even though proximity to the ground is a stabilizing factor. It may also be worthwhile to optimize the profile shape of the single or tandem wing to achieve maximum lift while maintaining pitch stability. The tandem configuration should be analyzed numerically. the full three-dimensional body of the train and the guideway should be investigated. a single wing near the ground is still unstable because the center of gravity lies aft of the aerodynamic center.

a system of passenger trains operated by the National . Abbreviated as AGE.a train developed in the United Kingdom that tilts as it passenger train negotiates curves. aerodynamic .a raceboat that has wings to lift it just above the water surface.tracked air cushion vehicle with lifting fans. aerodynamic lift . Abbreviated as APT.a flying boat that has several modes of operation. as planing boat.lift due to pressurized static air (lifting fans). as ram-wing. It can act as displacement catamaran or trimaran. The wing is called the active wing and its image is the image wing.the effect of the ground upon a vehicle operating in proximity ground effect of the ground. •296 GLOSSARY OF TERMS USED IN THE DISSERTATION active wing . The plane of symmetry between the wing and its image then models the ground plane. Aerotrain . Amtrak . or as free-flying airplane. as wing-in-ground-effect. aeroskimmer .lift due to pressures created by the forward motion of a wing through the air. aerofoil boat . advanced .instead of testing a wing flying close to a surface one can test the wing and its mirror-image. various vehicles were tested in France between 1965 and 1975. aerostatic lift .

a measure of a wing's slenderness (when viewed from above or below). chord . dynamic air .a measure of the curvature of a wing profile. aspect ratio . electrodynamic . drag the aerodynamic retarding force on a body. 297 Railroad Passenger Corporation on leased freight lines in the United States.when a ram-air cushion is created between a wing and the cushion underlying ground due to the forward motion (dynamic) of that wing. EDS vehicles utilize on-board superconducting magnets.EMS. it is due to the streamwise component of the pressure forces (pressure drag). camber .angle between the velocity vector and the chordline of attack a wing profile. electromagnetic . a descending or pitching aircraft near the runway).the time-dependent effect of the ground upon a vehicle effect moving in its proximity (e.EDS.g. a noncontact suspension system in which the suspension vehicle rides on an electromagnetic cushion created by . to friction with the air (skin friction drag). but with larger ground clearance (typically 0.the distance between the leading and trailing edges of a wing. angle of . boundary . For a wing profile. and require a certain forward motion before the repulsive force becomes effective. a noncontact suspension similar to magnetic suspension levitation. dynamic ground .a thin layer of fluid near any physical boundary in which layer the flow velocity varies between zero at the boundary to a value near the freestream velocity. and to flow separation.25 meters for the full-scale case) and with a smooth flat undersurface.

downwind. leeward .towards the side away from the wind.altitude stability.same as a hovercraft. The craft is propelled by propellers.vehicle that rides on a cushion of air pressurized by fans.same as the angle of attack. 298 conventional electromagnets. image wing . downstream. interface vehicle . . linear induction .a three-dimensional wing with wingtips at the ends ( as opposed to an infinite wing that is assumed to have an infinite span). magnetic . hovercraft . heave stability .see electromagnetic suspension. distal. levitation nonplanar .instead of testing a wing flying close to a surface one can test the wing and its mirror-image. finite wing . the tendency of a wing to return to its initial altitude following a perturbation. The cushion is sealed off by rubber skirts under the edges of the vehicle.when a vehicle operates close to a wavy terrain. hydroskimmer . The train must wrap around an elevated track with extremely small clearances (typically one centimeter for the full-scale case). incidence . This ground effect includes flight over waves and rough terrain.a vehicle that operates close to a solid or liquid surface. The plane of symmetry between the active and image wings then models the ground plane.a non-contacting propulsion system utilizing motor electromagnetic forces between the vehicle and the guideway.

span .a wing moving just above a solid or liquid surface. static ground . A subset of TLV which . TACRV .a theoretical point that continuously produces/ejects mass. the distance between the two wingtips. tilt-train . 299 nonplanar wing .the time-independent effect of the ground upon a vehicle effect moving in its proximity at constant speed and attitude (this includes hovering craft). singularity . sources.a phenomenon where flow can no longer follow the curvature of the body and breaks away from the surface. doublets. TACV Tracked Air Cushion Vehicle. rather than on the actual surface. Same as the TLRV. leaving behind a deadwater region.Tracked Air Cushion Research Vehicle.g.wingtip.a theoretical point at which some property tends to either zero or infinity (e.conventional train which tilts as it negotiates turns. sinks. tip .a wing whose silhouette is not a rectangle when viewed from the front (upstream). the spanwise end of a wing.a theoretical point that continuously absorbs mass..the width of a wing. separation . ram-wing . Built for the US DOT by Grumman Aerospace between 1971 and 1974.a ram-wing moving over a solid ground surface. and vortices). thin airfoil . The distance from the trailing edge to the surface is between five and fifteen percent of the chord length.a theoretical airfoil that is so thin that the tangency boundary conditions can be applied on the chordline. sink . terrafoil . source .

300 utilizes fan-fed static air cushions for levitation and guidance. upstream. Same as the TACRV. towards the side from which the wind is blowing. a theoretical point that causes the fluid in its vicinity to move in a circular pattern. levitating) form of suspension. Tracked Ram Air Cushion Vehicle. Tracked Levitated Research Vehicle. upwind. It was proposed by the Transportation Systems Center and by Mitre Corporation in 1974.e.. Any high speed ground transportation vehicle using a non-contacting (i. Tracked Levitated Vehicle. . Built for the US DOT by Grumman Aerospace between 1971 and 1974. proximal. A subset of TLV which is aerodynamically suspended (dynamic air cushion).

U. 1960. 301 BIBLIOGRAPHY [1] "Assessment of the Potential for Magnetic Levitation Transportation Systems in the United States. New Opportunities by the Federal Railroad Administration (FRA) of the United States Department of Transportation (USDOT). also "The "Aerodynamic Ground Ef­ fect" and the Development of the Aerofoil Boat. The Paradox of Negative Drag on Planing Surfaces. . [6] Collins Radio Co. [4] Lippisch. Iowa." Report and Report Supplement to the U. Alexander M.Aerodynamische Bodeneffekt" und die Entwick- lung des Flugflachen. part of the series Moving America: New Directions. Research on the Aerofoil Boat Concept Engineering Proposal. Alexander M." Luftfahrttechnik-Raumfahrttechnik Band 10. The Hydrodynamic Characteristics of Planing Surfaces^ Collins Engineering Report CER-1117-2.S. U.. [7] Lippisch. Colton Tow Tank Tests of a Low Aspect Ratio Ground Effect Surface. Jan 1964. Ground Effects Utilizing And Transition Aircraft. part of the Alexander Lippisch Papers at the Department of Special Collections at the Parks Library of Iowa State University (English translation). Research in the Field of Planing Surfaces at Collins Hydrodynamic Laboratory^ Collins Engineering Report CER-1117-1. and R.(Aerofoil-) Bootes.A. Nr.. [3] Lippisch. [2] Lippisch.190. Patent 3. 9 pages. 1960. Alexander M. Alexander M. 1 May 1960. Collins Radio Engineering Report (CER) 1117-8. Alexander M. 1963. Alexander M. [8] Lippisch.582 (22 June 1964).A." Unknown source. [5] Lippisch. Congress.F. 10 (Oct 1964): 261-269.S. June 1990. Collins Engineering Report CER-1117-4.S. Cedar Rapids. "Der .

Sherman M. "The Aerofoil Boat: A Free-Flying Ram-Wing/Surface Effect Vehicle.S. Ames. U.627. Patent 3." Sailplane and Glider (Lon­ don) Vol.A.S. Hydro-Space Systems Corporation Report 65-121. Society of Naval Architects and Marine Engineers. 11." in Jane's Surface Skimmers. The Aerodynamic Ground Effect Vehicle. New York: American Institute of Aeronautics and Astronautics. "Deep Water Birds and Gust Soaring. U. [16] Lippisch.582 (. telephone (515) 294-6672. Some Full-Scale Test Results." Flight International. 302 [9] Lippisch. and R.A. Alexander M." Department of Special Collections. [1. [19] Dugdale.S.A. [11] Lippisch. James L. [17] Lippisch.111 (May 9 1972).448 (Aug 20 1974). U. [15] Lippisch. Patent 3. [20] Corbett. Wing Arrangement. Alexander M. Technical &: Research Report R-9. Patent 3. also issued under former name: Institute of the Aeronautical Sciences.S. Colton Stepped Planing Boats.3] Lippisch. Alexander M.661. Ltd. Patent 3.235 (Dec 14 1971)." Sailplane and Glider (London) Vol. P.A. U.Dr.F. 14 (Jul 1946): 7-8. [21] Fitz Patrick. U. Aerofoilboat. 403 Parks Library. Alexander M. 1975: 368-375.190. Alexander M. 13 (Feb 1945): 6-7. [10] Lippisch. U. 1974-75 Edition.S. E.830.830. the "'Aerofoil Boat". Alexander M. Iowa 50011. L.179 (Aug 20 1974). Ground Effect Flying Surface. Alexander M.. "The Soaring Flight of the Albatross. Alexander M. 24 Nov 1965. London: Jane's Publishing Company. Alexander M. . Patent 3. Jun 1952: 118 pages. ISU. Aero Skimmer.A.. Fairchild Pub­ lication Fund Paper FF-7.G. Natural Flight and Related Aeronautics. Iowa. May 1970. "Dynamic Air-Cushion Vehicle . [12] Lippisch.A.June 1965). Cedar Rapids. [18] Iowa State University "The Alexander Martin Lippisch Papers. Lippisch Sends News of a Remarkable Craft. [14] Lippisch. Air-Cushion Vehicles Sup­ plement (London) (25 June 1964): 80-81. W.S.

C. [30] Nichols. "The Aeronautics of the Flying Fish. "On the Structural Specialization of Flying Fishes From the Standpoint of Aerodynamics. [28] Hubbs. Jr. E. 22. Carl L. 303 [22] Ahlborn. Washington. Hamburg. 22 (1937): 641." Zoologica (New York) Vol.A." Natural History (New York) Vol. [24] Hankin. Jr. [31] Breder. "Observations on the Flight of Fishes. 28 (Jan-Feb 1928): 64-77. H. [32] Breder. 17 (1932): 575-611. Papers Vol.S. C. U. Smithsonian Institution Annual Report. D. Breder. 1895: 35 pages. "The Flight of the California Flying Fish. Arts and Letters. 4 (31 Dec 1930): 114-121. M.. California. 9 (1929): 295-312." Michigan Academy of Science. U. 1914: 405 pages. Jr. [33] Shouletkin. 1933: 333-348. C. R. Was "Airdynamics of the Flying Fish. [25] Hubbs. M.) No. a Suggestion of Meth­ ods. London: IlifFe and Sons. "Observations on the Flight of Flying Fishes.S. T. "Field Observations on Flying Fishes. "Further Observations and Statistics on the Flight of Fishes." Copeia (Northridge. Carl L. a Record of Observation. [23] Hankin.. 2 (1920): 467-474.) No. [27] Hubbs.A. E. and C. 62 (1918): 85-88. 12 (10 Jan 1921): 464-465. Ger­ many) Vol. Germany) Vol. 1/2 (1929): 102-110. ." Flugsport (Frankfurt am Main. 25 (8 Nov 1933): 497-498. "About Flying Fishes. Nature's Own Seaplanes. [26] Hubbs. M. Carl L. Proceedings No." Aerial Age Weekly (New York) Vol." Copeia (Northridge." Internationale Revue der gesamten Hydrobiologie und Hydrographie (Leipzig. J. Animal Flight." Michigan Academy of Sciences. Norbert "Fliegende Fische. Germany: Real-Gymnasium des Johanneums.. H. Carl L." Zoological Society of London. Papers Vol. No. E. [34] Welkoborsky. [29] Dowd. Arts and Letters. With a Statistical Study of the Flight of the Cypselurinae and Remarks on the Evolution of the Flight of Fishes. Friedrich Der Flug der Fische. California.

591.475. [39] "Ground Effect. 82 Air-Cushion Vehicles Supplement Vol. also abridged in: Franklin Inst. 66 (1 Dec 1936): 623-625. Claudius Vortrdge und Ahhandlungen aus dem Gebiete des Flugzeug- baues und Luftschiffbaues. 10 (27 Jun . L.. Germany) Vol. N. [36] Latimer-Needham." Science et Vie (Paris).C. H. [40] Dornier. 60 (25 Oct 1951): 535-536.S. C. Germany. [38] Fontaine. Navy Yard. Claudius Flying Boat. 304 [35] Latimer-Needham. issued un­ der former name: The Sailplane. 1925 (English Translation). Albert "Auftrieb and Widerstand einer Tragflache in der Nahe einer hor- izontalen Ebene (Erdboden) (Mitteilungen aus der Gottinger Modellversuch- sanstalt).. abstracted by Klemin. H. M. Jan 1930: 145 pages. Washington. also issued under its former name: Construction Dept. 150 (Jan 1934): 37. W. 174-175. [42] Betz. D. "Flying-Fish Aerodynamics. 191 (May 1921): 687-693. Dayton." The Sailplane and Glider (London) Vol. and C. 6 (20 Dec 1962): 117 (bound in).C." Flight International (London) Vol. D. Jul 1926: 4 pages. "The Aerodynamics of the Flight of Flying-Fishes. 10 plates. Great Britain Aeronautical Research Council Report and Memorandum ARC R&iM 754 (July 1920): 10 pages." Zeitschrift fur Flugtechnik and Motorluftschiffahrt Jahrgang 3. Lock Cushioning Effect on Aeroplanes Close to the Ground.A. Karl "Beobachtungen iiber den Flug fliegender Fische. [37] Krejci-Graf. the Whys and Wherefores.. Maurice "Quand les Poissons Quittent Leur Domaine Aquatique. Vol. 1914-1930. Patent Office. H. 1. Washington. Ohio. Bear Ground-Plane Influence on Airplane Wings.S. 4 (11 Aug 1933). [41] Dornier.. Patent No. U. Albert F. Friedrichshafen. 1.." Natur und Volk (Frankfurt am Main. Alexander Scientific American (New York) Vol. [44] Zahm. and R. [43] Cowley. Vol. U. Journal.A. McCook Field Memorandum Report 167 of the Engineering Division of the Air Force Air Materiel Command. David Taylor Model Basin Aero Report 173 (Feb 1921): 6 pages. C." Flight International (London) Vol. 82 (Jul 1952): 25-30. Vol. No. Heft 17 (14 Sep 1912): 217-220. also abridged in: Aviation and Aircraft Journal (New York). also The Lift and Drag of a Wing in Proximity to the Ground.

Effect of Nearness to the Ground on the Aerodynamic Charac­ teristics of the Wing. 1923: 41-42 (Abridged Version). CAHI Technical Note 47. [45] Wieselsberger. Jahrgang 12 (31 May 1921): 145-147. A. [52] Serebrisky. Central Aero-Hydrodynamical Institute. Indiana. Moscow (1939): 12 pages. Germany) Heft 10. NACA TN 67 (Dec 1921): 8 pages. [53] Recant. National Advisory Committee for Aeronautics Technical Note NACA TN 705 (May 1939): 17 pages.S. Report (Vol. 305 1921): 807-808. also "Der Einfluss der Erdbodennahe auf den Fliigelwiderstand. Verlag Leemann." Ergebnisse der Aerodynamischen Versuchsanstalt zu Gottingen. II. Vol. Vol. Notre Dame. April 1922: 7 pages (English Translation). William R. . Moscow (1935). Lieferung. Arthur E. Susumu." Transactions of the Central Aero-Hydrodynamical Institute. Mitteilungen (I) aus dem Institut fiir Aerodynamik der Eidgenossischen Technischen Hochschule Zurich. M. Zahm Aeronautical Papers (1885-1945). CAHI Report No. 437 (Transactions 437) of the Central Aero-Hydrodynamical Institute Moscow (1939). "Experimental Investigation of the Vertical Approach of a Plate and the Inclined Approach of a Wing to the Ground. [48] Datwyler." Zeitschrift fiir Flugtechnik und Motorluftschiffahrt (Miinchen. Y." Aircraft Engineering 7 (Jan 1935): 24 (English Abstract). Carl. Zurich. [47] Tomotika. Y.. Gottfried Untersuchungen iiber das Verhalten von Tragfliigelprofilen sehr nahe am Boden. [49] Ushakov. CAHI Transaction 422. Ground Influence on Aerofoils. Sep 1946: 32 pages (English Translation). Wind-Tunnel Investigation of Ground Effect on Wings With Flaps. also "Investi­ gations about the Behaviour of Aerofoil Profiles very near the Ground. U. 2. also Wing Resistance Near the Ground. [50] Sears. also in: Albert F. "Uber den Fliigelwiderstand in der Nahe des Bodens. "Ground Effect With Special Reference to Pitching Moments. Isidore G. and S. [51] Serebrisky. NACA TM 77. [46] Raymond. A. Switzerland (1934): 109 pages. University of Notre Dame (1950): 625-628. also NACA TM 1095. M. B. Biachuev Wind Tunnel Investigation of the Hori­ zontal Motion of a Wing Near the Ground." Journal of the Aeronautical Sciences (New York). 7) of the Aeronautical Research Institute of Tokyo Imperial University (1933): 357-393.A. 5 (May 1938): 281-285. Promotionsarbeit 773 (disser­ tation).

A.. Lockheed-California Co." La Recherche Aerospatiale (Paris). Arthur W. [59] Carter. "The Forces Acting on a Circular-Arc Aerofoil in a Stream Bounded by a Plane Wall. [58] Bagley. ARC 22. Wind Tunnel Investigation of Single and Tandem Low- Aspect-Ratio Wings in Ground Effect. California. Heft 5 (Sep/Oct 1939): 220-226. [63] Onspaugh. and James L. NTIS N71-33494. Great Britain Aeronautical Research Council Report and Memorandum ARC R&M 3238 (Feb 1960): 40 pages. Horst "Stromungsbilder zum Bodeneinfluss. The Pressure Distribution on Two-Dimensional Wings Near the Ground. Part One Vol. Navy Department. 46 (14 Nov 1939): 19-54. Aug 1971: 18 pages. D.. NASA TN D-970 (Oct 1961): 27 pages. [56] Licher. Marvin P.S. Taylor Model Basin Aerodynamics Laboratory. Burbank.C. Lastinger Aerodynamic Characteristics of Low- Aspect-Ratio Wings in Close Proximity to the Ground. Wind-Tunnel Tests of a Low-Aspect-Ratio Wing in Close Proximity to the Ground. Herbert E. July 1961: 37 pages. Henri "Simulation de I'Effect de Sol au Tunnel Hydrodynamique. 95 (Jul-Aug 1963): 7-15. U. [61] VVerle. Douglas Aircraft Company Report SM-22615. NTIS N63-19883. Low-Speed Tunnel Tests on a Two-Dimensional Aerofoil With Split Flap Near the Ground. Carl M. A. . Jun 1963: 25 pages. [60] Fink.P. [62] White. J. No.. J. Washington. David W. Rose M. Santa Monica. 306 [54] Miiller. 63 (1963): 15 pages. A. Second Se­ ries. Technical Translation NASA-TT-F-13799. U. (Oct 1956): 46 pages. [55] Green.S. Lock­ heed Fluid Dynamics Laboratory (LFL) Report L-53." Proceedings of the London Mathematical Society. NASA TN D-926. also reprinted as: Office National d'Etudes et de Recherches Aerospatiales T. Increase in Lift for Two and Three Dimensional Wings Near the Ground. A. [57] Bagley." Forschung auf dem Gebiete des Ingenieurwesens Band 10. also Simulation of the Ground Effect in the Hydrodynamic Tunnel English translation.A. Effect of Ground Proximity on the Aerodynamic Characteris­ tics of Aspect-Ratio-1 Airfoils With and Without End Plates. E. Royal Aircraft Establishment Technical Note Aero 2636 (Mar 1961): 15 pages. May 1963.060 {Yeh 1960): 40 pages. No. also Royal Aircraft Establishment Report Aero 2625. Aero Report 1056.

Brainerd Design of Wing Sections for Use Near the Ground. J.C. Hansen Wind-Tunnel Investigation of Single and Tandem Low-Aspect-Ratio Wings in Ground Effect Lockheed Report LR 16906. AD-465949. [65] Harry. California. NTIS N68-32300. Torstein and J. U. and Samuel G. U. for the Federal Rail­ road Administration. San Diego.. D. Mar 1968: 54 pages. Fort Eustis. 15 (Feb 1972): 10-19. Taylor Model Basin Aerodynamics Laboratory. Burbank.. with and without Endplates.A. also Report SRL-TR-71-0012 US Air Force Systems Command (July 1971).S. Cambridge. March 1964 (analysis of data from Lockheed Report L-53): 100 pages. David W. [71] Boccadoro. Air Vehicle Corporation. Miller and Woodrow D. [66] Schaeffer. Mark K. in Ground Effect prepared at the College of Aeronautics. An Experimental Investigation into the Aerodynamic Character­ istics of a Wing. U.. Aero Report 1086. Washington. Massachussetts.LT. E.C.. U.. 1972: 11 pages. Fink Evaluation of Airfoil Section Char­ acteristics in Ground Effect Navy Department. Report COA-AERO-201. Report FRA-RT-73-34. P.476. NTIS PB 222. Mar 1965.A.. Cranfield. Smith "The Ram-Wing Surface Effect Vehicle: Comparison of One-Dimensional Theory with Wind Tun­ nel and Free Flight Results. also TRE- COM Technical Report 63-63.S. DTMB Technical Note AL 2 (Oct 1965): 22 pages. D. Viginia. [68] Kumar. Charles W. Taylor Model Basin Aerodynamics Lab­ oratory.S. No. United Kingdom. . Lockheed-California Company. 11. NTIS AD 474 257. Ram Wing Surface Effect Boat AIAA Paper 72-601." Hovering Craft & Hydrofoil Vol. [69] Gallington. William P. U. Washington. Army Transportation Research Command (USATRECOM).. Roger W. 307 [64] Stevens. Yves A. N65- 28802.S. Richard L. Towing Tank Tests on a Ram Wing in a Rectangular Guide- way prepared at M.A. Wind Tunnel Investigation of Ground Effect on a Rectangular Wing of Several Moderate Aspect Ratios Bureau of Naval Weapons (BuWeps) Problem Assignment 1-34-52. U.A. [70] Gallington.A. Jul 1973: 113 pages. David W. and Martin D.. [67] Strand.S. Report 349. Jul 1965: 65 pages. Roger W.S.

" Zeitschrift des Vereins Deutscher Ingenieure (Berlin. also AMS Technical Report 1318.. Germany) Band 84.12. [78] Chaplin. for the Office of the Secretary of Transportation at the Transportation Systems Center of the U. Germany. [74] Steinbach. Buffalo.S. B. T. Feb 1979: 111 pages. Department of Transportation at M.A.A. NTIS N86-10026. Maurice H. Cahfornia.A. U. also revised as: Army Transportation Research Command TCREC Technical Report 61-108. L.. Aeronutronic Division of Ford Motor Co. [79] Mankuta. Nr. Harvey R. Jan 1978: pages. New York. and W. also ISSN 0171-1342. R. [73] Haake. U-926. (Jun 1960): 6. K. for the German ministry of defense. Interim Report DOT-TSC-OST-78- 1.S.. U. Putman Experimental Investigation of Aerody­ namic Characteristics of a Tracked Ram Air Cushion Vehicle prepared at Prince­ ton University. (Apr 1961): 39 pages. [77] Carmichael.A. E. Dieter Experimental and Theoretical Investigations of the Flow Around Airfoil Systems with Ground Effect prepared at the Institute for Theo­ retical Fluid Mechanics at Gottingen. U. Rawlings State of the Art Summary: Air-Cushion Vehicles. Publ.und Raumfahrt at the RWTH Aachen.S. M. David Taylor Model Basin Report 1463 and Aero Report 994. Staufenbiel. [80] Liberatore. "Stromungsbilder zum Bodeneinfluss.. Massachussetts. Revision 2.I.S. Newport Beach.S. New York. D. Report BMVg- FBWT-79-4. Bruce H. Report DFVLR-FB-85-17. Bell Aerosystems Co. M. Literature Research 15. [76] Smith.A. Report 2017-945002. .. Feb 1985: 63 pages..A. and Mary Lee Bibliography on Ground Effects.1-6. Ground Effect Machine Research and Development in the United States. [75] Huber. Harry Ground Effect Machines Morphology Study. Princeton University James Forrestal Research Center Library (Oct 1959): 9 pages. Buffalo. Weckesser. Yeh Flugmechanische und aerodynamische Untersuchungen zum Staufiiigelprinzip prepared at the Lehrstuhl fiir Luft.S. Dobson and William L. Washington. also Report DOT-TSC-OST-77-35. U. (Dec 1960): 14-16. U. 29 (20 Jul 1940): 522- 523.S. 308 [72] Curtiss.C... (Jan 1961): 34-36. NTIS N78 24099 or PB-277 674. F. NTIS N80-17037. GEM Activities and Bibliography. Bell Aerosystems Co. U.T.C. Jr. Kleineidam. Wessels and B. Cambridge. H. Germany. G. Frank A.

309

[81] Rawlings, William L. and Donald H. Seiveno State-of-tlie-Art Surmnary;
Air-Cushion Vehicles. Revision i, Army Transportation Research Command
TCREC Technical Report 61-108, Fort Eustis, Virginia, U.S.A. (Aug 1961):
6.1-7.1; also Aeronutronic Division of Ford Motor Co. Publication U-926-RV 1.

[82] Earl, T. D. Ground Effect Machines NATO Advisory Group for Aeronautical
Research and Development AGARDograph 37, Paris, France (Jan 1962): 34.

[83] Park, Ford "Near-Surface Vehicles." International Science and Technology (New
York) (Feb 1962): 14-15.

[84] Borgenstam, Curt "Hovering Craft; Their Future Use and Development." Hov­
ering Craft & Hydrofoil {London) Vol. 2 (Oct 1962): 9, 11.
[85] Hayward, L. H. "The History of Air-Cushion Vehicles." Hovering Craft & Hy­
drofoil (London) Vol. 2 (Dec 1962): 12-17 and (Jan 1963): 12-18; also issued as
monograph by Kalerghi-McLeavy Publications, London, United Kingdom.

[86] Armfield, William J., IV et. al. Air Cushion Vehicles; Transportation of the
Future., New York, Transportation Research Associates (1962): 1, 2, 26-27, 43-
45.

[87] Tinson, Clifford W. "Ground Effect." Flight International (London) Vol. 83, Air
Cushion Vehicles Supplement Vol. 2, No. 7 (24 Jan 1963): 14.
[88] Foshag, William F. Literature Search and Comprehensive Bibliography of Wings
in Ground Effect and Related Phenomena DTMB Report 2179 (Mar 1966).
[89] Reid, Elliott G. >1 Full-Scale Investigation of Ground Effect, NACA Report 265
(1927): 233-237.

[90] Reid, Elliott G. and Thomas Carroll A Warning Concerning the Take-Off With
Heavy Load, NACA TN 258 (Jul 1927): 6 pages.
[91] Tonnies, E. "Der Boden-Effekt beim Fluge in Erdnahe." Zeitschrift fiir Flugtech-
nik und Motorluftschiffahrt (Miinchen, Germany) Vol. 23 (29 Mar 1932): 157-
164; also Effect of the Ground on an Airplane Flying Close to it, NACA TM 674
(Jun 1932): 16 pages (English Translation).

[92] Wetmore, J. W. and L. L Turner, Jr. Determination of Ground Effect From
Tests of a Glider in Towed Flight, NACA Report 695, 1940: 13 pages.

310

[93] Buell, Donald A. and Bruce E. Tinling Ground Effects on the Longitudinal Char­
acteristics of Two Models With Wings Having Low Aspect Ratio and Pointed
Tips, National Advisory Committee for Aeronautics Technical Note NACA TN
4044 Washington, D.C. (Jul 1957): 48 pages.

[94] Buell, Donald A. and Bruce E. Tinling The Static Longitudinal Stability and
Control Characteristics in the Presence of the Ground of a Model Having Trian­
gular Wing and Canard, National Aeronautics and Space Administration Memo.
3-4-59A, Washington, D.C., 1959.

[95] Doetsch, K. H., G. C. Howell and W.F.W. Urich The Landing Flare of a Gothic-
Wing Aircraft of Aspect Ratio 1.0 Taking Account of the Aerodynamic Ground
Effect, Great Britain Royal Aircraft Establishment Technical Note Aero 2600,
Revised, Farnborough, United Kingdom, Jan 1960: 7 pages.

[96] Peckham, D. H. Ground Effect on a Gothic Wing of Aspect Ratio 1.0, Great
Britain Royal Aircraft Establishment, unpublished.

[97] Pitkin, M. Preliminary Technical Concept Evaluation - the Aerofoil Boat, Mar­
itime Administration (MARAD), 1972.

[98] Kaario, Toivo J. Translations from the Finnish of Two Patents and One Paper
Describing T. J. Kaario's Early Work on Air Cushion Vehicles (1932-1949),
Aerophysics Company Report AR-595 , Nov 1959: 6-15; includes "Ilman Kan-
tama Pintakulkuneuvo [Air Cushioned Surface Vehicle]." Teknillen Aikakauslehti
(Helsinki) Vol. 32, (Feb 1942): 43-47.

[99] Kaario, Toivo J. "The Principles of Ground Effect Vehicles." in: Princeton Uni­
versity and U.S. Army TRECOM Symposium on Ground Effect Phenomena:
a Compilation of the Papers Presented October 21-22-23, 1959, Philadelphia,
U.S.A.: IVY-CURTIS Press, March 1960: 253-262.

[100] Princeton University and U.S. Army TRECOM Symposium on Ground Effect
Phenomena; a Compilation of the Papers Presented October 21-22-23, 1959,
Philadelphia, U.S.A.: IVY-CURTIS Press, March 1960.

[101] Anonymous "New Canoe has Wings." Mechanics and Handicraft (New York)
Vol. 3 (Oct 1936): 27.

[102] Crook, Louis H. Warner Jet Propelled Compression Plane, Aerodynamic Lab­
oratory. Aeronautical Report 560, Washington, D.C., U.S.A., Jun 1943: 6 pages.

311

[103] Warner, Douglas K. Compression Airplane. Washington, D.C., Dec 1944: 11
pages (Patent Off. 2,364,677).

[104] Warner, Douglas K. Skimming and Flying Vehicle^ Washington, D.C., Dec 1944:
9 pages (Patent Off. 2,364,676).

[105] Daniels, Charles J. Tests of a Model of a Flying-Boat Hull Incorporating D. K.
Warner's "Compression-Plane" Principle, NACA 171, Washington, D.C., 1944;
also National Advisory Committee for Aeronautics Memorandum Report L4F24.

[106] Mottard, Elmo J. and Robert D. Ruggles Tank Tests of a Powered Model of
a Compression Plane; NACA Model 171A-2, National Advisory Committee for
Aeronautics, Research Memorandum SL8G02, Washington, D.C., Jul 1948: 12
pages.

[107] Strand, Torstein and Scott Rethorst Interim Report on VRC Channel GEM
Concept, Vehicle Research Corporation, Pasadena, U.S.A., Nov 1960; presented
at ONR/TRECOM GEM Conference, Ft. Myer, Va., U.S.A., 16-18 Nov 1960.

[108] Strand, Torstein, Winston W. Royce and Toshie Fujita Performance Theory
for High Speed Ground Effect Machines, Vehicle Research Corporation Report
11, Pasadena, U.S.A., Jun 1961: 23 pages.

[109] Royce, Winston W. and Scott Rethorst Translational Characteristics of Ground
Effect Machines, IAS Paper 61-79 (Jan 1961): 24 pages; presented at IAS 29th
Annual Meeting, New York, 23-25 Jan 1961 (Institute of Aerospace Sciences
later became the American Institute of Aeronautics and Astronautics).

[110] Strand, Torstein "150 Knot GEM Cruise." in Institute of the Aerospace Sci­
ences. Proceedings of the Ground Effect Machines Forum, January 22-24, 1962.
New York, 1962: 32-37; also American Institute of Aeronautics and Astronau­
tics, Sherman M. Fairchild Publication Fund Paper FF-32; also published in
Aerospace Engineering (New York) Vol. 21 (Apr 1962): 38-44.
[111] Strand, Torstein, Winston W. Royce and Toshie Fujita "Cruise Performance of
Channel-Flow Ground-Effect Machines." Journal of the Aerospace Sciences Vol.
29, No. 6 (Jun 1962): 702-711, 718.

[112] Rethorst, Scott. "The 100-knot Columbia." Flight International (London) Vol.
82 (25 Oct 1962) Air Cushion Vehicles Supplement Vol. 1, No. 4: 73-76 (bound
in).

312

[113] Rethorst, Scott. "VRC Surface Effect Ship Columbia." Proceedings of the Na­
tional Meeting on Hydrofoils and Air Cushion Vehicles, Washington, D.C.,
September 17-18, 1962, New York: Institute of the Aeronautical Sciences (IAS)
and US Navy, 1962: 135-140; abstracted in: Hovering Craft & Hydrofoil (Lon­
don) Vol. 2 (Oct 1962): 22-23. (IAS later merged with the American Rocket
Society to become the American Institute of Aeronautics and Astronautics).

[114] Strand, Torstein, Roy G. Fowler and Hideo Yoshihara Increased Airplane En­
durance Through Employment of Wing End Plates in Ground Proximity, Air
Vehicle Corporation, San Diego, U.S.A., Dec 1962: 14 pages.

[115] Rethorst, Scott and W. T. Potter "Maritime Administration Surface Effect
Ship." New York: Society of Automotive Engineers, Preprint 697D, Apr 1963: 6
pages; presented at the SAE-ASNE National Aeronautical Meeting, Washington,
D.C., 8-11 Apr 1963.

[116] Jane's Yearbooks Jane's Surface Skimmers, 1967/68 Edition, London: Jane's
Publishing Company, Ltd., 1968.

[117] Anonymous "Aeromobilia." Hovering Craft & Hydrofoil (London) Vol. I (Dec
1961): 9-11.

[118] Bertelsen, William R. "The Arcopter GEM Project." Hovering Craft & Hydro­
foil (London) Vol. 1 (Jul 1962): 21-23.
[119] Anonymous "Across the Mississippi." Flight International (London) Vol. 82 (19
Jul 1962) Air-Cushion Vehicles Supplement Vol. I, No. I: 3-4 (bound In).

[120] Strang, C. "With the Outboarders." Motor Boating (New York) Vol. 108 (Aug
1961): 32, 54.

[121] Ando, Shigenori and Jun-ichi Miyashita "Comments on Aerodynamic Drag
of Ground Effect Machines." Aerospace Engineering (New York) Vol. 20 (Nov
1961): 24-25, 79, 81-83.

[122] Anonymous "Japanese Ram-Wing." Flight International (London) Vol. 82 (25
Oct 1962) Air-Cushion Vehicles Supplement Vol. 1, No. 4.: 70 (bound in).

[123] Anonymous "Kawasaki's Ram Wing." Flight International (London) Vol. 82
(22 Nov 1962) Air-Cushion Vehicles Supplement Vol. 1, No. 5: 88 (bound in).

[124] Ando, Shigenori, Jun-ichi Miyashita and K. Teraj "Design Philosophy and
Test Results of Two-Seated Experimental Simple Ram-Wing with Side Floats."
Kawasaki Review No. 14 (Mar 1964): 3-34.

313

[125] Ando, Shigenori "Japan's Ram-VVing Research." Flight International (London)
VoL 85 (28 May 1964) Air-Cushion Vehicles Supplement Vol. 4, No. 23; 72
(bound in).

[126] Ando, Shigenori, Jun-ichi Miyashita and K. Teraj "Summary of the Model Tests
for Simple Ram Wing KAG-3." Hovering Craft and Hydrofoil Yo\. 3 (Aug-Sep
1964): 38-50.

[127] Ando, Shigenori "An Idealised Ground Effect Wing." Aeronautical Quarterly
(Feb 1966); 53-?.

[128] Ashill, P.R. "The State of the Art of Ram-Wing Research." Hovering Craft and
Hydrofoil Vol. 2 (May 1963); 18-20, 22; also Note 159 of the Engineering College
of Aeronautics, Cranfield, United Kingdom (Nov 1963); 14 pages.

[129] Ashill, P.R. "Kawasaki KAG-3." Flight International (London) Vol. 85 Air-
Cushion Vehicles Supplement Vol. 4, No. 20 (27 Feb 1964); 29-30, 17 (bound
in).

[130] Ashill, P.R. On Some Aspects of the Aerodynamic Performance of Ground-
Effect Wings. Ph.D. Thesis, The University of Southampton, 1968.
[131] Ashill, P.R. "On the Minimum Induced Drag of Ground-Effect Wings." The
Aeronautical Quarterly Vol. XXI (Aug 1970); 211-232,
[132] Lockheed-California Co. Low-Speed Wind Tunnel Tests of the Winged Hull.
Phase i. Report LFL L-47-1, Burbank, California, U.S.A., Aug 1961.
[133] Lockheed-California Co. Research Proposal of Response of Aerodynamic Lifting
Surface in Ground Effect to Passage Over Sinusoidal Surface., Report 15778,
Burbank, California, U.S.A., Mar 1962; 12 pages.

[134] Lockheed-California Co. Proposal for Evaluation of the Effects of Surface Waves
on the Performance of a Winged Hull Test Vehicle., Report LR 15785, Burbank,
California, U.S.A., Mar 1962; 18 pages.

[135] Anonymous, A Design Study of Winged Hull Vehicles Configured for
Transoceanic Logistics^ Lockheed-California Co. Report LR 16027, Jul 1962:
110 pages.

[136] Anonymous, Operational Analysis of the Winged Hull Concept in ASW Oper­
ations., Lockheed-California Co. Report LR 16243, Sep 1962.

314

[137] Anonymous Preliminanj Economic Analysis of Winged Hull Vehicles.
Lockheed-California Co. Report LA/ME/2148, Sep 1962.

[138] Anonymous, Preliminary Economic Analysis of Passenger Carrying Winged
Hull, Lockheed-California Co. Report LA/ME/2161, Oct 1962.
[139] Anonymous "Air-Cushioned, This Lockheed "Winged-Hull" .. Japanese Ram-
Wing." Flight International (London) Vol. 82 Air-Cushion Vehicles Supplement
Vol. 1, No. 4 (25 Oct 1962): 70-71 (bound in).

[140] Lockheed-California Ca. Technical Proposal for a Stability and Control Study
of Winged-Hull Vehicles, Report 16429, Burbank, California, U.S.A., Nov 1962:
25 pages.

[141] Lockheed-California Co. Low-Speed Wind Tunnel Tests of a Winged Hull Con­
figuration, Report LFL L-50, Burbank, California, U.S.A., Jan 1963.
[142] Anonymous A System Analysis for Limited War, Lockheed-California Co. Re­
port LR 16812, May 1963.

[143] Lockheed-California Co. Quick Response Logistics Vehicle Design Study, Report
16811, Burbank, California, U.S.A., May 1963: 184 pages.

[144] Anonymous "Water Wind Tunnel." Flight International (London) Vol. 84 Air
Cushion Vehicles Supplement Vol. 3, No. 13 (25 Jul 1963): 3 (bound in).
[145] Swope, W. A. et al. Winged Hull Stability and Control Study, Lockheed Cali­
fornia Company Report, Burbank, California, U.S.A., AD 641901, Jul 1964.

[146] Herman, J. F. and M. L. Jager and P. F. Balla Winged Hull Stability and
Control Study, Lockheed-California Co. Report LR 17910, Jul 1964; revisions
Feb 1966 and Aug 1966.

[147] Alter, Horace J. and Vsevolod B. Koriagin Winged Hull Vehicle, U.S.A. Patent
3,216,673 (9 Nov 1965).

[148] Fridsma, G. Model Tests of the Lockheed Winged Hull, Davidson Laboratory,
Stevens Institute of Technology, Jan 1967.

[149] Stevens, William P. Feasibility Study of a Dynamic Interface Vehicle Config­
ured for an Amphibious Assault Mission, Lockheed-California Co. Report 15736,
Burbank, U.S.A., Feb 1962: 16 pages.

315

[150] Walters, Donald E., E. G. Stout, T. P. Higgins, Jr. and J.R. Brown The Dy­
namic Interface Vehicle, New York: American Institute of Aeronautics and As­
tronautics Paper 62-140, Jun 1962: 8 pages; also presented as IAS Paper No.
62-140 at the IAS National Summer Meeting, Los Angeles, California, 19-22 Jun
1962.

[151] Anonymous "You can Fly This $300 Boat!" Mechanix Illustrated Vol. 58 (Feb
1962): 65-67 & cover.

[152] Barlow, Gerald R. and R. G. Huntington Preliminary Study of Hover Perfor­
mance of a Ram-Wing GEM, David Taylor Model Basin Report 1662 and Aero
Report 1015, Washington, D.C. (Oct 1962): 39 pages.

[153] Barrows, Timothy M. and S. E. Widnall The Aerodynamics of Ram-Wing Ve­
hicles for Application to High Speed Ground Transportation, AIAA Paper 70-142
(Jun 1970).

[154] Barrows, Timothy M. Analytical Studies of the Lift and Roll Stability of a Ram
Air Cushion Vehicle, performed for the DOT by the Transportation Systems
Center, Interim Report DOT-TSC-FRA 72 10; also Report FRA-RT-73-21 (Dec
1972), NTIS PB-219 820: 67 pages.

[155] Anonymous "Causing a Mild Sensation on the Connecticut River From Essex
South to Long Island Sound is a 26 ft. Ram-Wing Boat Designed by Norman
Dickinson." in People and Projects, Hovering Craft & Hydrofoil (London) Vol.
2 (Sep 1963): 4.

[156] Kumar, P.E. Stability of Ground Effect Wings. College of Aeronautics Cranfield
Report Aero 196, AD 853136, Jul 1964.

[157] Yaffee, M. L. "Ram Wing studied as Transit Vehicle." Aviation Week & Space
Technology (2 January, 1967): 56-58, 60.
[158] Gallington, Roger W. "Ram Wing Surface Effect Boat." Journal of Hydronau-
tics Vol. 7, No. 3 (Jul 1973): 118-123.
[159] Gallington, Roger W. Vortex Shedding from the Ram Wing Vehicle, Paper pre­
sented to International Hovering Craft, Hydrofoil and Advanced Transit Systems
Conference, Brighton, England, (May 13-16, 1974).

[160] Miller, Jeffrey Arnot The Dynamic Compensation of a Ram Wing Surface Effect
Vehicle, Masters Thesis, Brown University (Nov 1974).

[167] Dunham. Aerophysics Co. D." in Sterk.. 316 [161] Anonymous "Ram Wing Vehicles as Water-Borne Transport Facilities. [168] Jane's Yearbooks Jane's Surface Skimmers. William R. [170] Anonymous "Ground-EflFect Goliath to Thunder Across Tundra.C. Proposal AP-28. Washington. [166] Dunham.321 Washington. William H. [163] Korvin-Kroukovsky. Patent 3. Davidson Laboratory.322. Daniel and William F. [171] 3a. 1971. 1981 Edition.. B. Stevens Institute of Technology. 13..S. Davidson Laboratory. Stevens Institute of Technology.) Unconventional Aircraft Concepts: Papers pre­ sented at a symposium organized by the Netherlands Association of Aeronautical .M. (Feb 1963): 18 pages. No.M. London: Jane's Publishing Company. D.C.V. David Taylor Model Basin Report 1571.. Savitsky. 2 (Nov 1974). [165] Foshag. Design and Fabrication of Two "Hydro-Skimmer" Models." Popular Me­ chanics (June 1992): 15. and E. [172] Lange. Lehman Wave Con­ tours in the Wake of a 10-degree Deadrise Planing Surface. Daniel and Joseph W. B.077.223 (30 May 1967). Ground Effect Machines.C. Report 344. Preliminary Results of Wind-Tunnel Tests of the HS-2 Hydroskimmer. Jun 1948. Neidinger Wetted Area and Center of Pressure of Planing Surfaces at Very Low Speed Coefficients.. Nov 1948. 1981. U. Savitsky.J. F. London: Jane's Pub­ lishing Company." U.F Publication Fund Paper FF-11. Davidson Labo­ ratory. Washington. Report 493.F Publication Fund Paper 170.A. D. 1970/71 Edition.ne^s Yearbooks Jane's Surface Skimmers. State of the Art and Future Development" Hovering Craft and Hydrofoil Vol. Ltd. Oct 1961: 34 pages. also S. Lehman Wave Con­ tours in the Wake of a 20-degree Deadrise Planing Surface. [164] Savitsky. Report 337. Patent 3. July 1954. Ltd.V. Jan 1961: 10 pages. [162] Korvin-Kroukovsky. William H. "Survey of Unconventional Aircraft Design Concepts. "Aerodynamically Designed Amphibious Vehicle. William F.S. Daniel and William F. Torenbeek (eds. [169] Bertelsen.. also S. Stevens Institute of Technology. Roy H.

Parametric Investigation of a Power Augmented Ram Wing with Load Alleviation Devices Over Water Waves of Various Sea States. Dec 1964. 1976. [173] Lange.E. [176] McCabe. (Feb 24-26. Charles W." Popular Mechanics (May 1992): 35-38. Design Concepts for Future Cargo Aircraft. DTNSRDC Report ASED TM 16-76-97." Popular Science (Oct 1980): 72-75 [179] Jane's Yearbooks Jane's Surface Skimmers. (Jul 1965). Dec 1964. Parametric and Conceptual Design Study of Aircraft Wing- in-Ground Effect (WIG) Vehicles. U. Report No. Abe "VVingships: Massive VVing-In-Ground-EfFect Flyers combine jet­ liner speeds with steamship Economy. Roger W. 76020-30 and Lockheed-Georgia LG77ER0049.W. [177] Moore. Ltd. Trobaugh "Wind Tunnel Investigation of Ground Effect on a Rectangular Wing of Several Moderate Aspect Ratios.M. 317 Engineers (NVvL) and the Students Society 'Leonardo da Vinci' on April 1987 at the Delft University of Technology. Earl F. presented at the AIAA 11th Annual Meeting and Technical Display. May 1977. [183] Harry. (Jul 1964).. Delft: Delft University Press. 123. J. Ben "New-Technology Monster Transports. Roy K.C.S. Army Transportation Research Command TRECOM Technical Report 64-54. 1975). and J. 1980 Edition. M. London: Jane's Pub­ lishing Company. [182] Condit. 75-306. 692. 1980. 1987: 1-50. D. al. DTNSRDC Report ASED-365. Jr." DTMB Report 1979. et. E. Parametric Investigation of a Power Augmented Ram Wing Over Water. and Lynn A. and Harvey R. P. Chaplin Theory of Power Augmented Ram Lift at Zero Forward Speed. Washington.. also Princeton University Report No. [175] Krause. Oct 1976.. [178] Kocivar. Feb 1976. Harrington An Investigation of a Dynamic Instability of a Winged G.H. F. ASED TM 16-76-95. . AIAA Paper No. Lockheed Report LR 17910. [181] Anonymous. [174] Gallington. [180] Dane.

Ltd. [192] Jane's Yearbooks Jane's Surface Skimmers. NTIS AD 641 246." Jane's Surface Skimmers (1975/6): 425-433." Popular Mechanics (July 1989): 57-59. 1972/73 Edition." Offshore (Mar 1991). Pennsylvania. 7 (Jul 1967): 74-84. 1977. P. Jul 1966: 71 pages. and Lynn A. Aerodynamics Labora­ tory..: Transportation Re­ search Institute at Carnegie-Mellon University (1969): 17-27. 75. Brasseur An Evaluation of the Wing-in-Ground Effect (WIG) Transport Aircraft Concept. 111.S.A. [195] Giraud. 1973. and G. DTMB Report C-2318. and Richard R. D. David "We Have Lift Off (Boat That Flies).S." DTMB Report 2259-1. [194] Shilkoff.. U. 1976/77 Edition. Washington. also Report Aero- 1111-1.A. 1 (Jan 1968): 149-164. Tim "License to Fly. Nov 1966. John H. [189] Jane's Yearbooks Jane's Surface Skimmers. Greer "A Conceptual Study for a New Winged Surface Effect Vehicle System. [187] Cameron. London: Jane's Publishing Company. London: Jane's Publishing Company. No. Pittsburgh.. 318 [184] Harry. PB 178 260. [191] McMasters. "The Boat that Flies. Charles W. [196] Lissaman. [185] Cole. and Richard R. S. Ltd. "How Fast can Trains go?" SAE Journal Vol. No. Greer "Large Winged Surface Effect Ve­ hicles. Trobaugh "Wind Tunnel Investigation of an Aspect Ratio 10 Tandem Wing Aircraft Configuration in Ground Effect. Marilyn J. Journal (April 1974): 41- 51. [188] Comisarow.C. John H. David Taylor Model Basin." in: High Speed Ground Transportation: Proceedings of the Carnegie-Mellon Conference on High Speed Ground Transportation. "Down to Earth Aerodynamics: Potential and Problems of the Fluid Mechanics of Rapid Ground Transportation. .Longitudinal Characteristics. [193] Stanford Research Institute Future Urban Transportation Systems: Technolog­ ical Assessment MR-2.. U. Kevin A. Part I . P. W. 124-125." Popular Science (April 1992): 56-60. May 1967. 2." High Speed Ground Transportation Journal Wo\. B. Francois "Tracked Air Cushion Vehicles. [190] McMasters. [186] Seidman." Naval Eng.

also in Journal of Fluid Mechanics Vol. August 1971. al. J. Davis and G. North-East Corridor Transportation Project NECTP-222. 1968..P. U. MIT Fluid Dynamics Research Laboratory Report 69-1 (Jun 1969). Pittsburgh.C.S. 4. 271-276. July 1970. 3 (1969): 352. A Cost Comparison of three Tracked Air Cushion Ve­ hicle Configurations..A. Part 4 (1970): 769-792. High Speed Ground Modes. PB 204 800. Engineering Study. [204] Berthelot. Large Cost Analysis for NECPT. [199] Widnall. and J. [206] Harris. P. London: Jane's Publishing Company. [203] Saito. De­ cember 1969. [205] Tracked Hovercraft Ltd.. Vol." Jane's Surface Skimmers {impOy. Hal Transportation in the World of the Future." in: High Speed Ground Transportation: Proceedings of the Carnegie- Mellon Conference on High Speed Ground Transportation. G. 2 (1970): 197-209. Jean M. North-East Corridor Transportation Project NECTP-219.: Transportation Research Institute at Carnegie-Mellon University (1969): 47-57. Evans and Company Inc. "Economics Of The Aerotrain Systems. Seemann "The Terrafoil: A New Concept in High Speed Ground Transportation. No. and Timothy M. 1968/69 Edition." High Speed Ground Transportation Journal Vol. No. [200] Dienemann.F.F. 1969. [202] Williams. Transportation Systems Technology . Ltd. Pennsyl­ vania.F. R. et. 3. Tstsuru "The New Tokaido Line: Past Performance and Future Prospects." High Speed Ground Transportation Journal Vol. and Gerald R. Tracked Air Cushion Vehicle. Barrows An Analytic Solution for Two- and Three-Dimensional Wings in Ground Effect. Gordon L. Federal Railroad Administration Report FRA-RT-71-68. I.A Twenty-Year Out­ look. New York: M. [198] Jane's Yearbooks Jane's Surface Skimmers. [207] Kovatch. Martin "Preliminary Investigation of the Lift and Drag of an Airfoil Operated between Parallel Planes. . Sheila E. 41. December 1969. [201] TRW Systems Group HSGT Systems. 319 [197] Hellman.

(1978).High Speed Ground Transportation of the Future. 3 (1974): 167-183. Barrows The Tracked Ram Air Cushion Vehicle (TRACV). Nov 1973: 224 pages. 126. [211] Fraize. Goss and Duane E. NTIS PB-230 485. [215] Lehl. 5. C. E. . Paul "Personal Rapid Transit. Report DOT- FRA-ORD&D-74-27. Frank and Lyle 0. William P.P. 320 [208] VVahl. and D. Rinehart and Winston. Howard T. Chilton. [217] Edward De Bono. No. W. No. USA (December 1972): 146 pages. David "Single-Rail Hovercars Geared For Low-Cost City Transit. 8. [216] Tan. Cromack "The Design of a Ram Wing Vehicle for High Speed Ground Transportation." Pop­ ular Science (Oct 75): 64. 3 (1971): 427-441. 136. U. Jane High Speed Trains. 2 (1974): 11-15. [210] Tan. Engnam Anthony The Design of a Ram-Wing Vehicle for High Speed Ground Transportation. A System Definition Study. performed for the Office of Research. W. 1976. and Timothy M. No. Engnam Anthony. New Jersey... editor Eureka! An Illustrated History of Inventions from the Wheel to the Computer.S. [212] Ellis. Masters Thesis. Herbert "New Turbotrains Pace the Comeback of Passenger Rail­ roads. Secausus. Hamilton The Lore of the Train." High Speed Ground Transportation Journal Vol. [220] Collins. London: Jane's Publishing Company.E. [213] Coffey." Popular Science (Nov 1971): 73-77. Cromack "Transportation in a Tube ." Jane's Surface Skimmers (1973/74): 351-357.A. Hoppie "Magnetic Levitation For Tomorrow's Transportation. [218] Scott. Holt. Elvest L." High Speed Ground Transportation Journal Vol. Amherst. New York (1974). 8. University of Massachusetts. 1975/76 Edition. Vol." Popular Science (Oct 1974): 66-69. Development and Demonstration of the Federal Railroad Administration (Department of Transportation). Zumwalt "The Airtrain: An Aircraft Technology Ground Transportation System. and Glen W." High Speed Ground Transportation Journal.: Chartwell Books Inc. Mitre Corporation Report MTR-6554. [209] Goss. [214] Shuldiner. Ltd. New York: Crescent Books (1973). [219] Jane's Yearbooks Jane's Surface Skimmers.

321 [221] Sato.C. Report by the US General Accounting Office (GAO) to the US Senate. D. 114. New Opportunities series." Popular Mechanics (Nov 1983): 90-93." Popular Mechanics (May 1985): 102- 105." Popular Mechanics (Apr 1980): 96-100. Report GAO/RCED-92-82 (February 1992): 50 pages." National Research Council.S. J. Washington. [231] Transportation Research Board "In Pursuit of Speed: New Options for Intercity Passenger Transport. [223] Hollingsworth. 188. 89. [228] Anonymous "Air-Powered El (Elevated) Train Would Hit 300 Mph. 1991. Gregory T. SP3. ." Popular Mechanics (Mar 1991): 17. [232] "High-Speed Ground Transport: Acquiring Rights-of-Way for Maglev Systems Requires a Flexible Approach"." Popular Mechanics (Jun 1988): 78-80. J. [230] "Safety Relevant Observations on the TGV High Speed Train" in Moving Amer­ ica: New Directions. New Opportunities series. Technological Advances in Vehicle Design Series (of books). Brian and Arthur F. [222] Eskow. ISBN 0-309-05122-3. "Supertrain: New Superconductors Rekindle Interest in Lev­ itated Locomotives. [225] Eskow.A. [226] Pope. TRB Special Report SR 233. Dennis "Japan Struts Its Stuff." Electric Lines (July-August 1990): 8-11. Dennis "Coming: The Fastest Train In The West. Cook The Illustrated Encrjclopedia of the World's Modern Locomotives. [227] Anonymous "TGV: The World's Fastest Train.. New York. [229] "Safety of High Speed Magnetic Levitation Transportation Systems: Prelimi­ nary Safety Review of the Transrapid Maglev System" in Moving America: New Directions. Akihiro "Beyond the Bullet: Japan's New Train shoots for More Speed. U. "Aerodynamics of Very High Speed Trains and Maglev Vehicles: State of the Art and Future Potential" International Journal of Vehicle De­ sign. USDOT/FRA (July 1991): 38 pages. 36.L. 190. Impact of Aerodynamics on Vehicle Design (1983): 308-341. USDOT/FRA (April 1991): 36 pages.: Crescent Books (1983). [224] Peters.

Mitteilung.A. 775-781.W. [237] Gabrielli. Developers Join Magnetic Rail Push. David "Fast Trains on a Slow Track. 261 (July 1935).[233] High Speed Ground Transportation Systems I: Planning and Engineering. 1-25. and K. I. Jahrgang. Leo "U. Na­ tional Advisory Committee for Aeronautics Seventh Annual Report [NACA TR 116] (1921): 183-184.Theory and Practice. also Ground Effect . Everitt "The Underbody Flow of an EDS-type Ad­ vanced Ground Transport Vehicle. II." Nachrichten von der Koniglichen Gesellschaft der Wissenschaften zu Gottingen . No. Wayson. 1993): 14-17." Mechanical Engi­ neering Vol. [235] O'Connor. Ludwig Applications of Modern Hydrodynamics to Aeronautics. also published in L'Aerotecnica Vol. 1924 (English Translation). United Kingdom: Pergamon Press. ISBN Q-S7262-927-9." Nachrichten von der Koniglichen Gesellschaft der Wissenschaften zu Gottingen . 4.Mathematisch- Physikalische Klasse (1919): 107-137. VI. edited by Murthy V. particularly Part A. Enrico "Teorie ed Esperienze sul Problema dell'Ala in Vicinanza del Suolo. [236] "Magnetschwebebahn: Freie Fahrt fiir den Transrapid?" StromThemen 11. Scuola d'Ingegneria di Pisa Ser. June 1937. 115." In These Times (Jun 14. 1923: 9-16. Nr. Orlando. Late to the Station. . Florida. 451-477. The Aerodynamic Design of Aircraft. G and Theodore von Karman "What Price Speed?" Mechanical Engineering (Oct 1950). [242] Prandtl. II. 15 (Apr 1935): 393-418. also The Induced Drag of Multiplanes. J. [241] Prandtl. [238] Pistolesi. Bondada and Roger L. [243] Prandtl. Oxford." Ergebnisse der Aerodynamischen Versuchsanstalt zu Gottingen. Ludwig "Der Induzierte Widerstand von Mehrdeckern." Pubblicazioni della R. Ludwig "Tragfliigeltheorie. Lieferung. [234] Moberg.P.Mathematisch- Physikalische Klasse (1918). 1978: 295-302. no. [244] Prandtl. [239] Kiichemann. USA. NACA TN 182." Journal of Wind Engineering and Industrial Aerodynamics Vol. Dietrich. April 1994. NACA TM 828. October 25-28. Mitteilung.S. 1992. 8 (Aug 1993): 74-77. Pro­ ceedings of the First International Conference on High Speed Ground Trans­ portation (HSGT) Systems. 8 (1981): 275-294. 40 pages. Ludwig "Tragfliigeltheorie. [240] Howell.

3 (Nov 1935): 40-42. Fukuchi and A. Widnall "Optimum Lift-Drag Ratio for a Ram Wing Tube Vehicle.Yamamoto "Aerodynamics of High Speed Train. 2 (Feb 1968): 1-26. Ludwig Induced Drag of Multiplanes. Edward A. With Special Reference to Tail Moment. 3 (March 1970): 491-497. NASA TN D-4228. and Sheila E. Liejerung. Germany) Vol. 1. Hideo Itokawa and Masuo Taima Further Studies of the Ground Ef­ fect on the Aerodynamic Characteristics of an Aeroplane. Roy H. Timothy M. [250] Ono. Albert "Die Gegenseitige Beeinflussung zweier Tragflachen. Ludwig "Der Induzierte Widerstand von Flugeln mit Endscheiben. [249] Heald. A Moving Belt Ground Plane for Wind-Tunnel Ground Simulation and Results for Two Jet Flap Configurations. 17-18.. Itiro. No. "Comparison of the Ground-Plane and Image Methods for Rep­ resenting Ground Effects in Tests on Vehicle Models. [252] Klemin. "Is 50 Miles Per Gallon possible with Correct Streamlining?" Jour­ nal of the Society of Automotive Engineering (New York) Vol. Thomas R. [253] Stalker. Masami "On the Representation of a Ground in the Wind Tunnel. Alexander "A Belt Method of Representing the Ground. 4 (Oct 1934): 198-199. 5 (17 Oct 1914): 253-258. 45. 32 (May 1933): 177-186. 1967. [254] Tani. Kawaquiti. E. "Aerodynamics of the Railway Train. 32 (Apr 1933): 144-156: also SAE Journal "Vol."' Ergebnisse der Aerodynamischen Versuchsanstalt zu Gottingen. 108 (Aug 1934): 282-286. [247] Betz." Railway Mech. No. 13) of the Aeronautical Research Institute of Tokyo Imperial University (Nov 1937): 117-145. W. [246] Prandtl. [248] Lay." Journal of the Aeronautical Sciences Vol. . 1927. National Advisory Committee for Aeronautics NACA TN 182 (1924). 8. [251] Hara T." Zeitschrift fiir Flugtechnik and Motorluftschiffahrt (Miinchen. Report 158 (Vol." Jour­ nal of the Aeronautical Sciences Vol. [256] Turner. Eng. III." Journal of Research of the National Bureau of Standards Vol. [255] Barrows. 13 (Dec 1934): 863-870. Vol." Monthly Bulletin of the International Railway Congress Associa­ tion Vol. M. No. G. 323 [245] Prandtl." AIAA Journal Vol.

Lucien Ground Effect on Airfoils.J. [268] Malavard. Kronast "Moving-Model Technique used in Automo­ bile Aerodynamics for Measurement of Ground Effects. 1970..N.. 1967. F. 2/3 (1991): 161-166. [267] Papenfuss. 324 [257] Turner. H." Journal of Aircraft Vol.its Interference with the Flow Underneath Cars.N. [258] Carr. Schwartz The Wind Tunnel's Ground Plane Boundary Layer . No.H. [265] Baker.J. NASA TN D-4229. Society of Automotive Engineers SAE 680121. G. W.R. Motor Industry Research Association Report 1966/13 (1966). B.J.J. John W. 5. [266] Kemmerly.. [264] Hucho. 1975. K. [262] Carr. NASA TN D-5935. Correlation of Pressure Measurements in Model and Full Scale Wind Tunnels and on the Road. L. Contract DA 91-591-EUC-1397 (Apr 1961): 30 pages. Pound A Moving Belt Rig for Ground Simulation in Low Speed Wind Tunnels." Journal of Wind Engineering and Industrial Aerodynamics VoL 24 (1986). 6 (June 1988): 557-562. Aerodynamic Characteristics of Vehicle Bodies at Crosswind Conditions in Ground Proximity. Society of Automotive Engineers SAE 750065. Guy T.D. . No. Moy and T. France: Bureau d'Analyse et de Recherche Appliquees. Rheoelectric Analog Study of The­ oretical Flows. the Role of the Lab­ oratory in Aerospace Design. "Train Aerodynamic Forces and Moments from Moving Model Ex­ periments. and M. [263] Beauvais.A. Roser Wind Tunnel Tests of Vehicle Models using a Moving Ground Surface. 11. [259] Grunwald. Turner Problems of Ground Simulation in Automotive Aerodynamics. [260] Butler Sidney F." Experiments in Fluids (Berlin. 1968. [261] Poisson-Quinton. S. G. Paulson Jr..W. and Michael Compton "Exploratory Evaluation of Moving-Model Technique for Measurement of Dynamic Ground Effects.C. Philippe "From Wind Tunnel to Flight.J. No. Janssen and G. 3 (May/Jun 1968): 193-214.W.. Thomas R. Tignor and T. and M. Germany) Vol. Wind Tunnel Investigation of a 3/8-scale Automobile Model over a Moving Belt Ground Plane. 227-251. 1975." Journal of Aircraft Vol. Aeronautical Research Council Report and Memorandum ARC R&M 3451 (1963). 25. C. Society of Automotive Engineers SAE 750066.

Harvey R. 1972. and Pitch. Thomas and Charles Dalton (eds): Aerodynamics of Transportation. [276] McDonald.R.T. 1975. New York: The American Society of Mechanical Engineers. Thomas and Charles Dalton (eds): Aerodynamics of Transportation. [274] Kramer C. U. also Ground Effect on the Take-Off and Landing of Airplanes.: The American Society of Mechanical Engineers. and Lewis W. [278] George. Stein "Wind Tunnel Tests on Underbody Aerodynamics. and G." in Morel. Sovran "Ground Plane Effects on the Aerodynamic Char­ acteristics of Automobile Models. 1973: Paper 17. Masters Rheoelectric Measurements of Some Theoretical Effects of Ground Proximity on Wings. Lucien "Remarque sur les Proprietes Aerodynamiques Theoriques des Profils d'Ailes en Presence du Sol.: The American Society of Mechanical Engineers. Takei and H.T. [275] Hamsten. New York. Society of Automotive Engineers SAE 750064. Washington." (German) Symposium on Industrial Aerodynamics. New York. Paris.A. [270] Chaplin.J. 27 Feb 1961. NACA TM 771. E.. and F. U." presented at Day of Studies on the Ground Effect Machine.. 325 [269] Malavard. Aachen." La Science Aerienne Vol. Fluids Engineering and Bioengineering Conference. "Aerodynamics of Simple Bluff Bodies including effects of Body Shape. Gerhardt. Niagara Falls. Cranfield. 1979: 61-69. Germany. [273] Mason W. D.S. Maurice "L'Influence du Voisinage d'un Sol sur I'Envol et FAtterrissage des Avions. N.M. [277] Morel.C. "A Historical Survey of Automotive Aerodynamics.S. Sakamoto Nissan Full Scale Wind Tunnel .its Application to Passenger Car Design. NSRDC Aero Report 1068. Thomas and Charles Dalton (eds) Aerodynamics of Transportation." in Morel. Society of Automotive Engineers SAE 720100. Ground Proximity. 1979: 71-81. New York. Based on: The Joint ASME-CSME Applied Mechanics. Christensen Correlation Tests in a Climatic Wind Tun­ nel." Advances in Road Vehicle Aerodynamics. 1935 (English Translation). . also listed as: ASME Symposium on Ground Vehicle Aerodynamics 1979. Jager and H. B.A. June 18-20. A. 1974: Paper 3. A. 1979: 248 pages. David Taylor Model Basin.. H. 1979. 3 (Jan-Feb 1934): 60-93. Jan 1964: 13 pages. [271] Le Sueur. [272] Ohtani K. United Kingdom.

and K. Based on: The Winter Annual Meeting of the ASME. 1983: 89 pages. sponsored by The Fluids Engineering Division. 925 (19 May 1938): 178-196.P. ASME. 18. No. "Force and Wake Characteristics of Simple Bluff Bodies in Ground Proximity. Miller (eds): Aerody­ namics of Transportation . "The Lift on a Flat Plate Between Parallel Walls. Miller (eds) Aerodynamics of Transportation - II. A. J." Mitteilungen aus dem Institut fiir Aerodynamik der Eid- genossischen Technischen Hochschule Ziirich No." Proceedings of the Royal Society of London." in Morel. 1979: 83-91. New York: The American Society of Mechanical Engineers. Vol. Thomas and Charles Dalton (eds): Aerodynamics of Trans­ portation. "The Two-Dimensional Aerofoil in a Bounded Stream. C. No.II. New York: The American Society of Mechanical Engineers. New York: The American Society of Mechanical Engineers. Series A Vol. Massachusetts." in Morel. [281] Morel. 132 (2 Jul 1931): 127-152. H. Thomas and James A. [287] Haller. Thomas and James A. L. Thomas and James A. "Luftwiderstand von Eisenbahn-Fahrzeugen. J.J. Series A Vol. E. [288] Hudimoto. [283] Howell.W. "The Effect of Turbulence.P." Memoirs of the College of Engineering. 326 [279] Howell. 5 (1936). [285] Green. New York: The American Society of Mechanical Engineers. 71 (Sep 1947): 167-177. [286] Havelock. "The Lift and Moment on a Flat Plate in a Stream of Finite Width. November 13-18.II. T. 1983: 81-89." in Morel. [282] Volgepohl G. Everitt "Gust Response of a High Speed Train Model.Simulation on the Wind-Induced Loads on Ground Vehicles. . Busuke "The Lift on an Aerofoil With a Circular Arc Section Placed Near the Ground. Kyoto Imperial Uni­ versity Vol. 1983: 1-9." Quarterly Journal of Mathematics (London). Pierre de "La Portance et la Trainee Induite Minimum d'une Aile au Voisinage du Sol (The Lift and Minimum Induced Drag of a Wing in Proxim­ ity to the Ground)." Zeitung des Vereins Deutscher Ingenieure (Feb 1934). [284] Rosenhead. Boston. Miller (eds): Aerodynamics of Transportation . [280] Baker." Proceedings of the Royal Society of London. 1983. 8 (Mar 1934): 36-41. 166. Also listed as: ASME Symposium on Ground Vehicle Aerodynamics 1983.

G. Susumu and Isao Imai "Note on the Lift and Moment of a Circular- Arc Aerofoil Which Touches the Ground With its Trailing Edge. [292] Tani." Proceedings of the Physico-Mathematical Society of Japan (Tokyo) Vol. Ziro Hasimoto and Kaoru Urano "The Forces Acting on an Aerofoil of Approximate ." AIAA JournalWo\. Itiro. 13) of the Aero­ nautical Research Institute of Tokyo Imperial University (Sep 1937): 23-76 (in English). Hermann and Erich Truckenbrodt Aerodynamics of the Airplane. No. Canada) Vol. United Kingdom) Vol. [290] Plotkin. and C. "Aerodynamic Characteristics of Wings in Ground Proximity. Report 156 (Vol. 96 (Aug 1932): 684-689 (in Japanese). Susumu. [294] Tani. Joseph and Allen Plotkin Low-Speed Aerodynamics: From Wing Theory to Panel Methods. Part 3 (Sep 1951): 289-307. [297] Schlichting. Lazar "Numerical Solution of the Equation for a Thin Airfoil in Ground Effect. 11 (1981): 1484-1486. New York: McGraw-Hill (1979). A. [298] Katz. ." AIAA Journal Vol 28. No. [291] Tomotika. [293] Saunders. Itiro "On the Effect of the Ground Upon the Lift of a Monoplane Wing.Joukowski Type in a Stream Bounded by a Plane Wall. New York: McGraw-Hill (1991). [296] Dragos." Journal of the Aeronautical Research Institute (ARI) of Tokyo Imperial Univer­ sity. 19. 12 (Dec 1973): 425-442. 20 (Jan 1938): 15-32. No." Canadian Aeronautics and Space Journal {Ottawa." Zeitschrift fiir Flugwissenschaften Vol." Journal of Applied Mechanics Vol. also Technical Note AIAA 81-4297. Dietrich "Nichtlineare Tragfiiigeltheorie in Bodennahe. 55 (Sep 1988): 735-736." The Quarterly Journal of Mechanics and Applied Mathematics (Oxford. 327 [289] Tomotika. Masuo Taimaand Sodi Simidu The Effect of Ground on the Aerody­ namic Characteristics of a Monoplane Wing. H. [299] Hummel. A. and the report's title as On the Effect of the Ground upon the Lift of a Monoplane Wing. Sometimes the second author's name is reported as M.G. Kennell "Thickness-Induced Lift on a Thin Airfoil in Ground Effect. 11 (June 1965): 185-192. No. 12 (1990): 2132-2134. 21. Tomika. 4. "Thin Ellipse in Ground Effect: Lift Without Circulation. [295] Plotkin.

" 1958 Jahrbuch der Wissenschaftlichen Gesellschaft fiir Luftfahrt (Braunschweig. 1193-1194. Germany): 53-61. C.und Deltaflugeln in Bodennahe. 1-138. 1950. No. Klaus "Berechnung der Aerodynamischen Beiwerte von Tragflugeln endlicher Spannweite in Bodennahe" Abhandlung der Braunschweigischen I'Fzs- senschaftlichen Gesellschaft XII (1960): 95-115.. Gunter "Nichtlineare Tragflachentheorie fiir schiebende und hangende Flugel in Bodennahe (Nonlinear Lifting-Surface Theory for Yawed and Banked Wings in Ground Proximity). and G. McGraw-Hill. [304] Binder. Jack An Introduction to Theoretical and Computational Aerodynamics. [301] Thomas. New York. 1980. No." Zeitschrift fiir Fluqwissenschaften (Braunschweig. [305] Gersten. 24. D. New York: John Wiley & Sons. 4 (Jul/Aug 1977). 7 (July 1986). 2353. and A. Bock "Betrachtungen zu einem Zwei-Wirbel- Flugelmodell." AIAA Journal Vol. Fundamentals of Aerodynamics. "Aerodynamische Eigenschaften von Pfeil. Boeing Document D3- 1861 (Oct 1958). Inc." Progress in Aeronautical Sciences Vol." Zeitschrift fiir Flugwissenschaften und Weltraum- forschung Vol. Wolfgang Computational Fluid Dynamics.H. [309] Kollmann. Plotkin "Lifting-Line Solution for a Symmetrical Thin Wing in Ground Effect. 1979. 241-249. . A Theoretical Method of Determining the Ground Effect on Lift and Pitching Moment for Wings of Arbitrary Planform. Washington: Hemisphere Publishing Corp. Royal Aircraft Establish­ ment RAE Report Aero. 328 [300] Tan." Great Britain Aeronautical Research Council Re­ ports and Memoranda ARC R&M 2884 (1955). Jr. [308] Chow. [303] Kohlman. [307] Hess. "Methods for Calculating the Lift Distribution of Wings (Sub­ sonic Lifting-Surface Theory).. H. U.O. Germany) Vol. John David. 1984. Chuen-Yen An Introduction to Computational Fluid Mechanics.L. [310] Moran. Smith "Calculation of Potential Flow about Arbitrary Bodies. F. and A. J. [306] Multhopp. also. [302] Ackermann. 8 (1967). Oct. 9 (Apr/May 1961): 100-104. [311] Anderson. New York: John Wiley h Sons. 1.M. 1984.

Australia. Germany): 108-125.und /Jawm/a/iri (Braunschweig. Applied Mathematics Department. Masters Thesis. and Gordon L.O. June 1979: 37 pages." Journal of Fluid Mechanics Vol. and Y. 152 pages. Frank Blackman. [317] Davis.. Potential Flow About 2-D Airfoils. Timothy M. 5 (May 1973): 308-312. A Non-Linear Unsteady One-Dimensional Theory for Wings in Extreme Ground Effect Report T7906. Klaus and Jan von der Decken "Aerodynamische Eigenschaften Schlanker Flugel in Bodennahe. New Jersey. [315] Bertin. Joseph P.. (NTIS AD 773 524). [314] Gersten. Inc. Englewood Cliffs. Joseph E. E. [320] Tuck. Part /.S. (NTIS AD 605 207).O. [316] Gray. U. U.. 98. [318| Barrows. 329 [312] Giesing. Joseph P. Part 1 (1980)." 1966 Jahrbuch der Wissenschaftlichen Gesellschaft fiir Luft. nearly identical to Report T7906: 33-47.S.C. 10. Report LB 31946 of the Lone Beach Division of Douglas Aircraft Co. Harris "Nonplanar Wings in Nonplanar Ground Effect. and Michael L. [321] Tuck. No. E. "A Nonlinear Unsteady One-Dimensional Theory for Wings in Ex­ treme Ground Effect. Uni­ versity of Adelaide. Miyai "A Theoretical Note on the Lift Distribution of a Non- Planar Ground Effect Wing. Widnall and Herbert H. also MIT Fluid Dynamics Research Lab Report T 70-3. 24 (Aug 1973): 227- 240. .S. sponsored by The Bureau of Ships Fundamental Hydrome­ chanics Research Program (related to the DTMB). [313] Giesing. Smith Aerodynamics for Engineers. Extension of the Douglas Neumann Program to Problems of Lifting. (Jul 1964): 71 pages. Jr. John J. 1979.A. (May 1968). The Use of Aerodynamic Lift for Support of High Speed Surface Transport Vehicles.: Prentice-Hall. T. (June 1970): 132 pages. Infinite Cascades. Washington.A. Report LB 31653 of the Long Beach Division of Douglas Aircraft Co. Sheila E. D." Journal of Aircraft Vol. Richardson The Use of Aerodynamic Lift for Application to High Speed Ground Transportation. Report FRA-RT-71-56 for the Office of High Speed Ground Transportation (OHSGT) of the Department of Transportation (DOT) of the U. (1971). Ohio State University. [319] Kida.A." Aeronautical Quarterly Vol.

15. (Jun 1987): 11 pages. 26. 330 [322] Tuck. Horace Hydrodynamics. and M. No. 1945. 2 (April 1981): 89-102. [324] Tuck. and S. Hawaii.N. New York: Dover Publications. California.. Takashi Ohta and Kunio Kuwahara Computational Study of Aerodynamic Behavior of a Car Configuration. A.: METS. Inc." AIAA Journal Vol. Bentwich "Sliding Sheets: Lubrication with Comparable Viscous and Inertia Forces.A." Journal of Engineering Mathematics Vol. Honolulu. paper presented at the AIAA 19th Fluid Dynamics and Lasers Conference. "Steady Flow and Static Stability of Airfoils in Extreme Ground Effect. June 8-10. Agarwal Euler Calculation for Flow over a Wing in Ground Effect.S. [328] Lamb. Fundamentals of Aircraft Design. [325] Deese." Journal of Fluid Mechanics Vol. E. sponsored by McDonnell Douglas (1986): 5 pages. "Analysis of Small-Aspect-Ratio Lifting Surfaces in Ground Ef­ fect.S.E. [326] Plotkin. . Dodbele "Slender Wing in Ground Effect. 1984 (revised from the 1975 edition). J. [329] Nicolai. [327] Hashiguchi. 4 (April 1988): 493-494.K." Journal of Fluid Mechanics Vo\. Masanori.S.A.O. 117 (1982): 305-314. No. [323] Newman. E. 1987. 135 (1983): 51-69. J. AIAA-87-1386.O. and R. San Jose. U. Leland M. AIAA-86-1765. U.