You are on page 1of 7

Aerospace Science and Technology 14 (2010) 557563

Contents lists available at ScienceDirect

Aerospace Science and Technology


www.elsevier.com/locate/aescte

Inatable kites using the concept of Tensairity


Joep C.M. Breuer, Rolf H. Luchsinger
Empa, Swiss Federal Laboratories for Materials Testing and Research, Center for Synergetic Structures, Ueberlandstrasse 129, CH-8600 Duebendorf, Switzerland

a r t i c l e

i n f o

a b s t r a c t
Inatable structures have very interesting properties such as low weight, compact transport volume and easy set up. Both manned and unmanned aircraft have been built using inatable wings. However, a severe drawback of inatable wings is the small load-bearing capacity of such structures which limits the aspect ratio of this wing type. Introducing the structural concept of Tensairity overcomes this deciency. Tensairity combines an inatable structure with struts and cables and thus increases the stiffness and maximal load of the inatable structure tremendously. A further improvement of the stiffness and ultimate load of Tensairity can be achieved by introduction of fabric webs into the airbeam. In this work, the concept of web-Tensairity was further developed into curved girders to be able to build wings with dihedral, sweep and twist. A comparison between a curved and a straight web-Tensairity girder proved that their load deection behaviour was very similar and superior to a curved and straight airbeam with the same dimensions and internal pressure. The stiffness of the web-Tensairity girder was estimated analytically. To demonstrate the concept of Tensairity wings, the goal was set to build a Tensairity kite which ies stable on a single tether. The latest kite prototype has a span of almost 8 m and a projected area of 11 m2 . Results of structural and aerodynamic tests of this kite are presented. 2010 Elsevier Masson SAS. All rights reserved.

Article history: Received 18 November 2009 Received in revised form 19 February 2010 Accepted 21 April 2010 Available online 24 April 2010 Keywords: Inatable wings Tensairity Kites Fabric structures

1. Introduction Inatable structures have very interesting properties such as low weight, compact transport volume and easy set up. They have found applications e.g. as boats or in emergency tents. Using inatable structures for wings was already suggested in a patent from 1933 by McDaniel [15] which introduced inatable spars for wings. There were two interesting developments in manned ight using inatable wings during the 1950s, namely the Inatoplane from Goodyear, which was further developed till the 1970s [15,4] and the British M.L. Utility Mk1, which was a tailless design with a symmetrical airfoil of 24% thickness [1]. Another example is the NASA Paresev 1C of 1963, a Rogallo wing with inatable beams, which was however not very stable in ight [17]. More recently, from 19902000, the Swiss company prospective concepts developed a series of aircrafts with inatable wings mainly as technology demonstrators. The tailless Stingray had webs in chord direction and a very thick prole allowing for a very low internal air pressure of 0.020.05 105 N m2 . Slender wings as the so-called Pneuwing were later realized by the company with

Corresponding author. Tel.: +41 44 823 4090; fax: +41 44 823 4211. E-mail addresses: joep.breuer@gmail.com (J.C.M. Breuer), rolf.luchsinger@empa.ch (R.H. Luchsinger). URLs: http://www.empa.ch (J.C.M. Breuer), http://www.empa.ch (R.H. Luchsinger). 1270-9638/$ see front matter doi:10.1016/j.ast.2010.04.009

webs in span direction and a higher internal pressure in the order of 0.7 105 N m2 [12]. Lately, the interest in inatable wings has shifted to unmanned aircraft with examples as the NASA Dryden I2000 inatable wing and the prototypes of the University of Kentucky in cooperation with ILC Dover [6,5,9], who also made the early example of the Apteron. Surf kites (tube kites) can also be seen as inatable wings. Compared to the Paresev, their design is today quite rened. The development of these kites is advancing quickly; however, the major focus is on the handling of the kite and not on the primary structure. In all these efforts to develop inatable wings the temporary aspect (compact transport, fast set up) coupled with the low weight were major driving forces. In some applications, the crash resistance of inatable structures and the ability to oat on water are further important aspects. On the other hand, a severe drawback of inatable wings is their very limited load bearing capacity. Inatable wings are much more exible than conventional wings. As a result, high air pressure, thick wing proles, low aspect ratio or a combination of these factors need to be used in their design limiting the aerodynamic and structural performance considerably. It is the purpose of this work to overcome this deciency by introducing the structural concept of Tensairity [13] in wing structures. Tensairity combines an inatable structure with struts and cables to increase the stiffness and maximal load of the inatable structure tremendously. First successful implementations of Tensairity have been realized in civil engineering applications

2010

Elsevier Masson SAS. All rights reserved.

558

J.C.M. Breuer, R.H. Luchsinger / Aerospace Science and Technology 14 (2010) 557563

such as bridges and roof structures [18]. Some initial work on the use of Tensairity in wings was done recently [3]. In this paper Tensairity kites are investigated. Currently the scientic interest in (inatable) kites is increasing for their potential to harness wind power for energy production or ship propulsion [10,16,8,7,22]. The aerodynamic eciency of these kites is crucial and Tensairity might allow for signicant improvements compared to the current kites. In Section 2, the Tensairity concept is explained. A major focus is set on web-Tensairity which is important for the applications in wing structures. The evolution of the design of an 11 m2 Tensairity kite as well as some tests and estimations of the performance are presented in Section 3. Flight tests of the 11 m2 kite are reported in Section 4 and conclusions are drawn in Section 5. 2. Tensairity The basic idea of Tensairity is to combine an airbeam with conventional cables and struts to improve the load bearing capacity of inatable structures. A basic Tensairity beam consists of a compression element, a low pressure airbeam which is tightly connected to the compression element and two tension elements which run from end to end of the compression element in a spiral way around the airbeam (Fig. 1). The low pressure airbeam (typical pressures for Tensairity structures are in order of 0.1 105 N/m2 ) pretensions the cables and stabilizes the compression element against buckling. Research has been done on the static response of spindle shaped Tensairity girders under bending loads by analytical, experimental and numerical methods [11]. Recently, Tensairity columns were investigated [19,21]. In this case three compression elements are equally spaced around a spindle-shaped airbeam. A further improvement of the stiffness and ultimate load of Tensairity is expected by introduction of fabric webs into the airbeam. The web will give more support to the compression element and has the ability to transfer shear forces. A cross-section of a plain Tensairity girder and a web-Tensairity girder can be seen in Fig. 2. So far, webs have been applied in Tensairity for wing structures [3] and columns [21]. 2.1. Evolution of web-Tensairity girders The rst Tensairity wing prototypes were built with symmetric, spindle-shaped webs as shown in Fig. 3a. The web, the compression element (thick line) and the tension element (thin line) are given in the gure. This approach limited the design to straight wings [3]. To be able to build a Tensairity kite which ies stable on a single line, the shape of the kite had to be changed resulting in severe modications of the web-Tensairity concept. The rst improvement to create more design freedom with Tensairity was to integrate the tension element in the web (Fig. 3b). A logical second step was then to integrate the compression element in the web as well (Fig. 3c). This allowed to give the girder a slight curvature and thus results in a slight dihedral in case of a wing. But the dihedral was too limited to be effective in wings. Furthermore, the lateral support of the compression element by the fabric decreases when it is positioned in the web. Therefore, the compression element needs to be curved to obtain a higher curvature of the girder (Fig. 3d). The hull has a curved shape and ination of the hull bends the initial straight compression element, which itself needs to have a small bending stiffness. The layout of Fig. 3d was further improved to be able to build wings with sweep and twist. It must be said that the amount of dihedral is limited by the allowable initial curvature of the compression element. Sweep and twist are also limited because otherwise the web is not properly aligned with the main load direction. Another approach to generate dihedral is shown in Fig. 3e. However, experimental investigations

Fig. 1. Basic setup of a cylindrical Tensairity beam.

Fig. 2. Cross-section of a standard spindle shaped Tensairity girder (left) and a spindle shaped web-Tensairity girder. The black dots indicate the tension and compression element.

Fig. 3. Different web layouts for web-Tensairity girders.

showed that the connection of the two parts at the central region is very unstable and prone to buckling even with additional stiffeners. Thus, the concept of Fig. 3e was not followed up in more detail. 2.2. Straight against curved compression element in Tensairity To evaluate the difference in the load bearing behaviour of a straight and a curved compression element in a Tensairity girder (Fig. 3 b and d), two similar models were made. Both models had a span of 1.96 m and a web height of 0.04 m at the tips and 0.07 m at the centre. A solid 4 mm carbon rod was used as a compression element and the tension element was composed of two 170 kg Dyneema lines. The models were made of Icarex rip-stop polyester kite fabric with PU bladders. The total weight of the models was about 320 g each, of which about 50 g was due to the tension and compression element. The radius of curvature of the bended compression element was 2.36 m leading to an angle between the centre and the tip of 12 degrees. To mimic the load distribution of a wing structure the specimens were suspended on a whippletree system with 8 evenly spaced attachment points (Fig. 4). The load consisting of lead weights weighing between 0.25 kg and 2.5 kg was introduced at the centre. The deection was measured at 5 positions along the span. Internal air pressure values of 0.05, 0.1 and 0.15 105 N m2 (50, 100 and 150 mbar) were applied. The measured load-deection curves for a pressure of 0.1 105 N m2

J.C.M. Breuer, R.H. Luchsinger / Aerospace Science and Technology 14 (2010) 557563

559

Fig. 6. Set up for the analytical model to estimate the deformation of the webTensairity beam.

Fig. 4. Bending test setup with the curved web-Tensairity model.

specimens. For the airbeam the initial stiffness is reduced at a load in the order of 20 N, where wrinkles start to appear which grow to a single big fold at the ultimate load of about 30 N to 40 N. The load-deection behaviour of the curved and straight airbeam is very similar. In relation to the airbeam, both the straight and the curved Tensairity specimen have an initial stiffness which is about a factor 3 higher for the same internal pressure, while the ultimate load is about a factor 4 higher. This is an impressive result considering that the tension and compression element add only 16% to the total weight. And the curved Tensairity girder performs as good as the straight Tensairity girder. This gives condence for the application of web-Tensairity in wings with dihedral. An analytical estimate of the deection of a straight webTensairity girder can be given [3]. The model relies on beam theory and assumes that the relative distance between the compression and the tension element is constant under the bending load which is reasonable for high enough pressure values. The experimental situation (Fig. 4) can be described as a cantilever with a homogeneous distributed load (Fig. 6). A straight web-Tensairity girder is considered. The differential equation

d2 y dx2

M (x) E I (x) q 2

(1)

needs to be solved with

M (x) =
Fig. 5. Load-deection behaviour of straight and curved web-Tensairity beams, straight and curved airbeams (no Tensairity) and an analytical estimation for the web-Tensairity beams for an air pressure of 0.1 105 N m2 .

(l x)2

(2)

and

are shown in Fig. 5. The deection is given as the difference between the centre deection and the average of the deection of the two outmost positions (at 1/16 and 15/16 of the span). An initial weight of 11 N was applied to tension the lines of the whippletree system. All tests were done 3 times to check the repeatability of the results and the average value of the 3 tests is given. The absolute deviations from the measurements to this average value had a maximum of 15%. These large deviations were measured during initial deection and buckling. In the linear part of the curves the variation was not more than 5%. To compare the performance of web-Tensairity with a simple airbeam the specimens were also tested with removed compression element as shown in Fig. 5. The overall behaviour of the straight and curved Tensairity girder is very similar. The straight and the curved web-Tensairity specimen have about the same stiffness but the ultimate load of the curved Tensairity girder is higher. For both specimens it was observed that as soon as the stiffness starts to decrease wrinkles start to build up in the upper side of the hull at the centre of the girder. At the ultimate load, the compression element buckles sideways in the central part causing a rotation of the girder. This behaviour has been observed for both the straight and the curved

with q the distributed load, l half of the span of the beam, E c A c , E t A t the product of the Youngs modulus and cross sectional area of the compression and tension element, respectively and h the vertical distance between the compression and tension element. Assuming a linear variation of the distance along the span

E I (x) = E A ct h(x)2 E c A c E t At E A ct = E c A c + E t At

(3) (4)

h(x) =

h0 l

(l x)

(5)

the right side of Eq. (1) becomes independent of x and Eq. (1) can be easily integrated. One obtains a parabolic deformation with the deection for 0 x L /2 given by

y (x) =

(q

where the span L = 2 l , the slenderness = L /h0 and the initial load F 0 have been introduced. The analytical estimate for = 28, E c A c = 1.49 106 N, E t A t = 5.32 104 N and F 0 = 11 N at x = 7 L /16 as a function of the total applied load is shown in Fig. 5. Despite the various simplications, the initial stiffness of the webTensairity girder is very well predicted by this analytical model.

16 E A ct

F0 ) L

x2

(6)

560

J.C.M. Breuer, R.H. Luchsinger / Aerospace Science and Technology 14 (2010) 557563

Fig. 8. The three views of the Tensairity kite with 11 m2 surface area.

Fig. 7. Development of the design of the Tensairity kite (views from the top with the same scale).

3. Tensairity kites The rst built Tensairity kites had a rectangular shape and no dihedral which led to a very poor aerodynamic stability [3]. The main goal of this initial project was to realize a Tensairity wing with a well dened airfoil. To improve the concept of Tensairity wings further the goal was set to build a Tensairity kite which ies stable on a single tether. To this end, the Tensairity concept needed to be adapted to more complex wing shapes. 3.1. Kites development The evolution of the Tensairity kites both in shape and size is shown in Fig. 7. The rst kites with a rectangular plan-form (Fig. 7a) had webs as shown in Fig. 3 a and b. As a rst step towards more complex shapes a structural model was made with the web layout of Fig. 3c and an elliptical plan-form (Fig. 7b). To decrease the weight the aft part of the airfoil of this model had a single skin which was tensioned by 5 thin carbon rods which were shaped and supported by inatable struts for most of their length. The skin in surf kites is also tensioned in a similar way by inated struts. Although this construction method saves weight and reduces the complexity of the kite, it also reduces the accuracy of the airfoil shape. In order to have an accurate shape of the nose, an inated structure with three webs was used in the model of Fig. 7b while the nose is given by a single air tube in a surf kite. Several attempts were made to y this model. A variety of xed control surfaces was added. However, the ight stability proved to be very poor. A search for a single line stable kite with a wing-like lifting surface with a reasonable aspect ratio resulted in the Pfeil Ente [2]. This design was modied with Tensairity wing spars as shown in Fig. 3 d and e (Fig. 7c). The resulting kites proved to be very stable with a range of different canard wings. Then followed a couple of canard kites with increasing size and complexity of wing shape using more and more air-supported structural members (Fig. 7d) while the number of tubes for the main spar was reduced to 2. Although this probably decreases the aerodynamic eciency, the reduction in overall complexity, weight, construction effort and the inated volume was found more important at this stage of the development. With increasing size, the Tensairity nose (fuselage) (see Fig. 7d) supporting the canard wing proved to be too exible, especially in torsion. A carbon rod was tried as

Fig. 9. Tensairity kite with 11 m2 surface area before launch during testing.

a nose, but it was too vulnerable. Also the long nose and the canard wing made ground-handling of the kite dicult. The decision was made to abandon the canard conguration and experiments were made with different tail congurations. In the end the centre section of the kite was elongated to the back and a reex was added to this part to have some extra pitch stability (Fig. 7e). In ight the central section of the sail bulges which generates a keel for directional stability (similar to a Rogallo wing). In this way the functions of the canard wing were integrated in the ying wing design. 3.2. 11 m2 prototype The latest Tensairity kite prototype has a span of 7.8 m and a projected area of 11 m2 (Figs. 8 and 9). It is quite slender for a kite with a projected aspect ratio of about 5.5. The bridle is only connected at two points to the centre strut. With a minimal weight of about 2.5 kg it is designed to generate an ultimate lift of 1000 N. The structure consists of one main Tensairity spar with a web in span-wise direction to take up the bending load and several Tensairity struts in chord direction to tension the lifting surface. The centre strut transfers the lift load to the bridle. The dihedral of the wing is 15 which is low for a kite but high for an airplane. The sweep of the main spar is 8 . The effective sweep is almost neutral due to the large chord in the centre section of the wing. The different airfoils also give the wing a small effective twist. All airfoils are based on the NACA 4-series as they are suitable for turbulent ow. The reex of the centre airfoil is achieved by adding an extra section to the existing airfoil of the canard kite (Fig. 7d). The airfoils have a thickness of 14% to be able to have large diameter air tubes and a high web to reduce the forces in the tension and compression element. Due to the low internal pressure in the order of 0.050.1 105 N m2 and the low wing loading very lightweight

J.C.M. Breuer, R.H. Luchsinger / Aerospace Science and Technology 14 (2010) 557563

561

Fig. 11. Schematic representation of the forces at the kite (top) and at the ground attachment point of the line (bottom).

Fig. 10. Load-deection behaviour of the Tensairity kite with 11 m2 surface area for an air pressure of 0.07 105 N m2 .

The wind speed was ideally almost zero during these tests. The following parameters were measured during these tests: line angle at the tow car, line force at the car, and speed and direction of the airow relative to the moving car. From these ight tests an estimate of the effective lift coecient and lift/drag could be made. Due to the short length of the line, its weight can be neglected and the line is almost straight [20]. The drag of the line D l , which is small compared to the drag of the kite, can thus be approximated by:

materials can be used for the kite. The skin of the pneumatic structure (tubes and struts) is made of Icarex, a rip-stop polyester kite fabric (34 g m2 ) and the lifting surface of the kite (sail) is made of Cuben Fiber, a Dyneema bre reinforced polyester lm (21 g m2 ). A lighter variant of Cuben Fiber (11 g m2 ) is used as reinforcement for the bigger diameter tubes. Heptax, an aluminium coated laminate designed for helium balloons (26 g m2 ) is used for the bladders while carbon bre pultrusions (8 mm diameter tubes with 1 mm wall thickness for the main spar and centre strut and 4 mm diameter rods for the other struts) are used as compression elements. The tension elements are composed of two 500 kg Dyneema lines. The tips of the wing and the front of the nose are unsupported by Tensairity to have a soft zone in case of a rough landing. The structure was subjected to a load test to conrm the structural reliability. The set up of the test was very similar to the one shown in Fig. 4. The wing was suspended on 8 unevenly spaced positions by a whippletree system, which approximated the loads of an elliptical lift distribution. The load was applied at the bridle at the centre with lead weights. The load deection curve is shown in Fig. 10 for an internal pressure of 0.07 105 N m2 . The deection is given as the difference between the deection of the centre and the tips. The test was stopped at a load of 700 N to prevent damage of the kite. The load deection curve is in this load range very linear as expected from the tests on the straight and curved web-Tensairity models (Fig. 5). The tip deection at 700 N is about 750 mm. The load-deection curve starts at an initial load of 100 N. This initial load was necessary to tension all lines of the whippletree system. The analytical model (Eq. (6)) predicts a deformation at 700 N of 510 mm, with = 27.2, E c A c = 2.75 106 N, E t A t = 9.75 104 N, F 0 = 100 N and L /2 = x = 3.525 m. Although the model is not very suitable in case of the kite, which has an elliptical lift distribution and a much more complex shape (main Tensairity girder has sweep and twist) it still provides a rst estimate of the stiffness and deection. 4. Flight tests To investigate the ight performance of the kite, tow test were made on two airstrips, the military airstrip in Dbendorf and the former military airstrip in St. Stephan, both in Switzerland. The kite was xed with a line of 54 m length and a diameter of 1.3 mm on a car and towed on the airstrip for a distance up to 1000 m.

D l C Dl

1 2

V 2 d sin 1 l

(7)

with C Dl the drag coecient of the line, the density of the air, V the velocity of the air, d the diameter of the line, 1 the line angle at ground and l the length of the line. The drag of the kite D k and the angle at the top of the line 0 are given by (Fig. 11)

D k = cos 1 F D l

(8) (9)

0 = cos1

Dk F

with F being the force on the line. The lift of the kite L k is dened as

where W is the weight of the kite. The ratio of lift over drag of the kite L / D is given by

L k = sin 0 F + W

(10)

L/ D =

Lk Dk

(11)

With C Dl = 1.2 and d, l, W , given, the measurement of F , V and 1 allows to determine D k , L k and nally L / D . The measured data for a typical test run are shown in Fig. 12. In this ight test the weight of the kite was increased to 3.6 kg due to additional measurement equipment and extra battens between the struts to improve the shape of the sail. Rather steady ight conditions were obtained between 60 and 90 s with a velocity of about 7 m/s and a line load of about 80 N. During this ight, the angle of attack was optimized in order to make the kite y as high as possible. The line angle 1 reaches values up to about 71 demonstrating the potential of the kite for high altitude ights. Given the projected wing area S of the kite, the drag coecient of the kite (C Dk ) and the lift coecient of the kite ( C Lk ) can then be calculated by means of

C Dk = C Lk =

Dk
1 2 1 2

V 2 S
Lk

(12) (13)

The values of L / D , C Dk and C Lk are plotted in Fig. 13 for the period between 60 and 90 s of the test shown in Fig. 12. The L / D ratio

V 2 S

562

J.C.M. Breuer, R.H. Luchsinger / Aerospace Science and Technology 14 (2010) 557563

Fig. 12. Experimental data of a tow test: the line load, the velocity and the line angle at the ground attachment point are given as a function of time.

ities of more than 15 m/s (plastic deformation of the aluminium connection pieces of the spars and aerodynamic instability), while the Tensairity kite had no problems at such speed with similar line loads. For the Tensairity kite better values of the L / D are expected if more attention is given to the sail design and the trailing edge tension. This will improve the effective lift coecient and the L / D . For a relatively simple single skin sail wing like the Tensairity kite it should be possible to reach L / D values of 10 while double skin sail wings might reach 20 or higher [14]. In separate tests, the 11 m2 Tensairity kite was own in the wind with an anchored winch. These tests demonstrated the goodnatured behaviour of the kite. It was own up to 2900 m (970 m over ground) near Axalp in the military airspace of Meiringen, Switzerland on September 4th, 2009. During these ights a GPS measured the altitude and a relieve valve together with a small pump regulated the pressure of the Tensairity structure. 5. Conclusions
Fig. 13. Calculated coecients from the experimental data of Fig. 12: Lift over drag L / D , drag coecient of the kite C Dk and lift coecient of the kite C Lk are given as a function of time.

value is about 5 in the period of steady ight. This is a bit lower than expected as an L / D value of 10 was reported in a comparable sail wing [14]. Probably, the difference is due to the lack of tension in the trailing edge of the sail, which bulges between the struts, reduces the local angle of attack and thus generates almost no lift. The average C Lk was about 0.35 while about 0.6 was expected [14]. But the L / D value of the Tensairity kite is still much better than a commercial Delta-kite we have tested. This kite with 19ft span and a weight of 2.9 kg from Premier Kites reached L / D values of about 2.53. The Delta kite could also not withstand wind veloc-

Tensairity is a new light-weight structural concept. The synergetic combination of an airbeam with cables and struts increases the stiffness and buckling load of an inatable structure with the same internal pressure by a factors 3 and 4 respectively by adding only 16% to the mass of the airbeam. The Tensairity concept can be directly adapted to inatable wing structures. To this end, curved web-Tensairity girders were investigated allowing for sweep, dihedral and twist of the wing. Experimental investigations show that a small curvature of the web-Tensairity girder does not deteriorate the performance compared to straight girders. As a rst application, a Tensairity kite with 8 m span ying stable on a single line was built and tested. The Tensairity kite performed better than a commercially available Delta kite. We expect that there is still a big potential to further enhance the performance of the Tensairity kite by an improved tailoring of the wing. The aerodynamic

J.C.M. Breuer, R.H. Luchsinger / Aerospace Science and Technology 14 (2010) 557563

563

and structural performance, the low weight, the crash resistance together with the small storage volume make the Tensairity wing not only interesting for kite applications but also for UAVs, hang gliders and ultra light aircraft. Acknowledgements The authors would like to thank Thomas Horvath for his insights into the design of kites. The nancial support of Festo is also gratefully acknowledged. References
[1] Airoplane, the Inatable-wing M.L. Utility Mk1 demonstrated, Flight Magazine (1957) 751752. [2] R. Beutnagel, Pfeil-Ente, Sport & Design Drachen 8 (2001). [3] J.C.M. Breuer, W.J. Ockels, R.H. Luchsinger, An inatable wing using the principle of Tensairity, in: 48th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, Honolulu, Hawaii, AIAA-2007-2117, 2007. [4] G. Brown, R. Haggard, B. Norton, Inatable structures for deployable wings, Vertigo, Inc., AIAA-2001-2068, 2001. [5] D.P. Cadogan, S.E. Scarborough, D. Gleeson, A. Dixit, Recent development and testing of inatable wings, in: 47th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference and 14th AIAA/ASME/AHS Adaptive Structures Conference, Newport, Rhode Island, AIAA-2006-2139, May 14, 2006. [6] D. Cadogan, T. Smith, R. Lee, S. Scarborough, D. Graziosi, Inatable and rigidizable wing components for unmanned aerial vehicles, in: 44th AIAA/ ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, Norfolk, VA, AIAA-2003-6630, April 2003. [7] M. Canale, L. Fagiano, M. Milanese, Power kites for wind energy production, IEEE Control Systems Magazine 27 (Dec 2007) 2538. [8] B. Houska, M. Diehl, Optimal control of towing kites, in: 2006 45th IEEE Conference on Decision and Control, San Diego, CA, 2006, pp. 26932697.

[9] J.D. Jacob, S.W. Smith, Design limitations of deployable wings for small low altitude UAVs, in: 47th AIAA Aerospace Sciences Meeting Including The New Horizons Forum and Aerospace Exposition, Orlando, Florida, AIAA 2009-1291, 2009. [10] M.L. Loyd, Crosswind kite power, Journal of Energy 4 (3) (1980) 106111. [11] R.H. Luchsinger, R. Crettol, Experimental and numerical study of spindle shaped Tensairity girders, International Journal of Space Structures 21 (3) (2006) 119 130. [12] R.H. Luchsinger, M. Pedretti, A. Reinhart, Pressure induced stability: From pneumatic structures to Tensairity, Journal of Bionics Engineering 1 (3) (2004) 141 148. [13] R.H. Luchsinger, A. Pedretti, P. Steingruber, M. Pedretti, The new structural concept Tensairity: Basic principles, in: Progress in Structural Engineering, Mechanics and Computations, A.A. Balkema Publishers, London, 2004. [14] M.D. Maughmer, A comparison of the aerodynamic characteristics of eight sailwing airfoil sections, NASA Langley Res. Center The Sci. and Technol. of Low Speed and Motorless Flight, 1979, Pt. 1, pp. 155176. [15] R.K. Norris, W.J. Pulliam, Historical perspective on inatable wing structures, in: 50th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, Palm Springs, California, AIAA-2009-2145, 2009. [16] W.J. Ockels, Laddermill, a novel concept to exploit the energy in the airspace, Aircraft Design 4 (2) (2001) 8197. [17] Paresev, the Paraglider Research Vehicle, http://www.nasa.gov/centers/dryden/ about/Organizations/Technology/Facts/TF-2004-09-DFRC.html. [18] M. Pedretti, R. Luscher, Tensairity-Patent Eine pneumatische Tenso-Struktur, Stahlbau 76 (5) (2007) 314319. [19] T.S. Plagianakos, U. Teutsch, R. Crettol, R.H. Luchsinger, Static response of a spindle-shaped Tensairity column to axial compression, Engineering Structures 31 (8) (2009) 18221831. [20] S.L. Veldman, C.A. Vermeeren, H.E.N. Bersee, O.K Bergsma, Conceptual design of a high altitude kite, AIAA-2002-1735, 2002. [21] T.E. Wever, T.S. Plagianakos, R.H. Luchsinger, P. Marti, Effect of fabric webs on the static response of spindle-shaped Tensairity columns, Journal of Structural Engineering 136 (4) (2010) 410418. [22] P. Williams, B. Lansdorp, W.J. Ockels, Nonlinear control and estimation of a tethered kite in changing wind conditions, Journal of Guidance, Control, and Dynamics 31 (3) (2008) 793798.

You might also like