Biomimetic Glider Wing Design for Future Transport

Summary

Flying vehicles are the key to many traffic problems such as congestions, due to their ability to travel at different heights. This would allow for better land use. As such, their needs are becoming increasingly pressing today. This study contributes to this field by identifying structures that are ideal for future transport. It focuses on biomimicry, looking to nature's best flyers and gliders and comparing their flight performances. The ten selected species were 3D modeled and analyzed by both XFOIL method and high-fidelity ANSYS Fluent simulation, named Computational Fluid Dynamics (CFD). Among the ten species, the best two were 3D printed and tested in an open return wind tunnel. The CFD results were compared with wind tunnel results. Scanning Electron Microscope (SEM) was used to validate CFD and wind tunnel data. Detailed analyzes suggested that Javan Cucumber was superior in flight performance. Then, the model was refined in XFLR5 and applied to a full-scale model which was further analyzed in ANSYS Fluent. The comparison between Javan Cucumber and current commercially available powered glider, XT912 Tundra-Merlin, showed the advantage of Javan Cucumber and its potential to be further studied. In addition, the Javan Cucumber model was prototyped by Computer Numerical Control (CNC) and installed on a Micro Air Vehicle (MAV) platform. Flights of this prototype suggested excellent flying capability. Lastly, Javan Cucumber was optimized using aerodynamic theories, and its CFD results suggested improved performance. A prototype of this optimized model was constructed and found to be significantly improved in its performances.

Question / Proposal

This research seeks to find a better design for powered gliders using biomimicry as well as CFD optimization. These wing designs would have the best aerodynamic performance on gliders by a broad comparison and selection across all flying or gliding species. It is anticipated that this research will go beyond theoretical conception and make great practical contributions to the development of personal flying vehicles for future transport. A trip to Lee Kong Chian Natural History Museum further made us more confident on the pursuit of biomimicry. (Picture below)

Among all the species to be studied, we hypothesized that the Javan Cucumber would perform best. The Javan Cucumber features elliptical wings, which was seen in the advanced Spitfire plane during the Second World War. According to Model Aircraft Aerodynamics written by Martin Simons, the elliptical wing produces a constant downwash at any speed with the minimum induced drag for a given lift.

For optimization, we hypothesized that the Javan Cucumber wing design, optimized with a smaller sweep angle and shorter chord length, will have a much higher aerodynamic performance.

Research

Personal flying vehicles have long been a staple of science fiction and human imagination. In reality, companies such as Uber are already working on flying cars which are set to be completed by 2020. However, apart from flying cars, there are also other flying vehicles that are currently already in use but which can be further improved. One such vehicle is the powered glider, which requires less distance to take off and land, can carry 1 to 2 people and use gliding as its primary power source. Currently, it is widely used as a sports activity, especially in the United States. Although it has been used for many years, its overall structure has remained mostly unchanged. Thus, there is room for improvement for future transportation. Biomimicry has been long applied in aircraft design; however, these aircraft usually have a tiny load capacity, or there has not been a broad comparison across species to discover which offers the best structure for flight. Hence, this study will use biomimicry, with the comparison across a wide range of species, to find a wind design suitable for future transportation.

 

In order to compare the aerodynamic performances of the selected biomimetic species, lift coefficient (CL) and drag coefficient (CD) was used to determine the aerodynamic performances of an aircraft.

In glider designing, engineers mainly regard the high CL/CD  as the criteria for better models. Considering the carrying ability of planes, the commercial aircraft industries usually place the lift as a major indicator of aircraft performance. In this research, the new powered gliders were designed for transportation. This use would require flights with high efficiency and a capacity to glide for a long distance with heavy loads, coinciding with the glider designers’ and commercial aircraft designers’ considerations.

For both high efficiency and long distance, the models should have a high CL/CD under a large range of Angle of Attacks (AoA). A higher CL/CD will result in a larger distance/height ratio, which suggests that the model can travel a longer distance when gliding down the same height. This model is thus more efficient. However, it is not sufficient to have a high CL/CD ratio at a single AoA, as the actual glide angle is subject to frequent changes of wind, temperature and pressure. Thus, the best performing species should have large CL/CD ratios in a broad range of AoAs.

To carry high loads, the glider simply needs to generate a relatively large lift under its working conditions, generally between AoA of 3 to 8 degrees. If the biomimetic designs can generate a large lift within or even beyond this range, its actual carrying ability can be considered as good.

Overall, the prioritized criterion of our selection for best models is the high CL/CD in a broad range of AoAs.  Being able to sustain a higher CL/CD in any AoA, the model is more efficient and capable of gliding further. Its performance will be even better if the lift generated by the model is large. 

 

 

 

Method / Testing and Redesign

1) Modeling 

We firstly used Autodesk CAD to sketch the planforms for all species, then we found the coordinates of the turning points and used XFLR5 to construct the wing using MH 60 airfoil.

We also reconstructed our models in an industrial-level 3D modelling software – the CATIA V5.

2) CFD simulation

We then used the Horseshoe Vortex Method in XFLR5 to conduct the simulation.

However, some experts then reminded us that XFLR5 is not widely used in industry due to its limitations. Hence, we proceeded to use a well-received software, ANSYS Fluent.

We used ICEM CFD to mesh the model and created the large enclosure using CH-Grid Topology.

After setting up everything, we conducted the Fluent simulation.

3) Manufacturing 

In order to validate our CFD results, we manufactured models for real-life testing.

We used 3D-printing to produce our models using PLA. Then, we polished our models carefully to reduce the surface roughness.

We also used CNC method with EPP material and Frame method with PP material to produce our prototype.

 

4) Wind tunnel experiment 

We 3D-printed our best two biomimetic models. They were half-models and set up vertically. The rotational plate below can change the angle of attack. The lift and drag measured by the sensors were displayed out.

We tested our models in the wind tunnel and collected the data

5) Structure testing

There was discrepancy observed between CFD and wind tunnel results (refer to RESULTS). It is necessary to account for the discrepancy.

We suspected that the skin friction was a reason for the discrepancy. In our CFD simulation, the surface of our model was assumed to be absolutely smooth. But in real life, our wind tunnel models may have relatively rough surfaces which will produce extra skin friction. Hence, we want to estimate the skin friction and investigate how much it contributed to the discrepancy. We then used a Scanning Electron Microscope to investigate the structure of our models. 

6) Prototype

We used CNC method to produce our Javan Cucumber wing with EPP foam. It was then installed on a model plane. It has ailerons which allow the plane to rotate for better maneuverability. We then tested it indoor.

7) Optimization

We mathematically modeled the shape of Javan Cucumber. The leading edge of the wing was abstracted as a segment of an ellipse and the chord length distribution is an ellipse as well. To change the shape of the wing, we changed the eccentricity of two ellipses. Therefore we had 30 distinct combinations, each one worth investigating.

Following the same testing methods mentioned before, we found out their performance by using XFLR5 and ANSYS Fluent. 

We then moved to construct the new prototype. We used the wing rib structure. We drew the planform and ribs in Autodesk CAD and cut the KT foam boards by a laser cutter. Later, ribs were assembled on the planform and coated with a layer of plastic film. The new prototype was done and even much lighter than the previous one.

 

 

 

 

 

 

 

Results

1) XFLR5

Through XFLR5 VLM analysis, a set of CL/CD data for our 10 tested modes was obtained. From the graph, it is clear that all these 10 models have relatively high CL/CD, which suggested their high flight capabilities. The most significant curve is the one representing Javan Cucumber, which is above all the other curves, indicating that, among these 10 models, Javan Cucumber has the highest CL/CD at any AoA. According to our criteria, having the highest CL/CD in a long range of AoA means that it has the best performance among all models.

2) ANSYS Fluent

The lift data produced by ANSYS Fluent accurately matched that produced by XFLR5

More importantly, two software both suggested that Javan Cucumber and the Pyralid Moth were the top 2 performing species, which corroborated the selection of the best performing species based on XFLR5 results. The selection was thus verified to be convincing, which made our comparison reasonable and reliable.

3) Wind Tunnel Experiments

It can be observed that despite some discrepancy, the CFD and wind tunnel data all show the coherent trend. Hence our CFD simulation was validated. However, the discrepancy may undermine the validity of the CFD and wind tunnel data. Hence it is necessary to account for the discrepancy in structure testing.

4) Structure testing

By using a Scanning Electron Microscope to investigate the structure of our models. We obtained the surface landscape of our model and estimated the surface roughness. Given the surface roughness, we did some calculations to estimate the skin friction. We found that skin friction constituted 25 percent of the discrepancy. Hence it was an important factor. With this discovery, we are able to account for the discrepancy between CFD and wind tunnel results.

 

6) Comparison with the Original Javan Cucumber Design and the Optimised Javan Cucumber Design

Following the same testing methods mentioned before, we found out their performance by using both XFLR5 and ANSYS Fluent. In this figure, their relative performances are illustrated by color, deeper means better. We can now conclude that the shape in row 3 column 1 is the best.

7) Comparison between full-scale Javan Cucumber (Original and Optimised) and Conventional Designs

Javan Cucumber, with its capacity to generate high lift, outperforms XT912 in this parameter. And the biggest weakness of its structure, which is the drag being too high, has been improved by optimization. The optimized model has a lower CD and higher CL/CD. It has reduced the gap between it and XT912. Because of the outperformance of Javan Cucumber and the success of the optimized Javan Cucumber, there is much confidence that the Javan Cucumber design, after further optimization, will be able to outperform the conventional design and be a feasible choice for future transportation.

 

 

 

Conclusion

In conclusion, the hypothesis has proven to be true: Javan Cucumber performs the best among all the selected species as it can sustain high CL/CD across a wide range of AoA, and produce enough lift force at full-scale. This excellent performance is achieved because the Javan Cucumber features elliptical wings and can experience constant lift under different AoAs. Besides Javan Cucumber, we have also found the models of the Pyralid Moth performed well in the simulation and testing.

In comparison with the existing wings for motor gliders, our best model Javan Cucumber has comparable or even better aerodynamic performance. That is, when AoA is 5°, the Javan Cucumber model can carry theoretically 234 kg of mass under a low speed of 20m/s while the XT912 Tundra-Merlin can only carry 165 kg of mass at the same speed. Although this is mainly because of the bigger reference area of Javan Cucumber, such design features can still possibly be applied in future powered glider designs.

The CL/CD of the optimized Javan Cucumber was higher than the original one, which aligned with the theory that reducing chord length could increase the maximum CL/CD. Reducing chord length also helped to reduce the drag.

It can be concluded that the optimised Javan Cucumber has better aerodynamic performances, in terms of higher CL/CD and lower drag.

For the future application, our best wing design we found, which is the one with planform of Javan Cucumber, can be used on gliders which are specifically used in cities. Our wind design is capable of generating high lift with small wing area, hence it will be convenient to be used in the urban area as it will occupy relatively small spaces. Such wind design can be applied to different forms of urban transportation, both unmanned and manned. Firstly, it could be mounted on the unmanned aerial vehicles for food and package delivery, hence reducing the number of delivery vehicles on the road. Moreover, our wing design could also be used for personal flying vehicles, enhancing their aerodynamic performances, making them powerful yet space-saving, being more feasible to urban commuters. Additionally, if our wing design could be combined with the vertical take-off and landing technology, it will hopefully give a boost to urban air transportation.

About me

We are Temasek Flyers, consist of two aircraft amateurs - Wang Yiyang and Zhang Haoyu. Since the first team meeting, we have been pouring myriad endeavor in biomimetic gliders. Yiyang is a semi-professional radio-controlled powered glider pilot, and being troubled by the limited lift and efficiency for a long time. Haoyu is eager to transfer his knowledge on math to the computational simulations. We thus gained the spontaneity of forming a team so as to pursue our dreams together.

Investigating, designing, optimizing, we are heading straight towards advancements in glider designs. The gigantic Internet provides us with invaluable opportunities in retrieving researches and experiences. Greatly inspired by numerous pieces of innovative works, we all firmly believe that the combined use of nature's time-tested wisdom and human's abundant experience will pave the best way for our journey. Founding for two years, we have projected the first glance into both fields and are on the cusp of bringing them together. However, none of us will be satisfied with the current achievement and we are to discover beyond the boundary unshrinkingly. 

If it is possible, talking about our future, we will pursue our dream in aeronautics unwaveringly in universities, institutes, and companies. We think and hope that any prize, or even the opportunity to participate in the Google Science fair could be the original stimulus for an encouraging and promising future. The prizes also add up to our courage and confidence, by which there is no barrier that we cannot surmount. 

Health & Safety

Wind Tunnel:

  • To prevent hazardous effects on hearing, we wear ear mufflers for noise.
  • To prevent any accidents, we keep objects and ourselves 2m away from the suction fan.

Scanning Electron Microscope:

  • To prevent injuries from liquid nitrogen, we wear safety gloves designed for low temperature.
  • To function the high-voltage electron gun, we strictly follow the printed users' handbook.

Bibliography, references, and acknowledgements

References:

[1] "Uber Elevate." Uber Elevate | The Future of Urban Air Transport. https://www.uber.com/info/elevate/. 
[2] "EHANG 184." EHANG|Official Site-EHANG 184 autonomous aerial vehicle. http://www.ehang.com/ehang184. 
[3] "Powered hang glider." Powered hang glider https://en.wikipedia.org/wiki/Powered_hang_glider._Wikipedia. 
[4] McMasters, John. A Legacy of Sustaining Innovations in Biomimetic Aircraft Design and Engineering Education. PPT. 
[5] Mueller, T. J. Fixed and flapping wing aerodynamics for micro air vehicle applications. Reston, VA: American Institute of Aeronautics and Astronautics, 2001. 
[6] Barrett, Ronald M., and Cassandra M. Barrett. "Biomimetic FAA certifiable, artificial muscle structures for commercial aircraft wings." Smart Materials and Structures 23, no. 7 (2014): 074011. doi:10.1088/09641726/23/7/074011. 
[7] Marks, Christopher R., James J. Joo, and Gregory W. Reich. "A Reconfigurable Wing for Biomimetic Aircraft." 54th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference, 2013. doi:10.2514/6.2013-1511. 
[8] Simons, Martin. Model Aircraft Aerodynamics. Chris Lloyd Sales & Marketing, 2000. 
[9] Antony Jameson, Thomas V. Jones. “Airplane Design with Aerodynamic Shape Optimization.” PPT. 
[10] Socha, John. J., and Michael LaBarbera. "Effects of size and behaviour on aerial performance of two species of flying snakes (Chrysopelea)." Journal of Experimental Biology 208, no. 10 (2005): 1835-847. doi:10.1242/jeb.01580. 
[11] Mark Drela, and H. Youngren. XFLR5 v6.02 Guidelines: Analysis of foils and wings operating at low Reynolds numbers. PDF. February 28, 2013. 
[12] "Similitude requirements and scaling relationships as applied to model testing." NASA. https://ntrs.nasa.gov/search.jsp?R=19790022005. 
[13] Brian Maskew, Program VSAERO Theory Document. NASA Contractor Report 4023, September 1987. 
[14] Bertagnolio, Franck. NACA0015 Measurements in LM Wind Tunnel and Turbulence Generated Noise. Report. Aeroelastic Design, Wind Energy Division, Risø National Laboratory for Sustainable Energy, Technical University of Denmark. Denmark. 
[15] Kaufmann, Wick, and Wohlfahrt. "Airfoils for Fly Wings." Airfoils for Flying Wings and Tailless Airplanes. aerotools.de/airfoils/foil_flyingwings.htm. 
[16] UrbanRail.net. UrbanRail.Net > Asia > Singapore > Singapore MRT (Metro). http://www.urbanrail.net/as/sing/singapore.htm. 
[17] U.S. Standard Atmosphere, 1976. PDF. National Oceanic and Atmospheric Administration, National Aeronautics and Space Administration, United States Air Force, October 1976. 
[18] G.U.N.T. Hamburg. “HM 170 Open wind tunnel. ” https://www.gunt.de/en/products/open-windtunnel/070.17000/hm170/glct-1:pa-148:pr-769 
[19] Alsoufi, Mohammad S., and Abdulrhman E. Elsayed. "Surface Roughness Quality and Dimensional Accuracy—A Comprehensive Analysis of 100% Infill Printed Parts Fabricated by a Personal/Desktop CostEffective FDM 3D Printer." Materials Sciences and Applications 09, no. 01 (2018): 11-40. doi:10.4236/msa.2018.91002. 
[20] Morgado, J., R. Vizinho, M.a.r. Silvestre, and J.c. Páscoa. "XFOIL vs CFD performance predictions for high lift low Reynolds number airfoils." Aerospace Science and Technology 52 (2016): 207-14. doi:10.1016/j.ast.2016.02.031. 
[21] MICROLIGHT WING GUIDE. PDF. Airborne Australia Pty Ltd.  
[22] Young, T. M. "Aircraft Design Innovation: Creating an Environment for Creativity." Http://www.fzt.hawhamburg.de/pers/Scholz/ewade/2005/JAERO2007_Young.pdf 
[23] Spalart, P. R. and Allmaras, S. R., 1992, "A One-Equation Turbulence Model for Aerodynamic Flows" AIAA Paper 92-0439 
[24] McCroskey, W. J. "A Critical Assessment of Wind Tunnel Results for the NACA 0012 Airfoil." NASA Technical Memorandum 100019, October 1987. 
[25] Ohtake, Tomohisa, Yusuke Nakae, and Tatsuo Motohashi. "Nonlinearity of the Aerodynamic Characteristics of NACA0012 Aerofoil at Low Reynolds Numbers." Journal of the Japan Society for Aeronautical and Space Sciences, vol. 55, no. 644 (2007): 439-445. 
[26] Eng.ucy.ac.cy. http://www.eng.ucy.ac.cy/EFM/Manual/HM%20170/Hm170e.pdf  
[27] Researchgate, https://www.researchgate.net/figure/SEM-images-of-porcelain-surface-a-Polished-controldemonstrating-some-scattered_fig6_263764265  
[28] Cool.conservation-us.org.  JAIC 1993, Volume 32, Number 2, Article 8 (pp. 177 to 206). http://cool.conservation-us.org/jaic/articles/jaic32-02-008.html 
[29] Mills, A. F., and Xu Hang. "On the Skin Friction Coefficient for a Fully Rough Flat Plate." Journal of Fluids Engineering105, no. 3 (1983): 364. doi:10.1115/1.3241008. 
[30] http://www.hzzoo.com/newsinfo.aspx?kid=893动物分类学基本知识-动物分类学-杭州动物园.  http://www.hzzoo.com/newsinfo.aspx?kid=893 
[31] NASA. Accessed September 06, 2018. https://history.nasa.gov/SP-367/f134.htm. 
[32] Bhushan, Bharat. "Surface Roughness Analysis and Measurement Techniques." http://home.ufam.edu.br/berti/nanomateriais/8403_PDF_CH02.pdf. 
[33]"Long Ca." Site Title Long Ca. Accessed September 06, 2018. http://www.pdhkt.com/2016/05/23/aircraftstability/. 
[34] Liu, Xu, Wei Liu, and Yunfei Zhao. "Unsteady Vibration Aerodynamic Modeling and Evaluation of Dynamic Derivatives Using Computational Fluid Dynamics." Mathematical Problems in Engineering2015 (2015): 1-15. doi:10.1155/2015/813462.]

 

Acknowledgment:
 
First and foremost, we would like to thank Dr. Victor Wang Peng Cheng of Singapore Institute of Technology, our mentor, for providing insights and expertise on CFD, as well as giving comments that greatly improved our manuscript. We would also like to thank Mr. Khoh Rong Lun of Temasek Junior College and Mrs. Jeanne Wan of Anglo-Chinese Junior College for their unwavering support and invaluable guidance through this year-long project. In addition, we would like to give special thanks to Dr. Dimitrios Chua Khim Heng for his patient guidance and assistance throughout the experiment session of our research. We also want to thank Dr. Wulf Hofbauer of Centre For Research And Applied Learning In Science (CRADLΣ) in Singapore Science Centre for his assistance during the SEM analysis and his help in determining surface roughness. We would also like to thank Mr. Jeggathishwaran S/O Panisilvam for his enthusiasm to assistant we use the workstation at SIT to run the Fluent simulation. Last but not least, we are grateful for all the teachers, seniors, and peers who gave comments and provided kind helps to us on this long journey, both online and offline.

Wang Yiyang mainly contributed to the identification and selection of the species that this project studied. He worked on the planform sketching in AutoCAD and XFLR5 simulations. The ICEM CFD meshing was mainly done by him and he prepared most data analysis forms. He contributed to the constructing of 3D models in CATIA and 3D printing for models being used for the wind tunnel experiments. Yiyang also studied the algorithm of Fluent and were in charge of the data presentation using Origin. Moreover, Yiyang constructed the MAV platform and did the flight test at Temasek Junior College and Dunman High School.

Zhang Haoyu mainly contributed to the study of the algorithm behind XFLR5 and was in charge of simulations run in Fluent. He also conducted the SEM analysis and calculation. Yiyang and Haoyu worked together to conduct the wind tunnel experiments, as well as the optimization of Javan Cucumber. Haoyu did the mesh independence study. The entire team worked together to write this research paper.