Comptational Flid Dynamics Simlation and Wind Tnnel Testing on Microlight Model Iskandar Shah Bin Ishak Department of Aeronatics and Atomotive, Universiti Teknologi Malaysia T.M. Kit Universiti Teknologi Malaysia shah@fkm.tm.my Abstract: This paper aims to highlight the wind tnnel testing techniqes and the Comptational Flid Dynamics (CFD) stdies on a scaled-down microlight model for its aerodynamic characteristics. The wind tnnel testing is condcted at a grace of Universiti Teknologi Malaysia Open Loop Sbsonic Tnnel facility with variations of angle of attack and flaps deflection angle. Wind tnnel corrections, sch as blockage effects, are considered dring data redction process in order to have reslts that will be almost precisely the same as in actal flight. In additional, the CFD simlation will be carried ot sing the software Flent 6.1 on the model. Comparison of these two methods later depicts that the reslt obtained from the wind tnnel testing is agreeable with the reslt simlated by the CFD. Key Words: Wind tnnel testing, Comptational Flid Dynamics (CFD), Microlight 1.0 INTRODUCTION Of late, the implementation of wind tnnel testing and simlation by CFD is becoming a trend in the stage of the design analysis process. This paper will present the wind tnnel testing techniqe on a 1:25 scaled-down model of single-seated microlight and the data redction procedres. CFD simlation is also carried ot for comparison prposes. 2.0 EXPERIMENTAL STUDY The objective of this testing is to determine the aerodynamic characteristics of a microlight which is designed previosly nder a stdent ndergradate project at Universiti Teknologi Malaysia. Its configrations as follow: Microlight Configrations 1) Wing specifications: Wing location Dihedral angle Wing plan form Aerofoil section High wing 3 0 Rectanglar NACA 2412 2) Aircraft gross weight 247.54kg 3) External dimension: Wing span, b 9 m Total length 5.5 m 4) Crising speed 92.67km/h 5) Service ceiling 3000 m Table 1: Microlight Configrations A 1:25 scaled-down model is fabricated for the prpose of this wind tnnel testing, as shown in Figre 1.
Figre 1: Microlight Model The testing is condcted at a grace of Universiti Teknologi Malaysia Open Loop Sbsonic Tnnel facility. This tnnel is a sction type which the fan is sitated at the exit of the tnnel. The test section is 0.457 m(h) x 0.457 m(w) x 1.27 m(l). The experiment is carried ot at the wind speed of 20 ms -1 (Reynolds Nmber, Re = 71 170). For this kind of testing, the measring device sed is a 3-Components Balance which is capable of measring lift force, drag force and pitching moment. Figre 2: 3-Components Balance Figre 3: Model dring wind tnnel testing
2.1 Wind Tnnel Data Correction Reslts obtained from wind tnnel testing need to be corrected following the blockage effect, boyancy, wall interference and STI (Strt, Tare and Interference) effects. However for this kind of testing, only blockage effect is considered as it contribtes qite significantly to the final reslt, the other corrections are assmed to be very less significant, and ths be ignored. The blockage is mainly comprised of solid and wake blockage. 2.1.1 Solid Blockage The present of the model in the test section will actally redce the area throgh which the air mst flow. From the Continity and Bernolli s eqations, this will increase the velocity of the air arond the model. This is called Solid Blockage [1]. It is a fnction of the model thickness, thickness distribtion and model size bt is independent of the camber. For example, as the velocity arond the model increase de to this effect, it will give the lift coefficient, C L a higher vale. Therefore the solid blockage correction needs to be performed to have the right C L vale. 2.1.2 Wake Blockage A real body withot sction type layer control will have a wake behind it and this wake will have a mean velocity lower than the freestream velocity. According to the Law of Continity, the velocity otside the wake in a closed tnnel mst be higher than the freestream velocity in order that a constant volme of flid may pass throgh. This higher velocity has a lowered pressre and as the bondary layer grows on the model, pts the model in a pressre gradient. Hence, the velocity of the air arond the model will increase and therefore, the reslt again needs to be corrected. 2.2 Data Corrections After calclating the solid and wake blockage effect, the total blockage effect is as follow: ε total = ε sb, w + ε sb, fselage + ε wb where ε sb, w = Solid blockage for wing ε sb, fselage = Solid blockage for fselage = Wake blockage ε wb Then the ncorrected freestream velocity and dynamic pressre can be corrected by applying these eqations: V q c c = V = q ( 1+ ε ) total 2 ( 1+ ( 2 M ) ε ) total where q = Dynamic pressre V = Freestream velocity M = Mach Nmber Hence, the corrected lift coefficient C Lc, drag coefficient C Dc and pitching moment coefficient C Mc can be extracted by a simple relation as follow:
q C Lc = CL CDc = CD CMc = qc qc q Others correction sch as temperatre, Mach nmber or pressre are not applicable here becase the flow is assmed to be incompressible flow with constant temperatre. 2.3 Reslts q q c C M Lift Coefficient vs Angle of Attack 1.5 CL 1.0 0.5-15 -10-5 0 5 10 15 20-0.5 Flap 0 Flap 10 Flap 20 Figre 4: Lift Coefficient vs Angle of Attack Drag Coefficient vs Angle of Attack 0.3 CD 0.2 0.1 Flap 0 Flap 10 Flap 20-15 -10-5 0 5 10 15 20 Figre 5: Drag Coefficient vs Angle of Attack
Pitching Moment Coefficient vs Angle of Attack CM 0.3 0.2 0.1-15 -10-5 -0.1 0 5 10 15 20-0.2 Flap 0 Flap 10 Flap 20 Figre 6: Pitching Moment Coefficient vs Angle of Attack 3.0 CFD SIMULATION As an alternative way to obtain the aerodynamic characteristic and to verify the wind tnnel reslt, Comptational Flid Dynamics (CFD) is sed. The first step of setting p Flent 6.1 is to draw a solid model. This task is carried ot by a SolidWorks software. The 3D drawing of microlight model has been generated and then be ct at the symmetry plane becoming a half model. This is to shorten the simlation time while the reslts obtained will be the same. The half solid model in shown in Figre 7 meanwhile reslts at zero angle of attack are shown in Figre 9 and 10 respectively. Figre 7: The Half of Solid Model
Figre 8: Meshed volme of microlight model Figre 9: Velocity contors by 0 0 angle of attack at 20 m/s Figre 10: Pressre contors by 0 0 angle of attack at 20 m/s 4.0 DISCUSSION It can be noticed from Figre 4, the graph for Flap 20 (i.e. flaps are deployed at 20 0 ) is located at the above, followed by the graph for Flap 10 and then at the bottom, the graph for zero deployment of flaps. This trend is correct as in principal; flap deployment will increase the lift coefficient and, nfortnately, same goes for the drag coefficient! From Figre 6, it can be conclded that this microlight is statically stable in longitdinal axes as the crve demonstrate a negative slope. Figre 11 and 12 depict the differences between reslts obtained form wind tnnel testing and CFD, respectively.
Lift Coefficient vs Angle of Attack At Flap Zero 1.5 1.0 CL 0.5-15 -10-5 0 5 10 15 20-0.5 WTT CFD Figre 11: Comparison between Wind Tnnel Testing and CFD Reslts for Lift Coefficient Drag Coefficient vs Angle of Attack at Flap Zero 0.3 CD 0.2 0.1 WTT CFD -15-10 -5 0 5 10 15 20 Figre 12: Comparison between Wind Tnnel Testing and CFD Reslts for Drag Coefficient From both graphs plotted, it can be seen how similar the reslts obtained from both methods. Both trend lines almost overlap each other. Means both methods are agreeable with each other. However, errors still occrred here and cased the reslts differential. Some errors predicted are: i. The model is not exactly 100 % same as in the CFD de to the delicate model-making process. ii. Different density between CFD (defalt 1.225 kgm -3 ) and Wind Tnnel Testing (1.1784kgm -3 ). iii. The flow qality of UTM-Open Loop Sbsonic Tnnel. iv. A slightly vibration of 3-Components Balance device dring testing. CONCLUSION As for plain speaking, the wind tnnel reslts that be presented throghot this paper can be accepted for ftre aerodynamic analysis as the reslts are following the correct trend and also agreeable with the CFD reslt. However for a better reslt, it is recommended to condct the wind tnnel testing at a higher Re in order to have dynamic similarity. Some advance tests sch as rolling and yawing testing are sggested to be carried ot in the ftre development. ACKNOWLEDGEMENT
Md Nizam Dahalan, Universiti Teknologi Malaysia. REFERENCES [1] Jewel B. Barlow, William H. Rae, Jr., Alan Pope, Low-Speed Wind Tnnel Testing, 3 rd Edition, Wiley- Interscience, 1999. [2] British Microlight Aircraft Association BMAA, Gide To Airworthiness Procedres, BMAA, Bllring Deddington Banbry, 1995. [3] John D. Anderson, Jr., Introdction to Flight, 4 th Edition, McGraw-Hill, 2000. [4] Jewel B. Barlow, William H. Rae, Jr., Alan Pope, Low-Speed Wind Tnnel Testing, 3 rd Edition, Wiley- Interscience, 1999. [5] Tang Mng Kit, Wind Tnnel Testing on Microlight Model, Universiti Teknologi Malaysia, Thesis, 2005. [6] Mohd Naemy Fahimy Bin Mstapa, Rekabentk Pesawat Ringan (Microlight), Universiti Teknologi Malaysia, Thesis, 2000.