BOUNDARY-LAYER CONTROL OF BLUFF BODIES WITH APPLICATIONTO DRAG REDUCTION OF TRACTOR-TRAILER TRUCK CONFIGURATIONSbyBIN YINGB.A.Sc., Nanjing Aeronautical Institute, 1982A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE STUDIESDepartment of Mechanical EngineeringWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIADecember 1991© Bin Ying, 1991In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature) Department of I^E The University of British ColumbiaVancouver, CanadaDate 3)e„<,, L7 DE-6 (2/88)11ABSTRACTEffectiveness of two fundamentally different concepts of boundary-layercontrol for the drag reduction of bluff bodies is studied experimentally. Themethods are:(i) Moving Surface Boundary-layer Control (MSBC) involvingmomentum injection through one or more rotating elements (lighthollow cylinders); and(ii) tripping of the boundary-layer using judiciously located fences tointerrupt pressure recovery.Wind tunnel tests with a two-dimensional wedge airfoil model suggest thatinjection of momentum can significantly delay separation of the boundary-layerresulting in a narrow wake and the associated reduction in the pressure drag.It also leads to a substantial increase in the lift at a given angle of attackresulting in a dramatic rise in the lift to drag ratio, from 2 to 80, underoptimum conditions. Effectiveness of the momentum injection process isprimarily governed by the gap-size, cylinder surface roughness and the ratioof the cylinder surface velocity (Uc) to the free stream velocity (U). A three-dimensional model of a rectangular prism and a 1/12 scale model of a typicaltractor-trailer truck configuration show that both the MSBC and fenceapproaches are promising in reducing the aerodynamic resistance. A kitconfiguration is proposed for ease of implementation of the concepts on new111and existing trailers. Road tests with a full scale cube-truck are recommendedto assess effectiveness of the boundary-layer control procedures in reducing thedrag during highway conditions.ivTABLE OF CONTENTSABSTRACT ^ iiLIST OF TABLES viLIST OF FIGURES ^ viiNOMENCLATURE xiiiACKNOWLEDGEMENT^ xv1. INTRODUCTION ^11.1 Background ^11.2 A Brief Review of the Relevant Literature ^21.3 Scope of the Present Investigation ^82. MODELS AND TEST PROCEDURE ^ 112.1 Two Dimensional Wedge Airfoil ^112.2 Tractor-trailer Truck Model ^132.2.1 Tractor-trailer truck model with twin rotating cylinders^132.2.2 Tractor-trailer truck model with fences ^152.2.3 Tractor-trailer truck model with a cylinder kit ^172.3 Test Procedure for the Tractor-trailer Truck Model ^202.3.1 Wind tunnel ^202.3.2 Model support system ^262.3.3 Cylinder rotation-rate measuring system ^262.3.4 Drag measurement system ^27V3. RESULTS AND DISCUSSION ^ 293.1 Two Dimensional Wedge Airfoil ^293.2 Tractor-trailer Truck Model with Twin Cylinders ^533.3 Rectangular Prism with Fences ^813.4 Tractor-trailer Truck Model with Fences ^813.5 Tractor-trailer Truck Model with both Twin Cylinders andFences ^873.6 Tractor-trailer Truck Model with a Cylinder Kit ^943.6.1 Cylinder kit 1 ^953.6.2 Cylinder kit 2 ^ 1024. CONCLUDING REMARKS 1094.1 Summary of Results ^ 1094.1.1 Application of the MSBC to a 2-D wedge airfoil model ^ 1094.1.2 Application of the MSBC and fences to a 3-D truck model 1114.2 Recommendation for Future Work^ 112REFERENCES^ 115viLIST OF TABLESTable 1. Wind tunnel tests conducted with different orientation ofthe twin helical-groove and spline cylinders: location of thefront cylinder was at the top leading edge and the rearcylinder at 25.4 cm downstream from the leading edge. . . . 59viiLIST OF FIGURESFigure^ Page1. The practical application of moving wall for boundary-layercontrol was demonstrated by Favre in 1938. Using an airfoil withthe upper surface formed by a belt moving over two rollers, hewas able to delay separation until the angle of attack reached55°, where the maximum lift coefficient of 3.5 was realized. . . . . 42. Schematic diagrams explaining principles of the MSBC and theboundary-layer trip devices in reducing drag of bluff bodies. . . . .^63. Schematic diagrams of the bluff bodies studied during the windtunnel test-program^ 104. A photograph showing the cylinders with different surfaceroughnesses used in the test-program: (a) smooth; (b) helicalgrooves; (c) roughness squares; (d) spline-1; (e) spline-2 . ^ 125.^Photographs showing the wind tunnel test arrangement for the2-D wedge airfoil. The model was supported by the lift-dragstrain gage balance with an oil damper to minimize vibration. .. 14^6.^Schematic diagrams showing arrangement of rotating cylindersand fences for the boundary-layer control:(a) MSBC using twin cylinders; ^(b) fences for tripping of the boundary-layer. ^^7.^Kit configurations developed for application to existing trucks asan add-on device:(a) schematic diagrams of kit 1 with the rotating elementprojecting 12.7 mm in the fluid stream; ^(b) schematic diagrams of kit 2 with the cylinder projecting31.8 mm in the flow; ^(c)^^close-up showing a splined cylinder serving as themomentum injection unit and two vertical fences for18192122viiitripping the boundary-layer with kit 2. ^ 23^8.^The boundary-layer wind tunnel, with a test cross-section of2.44x1.6 m, was used to study 1/12 scale models of the trailerand truck configurations:(a) schematic diagram showing details of the tunnel with themodel near the entrance to the test-section (first-bay); . . . . 24(b) photographs showing a tractor-trailer truck model (1/12scale), with boundary-layer control devices, being preparedfor wind-tunnel tests . ^ 25^9.^3-D tractor-trailer truck model test arrangement in the U.B.C.boundary-layer wind tunnel. ^ 2810. Aerodynamic coefficients for a two-dimensional wedge airfoil asaffected by the rotation of the smooth cylinder:(a) lift coefficient; ^ 31(b) drag coefficient; 32(c) CL 1 CD 3311. Variation of the aerodynamic coefficients with the angle of attackand helical groove cylinder rotation:(a) lift; ^ 35(b) drag; 36(c) CL / CD 3712. Plots showing effect of the cylinder surface roughness conditionon the lift and drag coefficient:(a) CL; ^ 39(b) CD; 40(c) CLI CD. 4113. Influence of axial surface grooves on the momentum injection(spline-1) as reflected through the variation of lift and dragcoefficients:(a) CL; ^ 42(b) CD; 43(c)^CI,/ CD. 44ix14. Effect of the surface roughness designated as spline-2 on theaerodynamic coefficients:(a) CL; ^ 45(b) CD; 46(c) CL / CD. 4715. Comparison charts showing relative merit of the surfacecondition on the boundary-layer control:(a) CLmax; ^ 49(b) C Dmin; 50(c) (CL /CD )max. ^ 5116. A schematic diagram of the closed circuit water channel facilityused in the flow visualization study. Slit lighting was used tominimize distortion due to three dimensional character of theflow. Long exposure provided path-lines with polyvinyl chlorideparticles serving as tracers. The dimensions are in mm . ^17. Typical flow visualization photographs for a wedge shaped airfoil,with a smooth surface cylinder, showing remarkable effectivenessof the MSBC concept for a = 30°, Rn = 3x104. Note theprogressive downstream shift of the separation point as the LIc lUincrease. Eventually the flow appears to approach the potentialcharacter^18. The concept of moving surface boundary-layer control continuesto be effective event at a high angle of attack of 55° as showndramatically by these flow visualization pictures (Rn = 3x104) . . . 5519. Effect of the moving surface boundary-layer control on the dragcoefficient of a tractor-trailer truck configuration. Note, anincrease in the sandpaper roughness contributes to the dragreduction through efficient momentum injection^ 5720. Effect of the twin helical cylinder configuration on themomentum injection and boundary-layer control:(a) Case 1: both cylinders flush; ^ 60(b) Case 2: front cylinder flush, rear cylinder raised 6.35 mm; 615254x(c) Case 3: front cylinder flush, rear cylinder raised 12.7 mm; 62(d) Case 4: front cylinder raised 6.35 mm, rear cylinder raised12.7 mm; ^ 63(e) Case 5: front cylinder raised 6.35 mm, rear cylinder flush; 64(f) Case 6: front cylinder raised 12.7 mm, rear cylinder flush;^65(g) Case 7: front cylinder raised 12.7 mm, rear cylinder raised6.35 mm; ^ 66(h) Case 8: front cylinder raised 12.7 mm, rear cylinder raised12.7 mm; 67(i)^Case 9: front cylinder raised 6.35 mm, rear cylinderraised6.35 mm. ^ 6821. Variation of the drag coefficient CD with the speed ratio for thetwin spline cylinder configurations:(a) Case 1: both cylinders flush; ^ 70(b) Case 2: front cylinder flush, rear cylinder raised 6.35 mm; 71(c) Case 3: front cylinder flush, rear cylinder raised 12.7 mm; 72(d) Case 4: front cylinder raised 6.35 mm, rear cylinder raised12.7 mm; ^ 73(e) Case 5: front cylinder raised 6.35 mm, rear cylinder flush; 74(f) Case 6: front cylinder raised 12.7 mm, rear cylinder flush; 75(g) Case 7: front cylinder raised 12.7 mm, rear cylinder raised6.35 mm; ^ 76(h) Case 8: front cylinder raised 12.7 mm, rear cylinder raised12.7 mm; 77(i)^Case 9: front cylinder raised 6.35 mm, rear cylinder raised6.35 mm ^ 7822. Effect of cylinder rotation on the flow past a tractor-trailer truckconfiguration. ^ 7923. Effect of cylinder rotation on the flow past a tractor-trailer truckconfiguration. Note, the tractor geometry and spacing betweenthe tractor and trailer are different here. ^ 8024. Effect of the fence width and height on the drag of a three-dimensional prism. ^ 8325. A schematic diagram showing application of the fences on thexifront, exposed face of the trailer of a truck^ 8426. Variation of the drag coefficient with the position of horizontalfence 1^ 8527. Variation of the drag coefficient with the position of horizontalfence 2 when fence 1 is fixed at its critical orientation. ^ 8628. Variation of the drag coefficient with the position of twin verticalfences 3 and 4 when fences 1 and 2 are fixed at their criticalorientations . ^ 8829. Stages in fine tuning of the fence dimensions and associated dragcoefficient:(a) The fences were extended to the edges of the trailer frontface in both horizontal and vertical directions^ 89(b) The twin vertical fences were truncated to horizontal 2fence position from the bottom edge of the trailer frontface. ^ 90(c) The twin vertical fences were truncated further tohorizontal 1 fence position from the top edge of the trailerfront face. ^ 91(d) The twin horizontal fences had offset from both side edgesof the trailer front face . ^ 9230. Variation in the drag coefficient with the cylinder speed ratio fora hybrid configuration involving two vertical fences and twinrotating cylinders in their optimum geometry. ^ 9331. Variation in the drag coefficient with the cylinder speed ratio asaffected by kit 1 vertical orientation:(a) absence of the front cover; ^ 97(b) with the front cover. ^ 9832. Variation in the drag coefficient with the cylinder speed ratioand kit 1 position with two vertical fences mounted optimally onthe front face of the kit. 99xii33. Variation in the drag coefficient with the cylinder speed ratio forkit 1 at the optimum vertical orientation. Note the influence ofcover and fence geometry^ 10034. Variation in the drag coefficient with the cylinder speed ratio forkit 2 at different vertical position. Note the effect of gap. ^ 10435. The drag coefficient as affected by position of the two verticalfences on the front face of kit 2, set at the optimum position,with the gap covered and stationary cylinder. ^ 10536. Variation in the drag coefficient as affected by the position of ahorizontal fence with two critically-oriented vertical fences on thefront face of kit 2. The gap was covered and the cylinder washeld stationary. ^ 10637. Variation in the drag coefficient with the cylinder speed ratio forkit 2 at its optimum vertical orientation, gap covered and fencesin critical geometry. 108NOMENCLATUREU^free stream velocityUc cylinder surface speeda^angle of attackc^wedge airfoil chordP^air density1^wedge airfoil spanA^projected area normal to the free streamCL^mean lift coefficient, lift / (1/2)pU2c1CD mean airfoil drag coefficient, drag / (1/2)pU 2c1 or drag / (1/2)pU 2AUcf front cylinder surface speedUcr rear cylinder surface speedd^hydraulic diameter, (4A/n) 1/2H^trailer heightB^trailer widthL^trailer lengthhf^fence heightbf^fence widthbfh horizontal fence widthbfv vertical fence widthyf^vertical coordinate of the horizontal fencexivxf^horizontal coordinate of the vertical fence( )cr critical value()min minimum value()max maximum valueHk cylinder kit coordinate representing distance from the top surface of thetrailer to the horizontal center line of the cylinderBe^kit front cover width, same as that of the trailerHe kit front cover heightGP gap between the cylinder kit and the cabp^standard air viscosity, 17.8 x 10 -6 kg/msRn Reynolds number, pUdipXVACKNOWLEDGEMENTA special thank is extended to my supervisor, Dr. V. J. Modi for his timeand guidance throughout the project. His insight and amiable nature has madethis project a thoroughly enjoyable experience. I would not have got the keypoints without his novel ideas and hints during the study.Assistance of Professor T. Yokomizo in the flow visualization study isgratefully acknowledged. It was carried out in his laboratory at Kanto GakuinUniversity, Yokohama, Japan.The assistance of Mr. E. Abell, Senior Technician, in construction of themodel is gratefully acknowledged.The investigation reported here was supported by the Science Council ofBritish Columbia, Grant Nos. AGAR 5-53628, 5-53698 and Natural Sciencesand Engineering Research Council of Canada, Grant No. A-2181.11. INTRODUCTION1.1 BackgroundIdentification and exploitation of energy sources such as fossil fuels,solar, wind, geothermal, tidal, biomass, etc. have received attention for quitesome time. The oil embargo emphasized a need for self-sufficiency in energy.At the same time, energy conservation measures as reflected in the improveddesign of engine, aircraft with high lift/drag ratio, aerodynamically shapedcars, better insulated houses, heat recovery in industrial processes, etc. havealso gained importance. The proposed project focuses on one such energyconservation area, which promises to have a significant economic consequence.It addresses the sector of energy consumption, the commercial roadtransportation industry, which is :(i) already significant in size;(ii) presently growing and promising to maintain that trend.The present investigation is directed at energy conservation through aneffective reduction in aerodynamic resistance of a typical truck configuration.It should be recognized that:(a) two-thirds of all the goods in North America are transported bytrucks;(b) on an average a truck covers around 130,000 - 150,000 km/year;2(c) 50 - 70% of the truck's power (at 80 - 100 km/hr speed) isconsumed in overcoming aerodynamic resistance, compared toaround 20 - 30% for rolling friction and 10 - 15% lost in thetransmission system.A simple analysis of this set of data suggests that even 1% reduction inthe aerodynamic drag can amount to a significant saving in the fuel cost [1].Aerodynamic resistance is a result of the complex three dimensional flow fieldassociated with the truck geometry, and is governed by the boundary layerseparation as well as reattachment regions. It is sensitive to speed and therelative wind direction (yaw incident angle). In general, the drag coefficient fora conventional truck may vary significantly depending on its geometry, speedand yaw [2].1.2 A Brief Review of the Relevant LiteratureEver since the introduction of the boundary-layer concept by Prandtl,there has been a constant challenge faced by scientists and engineers tominimize its adverse effects and control it to advantage. Methods such assuction, blowing, vortex generators, turbulence promoters, etc., have beeninvestigated at length and employed in practice with a varying degree ofsuccess. The vast body of literature accumulated over years has been reviewedrather effectively by several authors including Goldstein [3], Lachmann [4],3Rosenhead [5], Schlichting [6], Chang [7], and others. However, the use ofmoving surface for boundary-layer control has received relatively lessattention.Irrespective of the method used, the main objective of a control procedureis to prevent, or at least delay, separation of the boundary-layer. A movingsurface attempts to accomplish this in two ways:(i) it retards growth of the boundary-layer by minimizing relativemotion between the surface and the free stream;(ii) it injects momentum into the existing boundary-layer.A practical application of moving wall for boundary-layer control wasdemonstrated by Favre [8]. Using an airfoil with upper surface formed by abelt moving over two rollers (Figure 1), he was able to delay separation untilthe angle of attack reached 55° where the maximum lift coefficient CLmax = 3.5was realized. Alvarez-Calderon and Arnold [9] carried out tests on a rotatingcylinder flap to develop a high lift airfoil for STOL type aircraft. The systemwas tested in flight on a single engine high-wing research aircraft.Of some interest is the North American Rockwell designed OV-10Aaircraft which was flight tested by NASA's Ames Research Center [10-12].Cylinders located at the leading edges of the flaps were rotated at high speedwith the flaps in lowered position. The main objective of that test program wasto assess handling qualities of the propeller powered STOL type aircraft athigher lift coefficients. The aircraft was flown at speeds of 29-31 m/s, along Figure 1. The practical application of moving wall for boundary-layer control was demonstrated by Favrein 1938. Using an airfoil with the upper surface formed by a belt moving over two rollers, hewas able to delayer separation until the angle of attack reached 55°, where the maximum liftcoefficient of 3.5 was realized. 45approaches up to -8°, which corresponded to a lift coefficient CD .-= 4.3. In thepilot's opinion any further reductions in approach speed were limited by thelateral-directional stability and control characteristics.In terms of trying to understand the phenomenon at the fundamentallevel Tennant's contribution to the field is significant. Tennant et al. [13] haveconducted tests with a wedge shaped flap having a rotating cylinder as theleading edge. Flap deflection was limited to 15° and the critical cylindervelocity necessary to suppress separation was determined. Effect of increasingthe gap-size (between the cylinder and the flap surface) was also assessed. Noeffort was made to observe the influence of an increase in the ratio of cylindersurface speed (Uc) to the free stream velocity (U) beyond 1.2.Through a comprehensive wind tunnel test program involving a familyof airfoils with one or more cylinders forming moving surfaces, complementedby the surface singularity numerical approach and flow visualization, earlierstudies by Modi et al. [14 - 17] have shown spectacular effectiveness of theconcept, which increased the maximum lift coefficient by more than 200% anddelayed the stall angle to 48°.Yet another approach to boundary-layer control can be through itstripping by judiciously located fences on the front face of a bluff body. Thisinterferes with the pressure recovery thus promising to reduce drag.The basic concepts involved in the boundary-layer control through theabove two methods are illustrated in Figure 2. It shows a bluff body, a two-DragBluff BodyUand preventing the pressure recovery.Potential FolwWakeU CPf^ PbIt is apperent that by increasing Pb or reducing Pf we canreduce the pressure drag.(i) MSBC tends to increase Pb by keeping the flow attached;(ii) Fences tend to reduce Pf by tripping the boundary-layerFigure 2. Schematic diagrams explaining principles of the MSBC and the boundary-layer trip devices inreducing drag of bluff bodies.^ rn7dimensional prism, located in a fluid stream at zero angle of attack. Pt. and Pbare pressures on the front and rear faces, respectively. They are assumed to beuniform over the faces, in this illustrative example, for simplicity. Obviously,by increasing Pb and/or decreasing Pt-we can reduce the pressure drag. MSBCtries to increase Pb by keeping the flow attached. On the other hand, fencesreduce Pf by tripping the boundary-layer. These principles are explainedthrough diagrams of the flow past a circular cylinder in the same figure. At thestagnation point the pressure is the largest and pressure coefficient is 1. Theboundary-layer separates at O forming the wake. In the wake the pressure isessentially uniform at a lower value. This is what fences try to achieve. If theseparation is prevented, ideally the pressure will reach the stagnation value.This is what the MSBC tries to accomplish.A comprehensive literature review of the road vehicle aerodynamicssuggests that although aerodynamically contoured car design has become astandard practice lately, the trucks and buses have changed little during thepast 30 years [18 - 21]. Most of the modifications have been limited to roundededges with provision for vanes, skirts and flow deflectors. The benefit due tosome of the "add-on" devices is still a matter of controversy and, at best,marginal under conditions other than the specific ones used in their designs.Bearman [22] has presented an excellent review on the subject (with 54references cited). The thesis by Wacker [23] also discusses limited influence of"add - on" devices with a possibility of increasing the drag under non-optimal8conditions. On the other hand, it was found that judicious choice of groundclearance, gap-size between the tractor and the trailer, and back inclinationcan reduce the drag coefficient by a significant amount.A word concerning numerical analysis of the complex aerodynamicsassociated with road vehicles would be appropriate. A reliable and cost-effective methodology, if available, can assist in design with reduceddependence on time consuming and expensive wind tunnel tests. With theadvent of supercomputers, parallel processing and neural network concepts,considerable progress has been made in that direction. However, modelling ofthree dimensional boundary layers around a complex geometry at supercriticalReynolds numbers, with separation, reattachment and reseparation of unsteadyturbulent flows, still represents a challenging problem [24,25].1.3 Scope of the Present InvestigationThe present study builds on this background and explores application ofthe two concepts: (a) Moving Surface Boundary-layer Control (MSBC); and (b)trip fences; to a two-dimensional wedge airfoil and tractor-trailer truckconfigurations. The extensive wind tunnel test program, complemented by aflow visualization study, investigates effectiveness of:(i) the MSBC for 2-D wedge airfoil;(ii) the MSBC for 3-D tractor-trailer truck;9(iii) the trip fences when applied to the upstream face of a rectangularprism and the trailer; and(iv) combinations of the MSBC and fences as applied to a tractor-trailer truck configuration.A schematic diagram of the configurations studied is presented in Figure 3. Animportant parameter during the MSBC is the ratio of the cylinder surfacevelocity (Ut) to the free stream velocity (U), which was systematically variedduring the test-program conducted in the smooth flow condition. In the fencestudy, the variables of interest are the fence width and height (bf and hf,respectively) and their locations that would lead to a maximum reduction indrag.bfU •►• • •Hk11.5MMINIIIN DT.•MSBC for 2-D Wedge Airfoil^MSBC for 3-D Tractor-trailer TruckUcf Uc rO cal0 0Fences for Rectangular Prism and Truckhfhf^ --1.1 14—biCombination of MSBC and Fences for Truckhf---.1^ h f,eng C0 0Uc f Uc r^ --01 14--bfU •■►^ U ■■►ca .Y.1 INIMINIFigure 3. Schematic diagrams of the bluff bodies studied during the wind tunnel test-program.1 12. MODELS AND TEST PROCEDURE2.1 Two Dimensional Wedge AirfoilThe two dimensional wedge-airfoil model was tested in a 45 x 45 cmcross-section wind tunnel with a maximum speed of 50 m/s. The largeconverging nozzle at the entrance of the tunnel (contraction ratio = 10 : 1)makes the flow in the test-section uniform with a level of turbulence less than0.5%. The tunnel speed can be adjusted by a variac transformer and measuredusing a pitot static tube connected to an inclined alcohol manometer.The wedge model with a chord length of 11.5 cm and the tail angle of ,----:14° was constructed from aluminum with the nose replaced by a cylinder of2.54 cm diameter and of desired surface roughness. Five different surfaceconditions were used in the test-program (Figure 4) designated as:(i) smooth;(ii) helical grooves;(iii) roughness squares;(iv) spline-1;(v)^spline-2.The difference between (iv) and (v) is essentially characterized by theshape of splines. The (iv) is with sharp teeth whereas the (v) with round ones.The cylinder was driven by a variac controlled 1/8 H.P. a.c. motorFigure 4. A photograph showing the cylinderswith different surface roughnessesused in the test-program: (a)smooth; (b) helical grooves; (c)roughness squares; (d) spline-1; (e)spline-2.(a) (b)^(C)^(d)^(e)co.w.w.ww, w ,e,ww , w,u , w , w, wasmit,41111111141110111111111.013through a flexible belt system. The motor speed was monitored using a digitalstroboscope. In the present test-program the ratio Uc l U was varied from 0 - 4.This corresponds to a maximum cylinder speed of around 12,000 rpm at a freestream speed of 3.85 m/s (R n = 3x104). To ensure two dimensionality of the flowthe model was fitted with end plates. The lift and drag forces were recordedover a range of the angle of attack of 0° - 55° with 5° increment. The force canbe measured with an accuracy of 0.5 gm/mV.The models were susceptible to vibration, particularly at high angles ofattack, due to the turbulent wake created by the shedding vorticity. This wasminimized by an externally located viscous oil damper. The test arrangementis shown in Figure 5.2.2. Tractor-trailer Truck ModelA 1/12 scale tractor-trailer truck model was constructed out of Plexiglas.The model has a trailer with width B = 22.7 cm, height H = 26.2 cm, andlength L = 128.4 cm, with a hydraulic diameter of . It can be used to assesseffectiveness of the MSBC, tripping of the boundary-layer using fences, or acombination of the two.2.2.1 Tractor-trailer truck model with twin rotating cylindersFigure 5. Photographs showing the wind tunnel test arrangement for the 2-D wedge airfoil. The modelwas supported by the lift-drag strain gage balance with an oil damper to minimize vibration.15The two identical cylinders of diameter 6.35 cm were mounted on the topface of the truck-trailer model with the front one forming the trailer's leadingedge while the rear one was positioned 25.4 cm downstream. The verticalorientation of each cylinder can be adjusted to three different heights as: 0 (thecylinder flush with the trailer top face), 0.635 cm and 1.27 cm. Differentcombinations of the cylinder arrangement resulted in nine cases. The testswere conducted with helical grooves and spline surface conditions whichappeared more promising. As before, each cylinder was driven by a 1/8 H.P.motor, with an optical sensor in conjunction with a slotted disk to monitor itsspeed. A feedback control system was used to maintain the speed at a desiredlevel.2.2.2 Tractor-trailer truck model with fencesA typical fence is a thin flat plate of width b 1, height hf and negligiblethickness. In the test-program, fences made of aluminum were used. To assessthe effectiveness of fences as a drag reducing "add on" device, a three-dimensional rectangular prism (B = 22 cm, H = 23 cm, L = 101.5 cm, d = 25.4cm) with four fences forming a square about the geometric center served as themodel. The coordinate system to position fences has its origin at the geometriccenter of the front face. The two fence variables bf and hf determine the heightand size of square frame. In the test-program, size of the square geometry and16height of the fence were varied systematically to arrive at an optimumconfiguration (1)1 = 0, 12.7, 15.24, 17.78, 19.05 cm; lif = 0, 0.85, 1.45, 2.2, 2.7,3.35, 3.9 cm).As the fence results with the three-dimensional prism showedencouraging trends, the next logical step was to assess their effectiveness inreducing drag of a tractor-trailer configuration. To this end, a 1/12 scale modelof the truck with H = 26 cm, B = 22.7 cm and L =128.4 cm was constructedfrom Plexiglas, with hydraulic diameter, d, equal to 31.14 cm. The coordinatesystem to identify position of the fences had its origin at the geometric centerof the front face of the trailer. The fence positions were varied systematicallyto arrive at an optimum (critical) configuration as indicated below:(a) Position of the horizontal fence 1 varied along the y direction toarrive at the critical orientation (yfl )cr.(b) With the fence 1 fixed at (yfi )cr, the critical position of thehorizontal fence 2, (Y/2)cr, was established.(c) With fences 1 and 2 held fixed at their critical locations, the twinvertical fences were symmetrically located about the yf axis toarrive at the critical position (xf)u leading to a minimum dragcoefficient.(d) The fences were extended both vertically and horizontally to thefront face edges to assess the influence of their extension.(e)^The effect of increase and decrease of the fence lengths was17assessed systematically to arrive at a configuration leading to amaximum reduction in the drag coefficient.Figure 6 shows typical test arrangement for boundary-layer control throughmomentum injection and tripping devices.2.2.3 Tractor-trailer truck model with a cylinder kitIdeally, introduction of MSBC through rotating cylinders should proceedwith design of the new generation of trailers. The boundary-layer controlsystem in this way can become an integral part of the trailer configuration.However, until the concept has demonstrated its economic potentialconvincingly through exhaustive road tests, such far-reaching design changecannot be expected. This is understandable as the investment commitmentinvolved is indeed significant.An alternative would be to implement the concept on existing trucks inservice, however, this would entail modification of the trucks to accommodatethe cylinder(s), drive and control systems, cylinder speed monitoringarrangement and feedback device, etc. Obviously, this may not be alwaysattractive. To alleviate the situation, it was decided to design a self-containedMSBC-kit, which can be attached to the trailer's front face judiciously tocapture most of the desirable influence of the boundary-layer control withoutmodifying the trailer.Figure 6. Schematic showing arrangement of rotating cylinders and fences for the boundary-layer control:(a) MSBC using twin cylinders;Figure 6. Schematic showing arrangement of rotating cylinders and fences forthe boundary-layer control: (b) fences for tripping of the boundary-layer.1920Two different self-contained kits, each containing a cylinder with a 1/8H.P. motor and the speed control system, were designed for model tests. Thekits have the same width as the trailer and occupy the space near the gapbetween the tractor and the trailer. Position of the kit with reference to thetrailer is adjustable in the vertical direction.The difference between the two kits is primarily characterized by theextent of the cylinder exposed above the top face of the kit. This also reflectson the surface exposed to the fluid stream. In the case of kit 1, the cylinder isinitially raised 12.7 mm above the kit's top surface. For kit 2, half the cylinderis exposed to the fluid stream. Details of the two kits are shown in Figure 7.2.3 Test Procedure for the Tractor-trailer Truck Model2.3.1 Wind tunnelThe trailer and truck models were tested in the boundary-layer windtunnel at the University of British Columbia (Figure 8). The tunnel is an open-circuit type powered by an 80 kW three phase motor which drives an axial flowfan at a constant 700 rpm. The tunnel wind speed is varied using a pneumaticcontroller to alter either the rotating frequency of the fan or the blade pitch.The settling section contains a honeycomb and four screens to smooth the flowas it enters a 4.7 to 1 contraction section which accelerates the flow andimproves its uniformity. The tunnel has a test-section 2.44 m wide, 1.6 m highFigure 7. Kit configuration developed for application to existing trucks as an add-on device: (a) schematicdiagrams of kit 1 with the rotating element projecting 12.7 mm in the fluid stream;Figure 7. Kit configuration developed for application to existing trucks as an add-on device: (b) schematicdiagrams of kit 2 with the cylinder projecting 31.8 mm in the fluid stream;23Figure 7. Kit configuration developed for application to existing trucks as anadd-on device: (c) close-up showing a splined cylinder serving as themomentum injection unit and two vertical fences for tripping theboundary-layer with kit 2.2. 9x2. 9 mscreenTest - sect i on2. 44x1. 6 m1 honeycomb and 4screens 1 n 4x4 mset t I i ng sectionCont r act i onratio, ^4. 7: 24. 4 mPI-ModelFigure 8. The boundary-layer wind tunnel, with a test cross-section of 2.44x1.6 m, was used to study 1/12scale models of the trailer and truck configurations: (a) schematic diagram showing details ofthe tunnel with the model near the entrance to the test-section (first-bay);25Figure 8. The boundary-layer wind tunnel, with a test cross-section of 2.44x1.6 m, was used to study 1/12 scale models of the trailer and truckconfigurations: (b) photographs showing a tractor-trailer truck model(1/12), with boundary-layer control devices, being prepared for wind-tunnel tests.26and 24.4 m long consisting of eight 3.05 m long bays, and a variable height roofto allow for the boundary-layer correction. The stable wind speed of the tunnelis in the range of 2.5 - 25 m/s. The adjustable test-section roof was set for azero pressure gradient. The present set of experiments were carried out in thesecond bay which provided smooth flow with a turbulence lever less than 0.4%.The typical test Reynolds number based on the hydraulic diameter was 10 5 .2.3.2 Model support systemA typical truck model was supported by four steel guy wires which weresuspended from the ceiling and carried turnbuckles to help level the model. Asthe length of the wire (--: 145 cm) is much larger than the maximum horizontaldisplacement of the truck model 5 cm), the drag induced displacement wasessentially linear in the downstream direction.2.3.3 Cylinder rotation-rate measuring systemAs pointed before, the moving surface element used for the momentuminjection is a circular cylinder driven by a 1/8 H.P. variable speed D.C. motor.A system of pulleys with a belt connection transmits the motion from the motorto the cylinder. A slotted disk in conjunction with an optical sensor(photomultiplier) and an amplifier are used to measure the cylinder speed. The27amplifier signal was filtered and displayed on a DISA voltameter. The leastsquare fit of the results when calibrated using a digital stroboscope resultedin a linear plot of mV against the rpm.2.3.4 Drag measurement systemVariation in the drag due to the boundary-layer control devices beingrelatively small, required development of a sensitive transducer for itsmeasurements. The model was suspended from the ceiling by four wires asdescribed before to minimize frictional effects. The drag induced downstreammotion of the model was transmitted by an inelastic string to a cantilever beamwith a pair of strain gages near its root. The gages formed a part of theWheatstone Bridge (of the Bridge Amplifier Meter, BAM) and the amplifiedfiltered output was recorded using a DISA voltameter. The sensitivity of thedrag measurements was around 0.4 gm/mV. The Calibration of the cantileverusing static loads was performed, with the model suspended as described, twiceduring a test-session, before and after the tests; and the average calibrationvalue was adopted to account for any drift. Figure 9 shows schematically themodel support and drag measurement system.28Figure 5. 3-D tractor-trailer truck model test arrangement in the U.B.C.boundary-layer wind tunnel.293. RESULTS AND DISCUSSION3.1 Two Dimensional Wedge AirfoilThe lift and drag characteristics of an airfoil is significantly affected byits geometry, i.e., thickness and camber distribution. In order to focus on theeffect of momentum injection, the model selected for study was purposely takento be simple: a wedge shaped airfoil. The airfoil model is free of camber(symmetrical about the chord-line) and two dimensional. The amount ofinformation obtained through a systematic variation of the angle of attack,cylinder's surface condition and speed of rotation is really enormous. Influenceof the Reynolds number in the subcritical range 10 4 - 5x10 5 was found to benegligible. The results presented here for the wind tunnel tests conducted ata fixed Reynolds number of R n .---; 3x10 4 .The relatively large angles of attack used in the experiments result ina considerable blockage of the wind tunnel (.---. 20% at a = 55 ° ). The wallconfinement leads to an increase in the local wind speed at the location of themodel, thus resulting in higher aerodynamic forces. Several approximatecorrection procedures have been reported in literature to account for this effect.However, these approaches are mostly applicable to streamlined bodies withattached flow. A satisfactory procedure applicable to a "bluff body" offering alarge blockage and with separating shear layers is still not available.30With rotation of the cylinder(s), the problem is further complicated. Asshown by the pressure data and confirmed by the flow visualization [26], theunsteady flow can be separating and reattaching over a large portion of the topsurface. In absence of any reliable procedure to account for wall confinementeffects in the present situation, the results are purposely presented in theuncorrected form.To establish a reference which can be used to assess the influence ofcylinder rotation and surface condition, the first step was to obtain lift anddrag characteristics of the wedge-airfoil with the smooth stationary cylinder(U, / U = 0) forming its nose and the gap between the cylinder and rest of thewedge sealed. The results are presented in Figure 10 which also shows effectof the cylinder rotation. The reference configuration gave a maximum liftcoefficient CLmax -- 1.47 at a = 55°. However, with the cylinder rotation, slopeof the lift curve dramatically increase and the stall is delayed significantly. AtU,/ U -• 4 and a = 55° the peak lift coefficient reaches 3.95, an increase ofaround 168% ! The corresponding results for drag are presented in Figure 10b.It is of interest to recognize that, in general, the drag coefficient also shows afavourable trend. With an increase in speed of the cylinder, there is a distinctdrop in the drag coefficient at a given angle of attack. For example, at a = 55°,the decrease in CD from -- 1.85 for U,/ U = 4 to the reference value of 1.18corresponds to a reduction of 36% ! Of course, one way to assess effectivenessof the momentum injection in controlling the boundary-layer separation would5^10^15^2 0^2 5^30^35^40^45^50^55a 0Figure 10. Aerodynamic coefficients for a two-dimensional wedge airfoil as affected by the rotation of thesmooth cylinder: (a) lift coefficient;Smooth Cylinder--2:1-Uc/U = 0Uc/U = 1.1Uc/U = 2.1Uc/U = 3.1Uc/U = 4.2Rn = 3x1042.50.511.5CD0^5^10^15^20^25^30^35^40^45^50^55a °Figure 10. Aerodynamic coefficients for a two-dimensional wedge airfoil as affected by the rotation of thesmooth cylinder: (b) drag coefficient;CLCD0^5^10^15^20^25^30^35^40^45^50^55a °Figure 10. Aerodynamic coefficients for a two-dimensional wedge airfoil as affected by the rotation of thesmooth cylinder: (c) CL/CD.34be to study the variation of CL / CD with a as affected by the cylinder rotation.This is shown in Figure 10c. Note, in absence of the cylinder rotation (referencecase, U,/ U), the peak value of lift to drag ratio, (CL / CD )max, is around 1.58 (ata = 20°). It attains a value of (CL / CD )max - 10.5 at a = 15° and U,/ U - 3, anincrease of around 580% !It seems logical that character of the cylinder surface roughness shouldimprove the efficiency of the process of momentum injection. Hence, asmentioned before, the experiments were carried out with five distinctlydifferent rough surfaces (Figure 4).For the helical-groove cylinder (Figure 11) at a = 55° the maximum liftchanges from CLmax - 1.98 to 4.58 with the speed of the cylinder increasingfrom Uc/ U = 0 to 4, an increase of around 130% (Fig. 11a). The correspondingdrag deduction is around 21% from the CDmax -• 1.8 down to 1.4 at the sameangle of attack (Figure 11b). It is interesting to note that now the (CL / CD)max..---, 16 at a = 15° and Uc / U --. 2 (Fig. 11c), an increase by a factor - 10.5(compared to the reference value of 1.58). This suggests that an optimumchoice of cylinder surface can improve the momentum injection and hence delaythe boundary-layer separation.To that end, the cylinder surface characterized by slotted squares(roughness squares) and splines running parallel to the cylinder axis appearedpromising. Figure 12-14 show some typical results for the three surfacecharacteristics: roughness squares, spline-1 and spline-2). The mechanism of10 15 205x Uc/U = 0--['- Uc /U = 1X Uc /U = 2—e-- Uc/U = 3-$ Uc/U = 4Rn = 3x10 4Helical-groove CylinderC ,25^30a °555035 40 45Figure 11. Variation of the aerodynamic coefficients with the angle of attack and helical groove cylinderrotation: (a) lift;555040 4520 3525 30a °10 155Uc/U = 1Uc/U = 2Uc/U = 3Uc/U = 4R n = 3x104CDFigure 11. Variation of the aerodynamic coefficients with the angle of attack and helical groove cylinderrotation: (b) drag;^ czcr)5^10^15^20^25^30^35^40^45^50^55a oFigure 11. Variation of the aerodynamic coefficients with the angle of attack and helical groove cylinderrotation: (c) CL / CD.38momentum injection in the two cases, roughness squares and splines, appearsto be fundamentally different. Square projections serve as large scaleroughness elements rendering the flow turbulent. Splines, on the other hand,act like turbine blades thus injecting momentum in a more direct way. Withthe roughness square case, the peak lift coefficient attained a value of 4.15 ata = 50° and U,/ U = 4 (compared to CLmax of 1.5 for the reference case) asshown in Figure 12a. In general, the drag coefficient also reduced as expected(Figure 12b). As before, the optimum performance appears to occur at a lowerangle of attack of a = 15° where CL I CD of around 22 is realized (Figure 12c),an increase by a factor of 13 !Effectiveness of the spline geometry as a momentum injection device isillustrated rather dramatically by the performance plots in Figure 13.Simultaneous increase in CL and decrease in CD leads to an optimum value ofa = 15° where CL /CD reaches a spectacular value of around 75 !Of course, this suggests that spline geometry itself should be studiedthoroughly through a systematic variation of the spline parameters: width,height, taper, number, etc. Obviously, this represents a project in itself andbeyond the scope and focus of the present study. However, to gain someappreciation, a cylinder with rounded splines was constructed as explainedbefore (spline-2). Although the further improvement in performance is onlymarginal with (CL /CD)max rising to - 80° at a = 15, it is indeed distinct anddefinite (Figure 14c). Such an increase in (CL / CD)max by a factor of around 525^10^15^20^25^30^35^40^45^50^55a oFigure 12. Plots showing effect of the cylinder surface roughness condition on the lift and dragcoefficient: (a) CL;CD0^5^10^15^20^25^30^35^40^45^50^55a °Figure 12. Plots showing effect of the cylinder surface roughness condition on the lift and dragcoefficient: (b) CD;25 30a °Figure 12. Plots showing effect of the cylinder surface roughness condition on the lift and dragcoefficient: (c) CL /CD .^ ,41-,0^5^10^15^20 35 40 45^50 5525^30a o5^10^15^20 35^40^45^50^55Figure 13. Influence of axial surface grooves on the momentum injection (spline-1) as reflected throughthe variation of lift and drag coefficients: (a) CL;5545 5035 4020 25^30a o10 155Uc/U = 0Uc/U = 1.1lic/U = 2.1Uc/U = 3.1CDFigure 13. Influence of axial surface grooves on the momentum injection (spline-1) as reflected throughthe variation of lift and drag coefficients: (b) CD;0 5 10 15 20 25 30 35 40 45 50 55a °Uc/U = 0U c/U = 1.1Uc/U = 2.1Uc/U = 3.1U c/U = 4.2Figure 13. Influence of axial surface grooves on the momentum injection (spline-1) as reflected throughthe variation of lift and drag coefficients: (c) CL/CD.CLUc/U = 0Uc/U = 1.1Uc/U = 2.1Uc/U = 3.1Uc/U = 4.2Rn = 3x10 4Spline-2 Cylinder5^10^15^20^25^30^35^40^45^50^55a °Figure 14. Effect of the surface roughness designated as spline-2 on the aerodynamic coefficients: (a) CL;2.5Spline-2 Cylinder1.5 Uc/U = 0Uc/U = 1.1Uc/U = 2.1Uc/U = 3.1Uc/U = 4.2Rn = 3x10 40.51^i^1^I^1^I^15^10^15^20^25^30 35^40^45^50^55a °Figure 14. Effect of the surface roughness designated as spline-2 on the aerodynamic coefficients: (b) CD;CD25 30a °35 40 45^50 555^10^15^20Figure 14. Effect of the surface roughness designated as spline-2 on the aerodynamic coefficients: (c)CI, 1CD.^ )4=•---148(compared to the reference case) can be used to advantage by the nextgeneration of highly manoeuvrable airplanes.Figure 15 summarizes the results concerning CLmax ,) C Dm i n and(CL /CD )max as affected by the cylinder surface roughness. It is interesting tonote that although the maximum lift coefficient is associated with higher a (a= 50° or 55°, Figure 15a) the (CI, /CD )max corresponds to a in the range of 15°to 20° (Figure 15c). Of course, this is because of the drag characteristics asshown in Fig. 15b. It is useful to recognize that before differences in shape ofthe splines, as with the spline-1 and spline-2 cases, have relatively little effecton CLmax, CDmin and (CI, / CD )max •To have better appreciation as to the character of the flow in presenceof the MSBC, a flow visualization study was undertaken with the help and testfacility of Professor T. Yokomizo*. The tests were carried out in a waterchannel (Figure 16) using slit lighting with polyvinyl chloride particles servingas tracers. The flow visualization pictures were taken over a range of modelorientation, cylinder surface condition and speed. A video was also taken. Thetest Reynolds number, based on the free stream speed and chord length wasaround 3x104 . Although this is close to the wind tunnel test condition, it is* Dr. T. Yokomizo, Professor, Department of Mechanical Engineering, KantoGakuin University, Mutsuura, Kanazawa, Yokohama, Japan 239.0Conventional Smooth Helical Grvs. Rough345Spline-1^Spline-2CLmax21uciu 0, a. 550^uc iu . 4, a. 550^uctu 4, a. 550uciu 4, a. 5o°^uciu . 4, a. 55°^uctu 4, a. 55°Figure 15. Comparison charts showing relative merit of the surface condition on the boundary-layercontrol: (a) CLmax ;ccuciu = o, a= 00uciu = 4, a= 15°uciu = 1, a= 5°uciu = 4, a = 15°Uc/U = 2, a= 15°uciu = 4, 0(= 15 °0.160.12CDmin0.080.040Conventional Smooth Helical Grvs. Rough^Spline-1^Spline-2Figure 15. Comparison charts showing relative merit of the surface condition on the boundary-layercontrol: (b) CDmin;^ uciu o, a = 20°ticiu = 4, a = 15° MINERMIR= Uc/U = 3, a = 15°Uc/U = 4, a = 15 °Uc/U = 2, a = 150uciu = 4, a= 15°80(CLCD max4004R n = 3 x10Uc4111%.%■.. _ _ _ ... ... •41\ .f7-7W-M...m.rm.........rammi.•.........Conventional Smooth Helical Grvs. Rough^Spline-1 Sp ine-2Figure 15. Comparison charts showing relative merit of the surface condition on the boundary-layercontrol: (c) (CL/Cdmax .Figure 16. A schematic diagram of the closed circuit water channel facility used in the flow visualizationstudy. Slit lighting was used to minimize distortion due to three dimensional character of theflow. Long exposure provided path-lines with polyvinyl chloride particles serving as tracers.The dimensions are in mm.53quite different from the real life situation. Hence the results should beconsidered only qualitative in character, however, they do show effectivenessof the moving surface boundary-layer control procedure.Figure 17 shows typical pictures of flow past the wedge airfoil at a = 30°as affected by the momentum injection. Effectiveness of MSBC is strikinglyapparent even at such a high angle of attack. In fact the concept continues tobe effective even for a as high as 55° (Figure 18) !3.2 Tractor-trailer Truck Model with Twin CylindersBased on the 2-D wedge airfoil results the concept of the MSBC has beenestablished and it was decided to apply this idea to a typical bluff body, atractor-trailer truck, in a real life. The 2-D wedge airfoil data showed theimportance of cylinder roughness in improving efficiency of the momentuminjection process and associated reduction in the drag. Therefore, it seemedreasonable to introduce the momentum more directly to the tractor-trailertruck model. This was achieved in several ways:(i) Modify the cylinder surface by coating it with sand particles(sandpaper of grades 80 and 40).(ii) Provide increased cylinder surface roughness through helicalgrooves or splines running parallel to the cylinder axis.(iii) Keep one cylinder at the top leading edge of the trailer (referreda = 30° ,^RT., = 3 x 104^ 54II, / U =U, / U =U, / U =Figure 17. Typical flow visualization photographs for a wedge shaped airfoil,with a smooth surface cylinder, showing remarkable effectiveness ofthe MSBC concept for a = 30°, R n = 3x104. Note the progressivedownstream shit of the separation point as the t/c / U increase.Eventually the flow appears to approach the potential character./ U/ U =/ U =a = 55° ,^RT, = 3 x 104^55Figure 18. The concept of moving surface boundary-layer control continues to beeffective enen at a high angle of attack of 55° as shown dramaticallyby these flow visualization pictures (Rn = 3x104 )56to as the front cylinder) and locate the second cylinder (rearcylinder) at an optimum distance downstream. Objective is toinject additional momentum in the boundary-layer to compensatefor dissipation of the momentum introduced by the front cylinderand thus counter the emergence of adverse pressure gradient.(iv) Raise the cylinders so as to immerse them in the boundary-layerand assess the effect of cylinder orientation.Tests with a 1/12 scale model of the truck were carried out in theboundary-layer tunnel with negligible blockage effect (blockage ratio = 1.2%).The trailer was provided with rotating cylinders at its top leading edge anddownstream locations. The LIH ratio for the trailer was approximately 3.75which suggested that rotation of the trailing edge cylinder will have virtuallyno effect on the drag reduction. The wind tunnel tests substantiated thisobservation. Figure 19 shows variation of CD with the cylinder speed rationUe / U for three cases: cylinder with smooth surface; cylinder surface roughnessof grade 80; and cylinder surface roughness of grade 40. In absence of themomentum injection (U,/ U = 0), the truck drag coefficient is around 0.81 andreduces to 0.765 at U,/ U = 2 for the smooth cylinder case. The surfaceroughness of the cylinder improves the performance further reducing CD toaround 0.73 at U,/ U = 2.1 for the roughness grade of 80. Increasing thesurface roughness to 40 drops the minimum CD to 0.7, a reduction of around13%.10.850.900.80 a --e—,, • --•-- •^• •^ lio- ---,^• - •"^C D0.75 -Uc/6• M.,/6•■ SmoothO Roughness 80 grade• Roughness 40 grade0.70 -■.•-• -• -•0.65 -UC/UFigure 19. Effect of the moving surface boundary-layer control on the drag coefficient of a tractor-trailertruck configuration. Note, an increase in the sandpaper roughness contributes to the dragreduction through efficient momentum injection.0.60 ^0.0 0.2^0.4^0.6^0.8^1.0^ 1.6^1.8^2.0^2.21.2^1.4 2.4^2.658For the subsequent studied in (ii), (iii) and (iv) a different model of thetractor trailer configuration was used (also 1/12 scale). It had corresponded tothe configuration selected for future road-tests. The cylinder orientationsstudied with helical and spline set of cylinders are indicated in Table 1.Figure 20 shows effect of the twin cylinders with helical grooves. At theoutset it is apparent that the front cylinder rotation reduces the drag as before,however, character of the plot is rather different. There is a monotonicreduction in the drag coefficient with Ucf/ U (Ucr / U = 0). Although the rearcylinder rotation further diminishes CD, the reduction is relatively small.Raising the cylinders does improve efficiency of the momentum injection,however, the maximum drag reduction attained was around 14.3% (Case 2,Figure 20b), not much different than that given by the grade 40 sandpaper(Figure 19).On the other hand, the spline cylinder reduces the drag coefficientdramatically. To begin with, it should be recognized that the base dragcoefficient of the truck with flush mounted spline cylinders in a absence ofrotation is higher than before (1.12 against 1.015 for the helical groovecylinders, Case 1). Note, with the spline cylinder raised and with only the frontcylinder rotating (Ucr/ U = 6.1), the drag coefficient reduces by about 37%(Figure 21f). With the rear cylinder rotation, the drag reduction jumps to 52%(Case 6, Ucf/ U = 6.1, Ucr / U = 5.7). Even with the speed ratio of 4, the CDreduced by around 26%. Thus the spline geometry with raised position of theTable 1.59Wind tunnel tests conducted with different orientation of the twinhelical-groove and spline cylinders: location of the front cylinderwas at the top leading edge and the rear cylinder at 25.4 cmdownstream from the leading edge.CaseCYLINDER ORIENTATIONFront Raised, mm Rear Raised,mm1 — —2 — 6.353 — 12.74 6.35 12.75 6.35 —6 12.7 —7 12.7 6.358 12.7 12.79 6.35 6.35*BAIN r.y.^• •0.9 -)1( Uc f/U = 0^Uct/U = 1.4 Ucf/U = 4.2A Uc f/U = 5.9^Rn = 10 5^CASE 1 (HELICAL)(JO Ucr0.)—L9)^1.05 -0.85 ^0 1^2^3^4^5^6U crFigure 20. Effect of the twin helical cylinder configuration on the momentum injection and boundary-layer control: (a) Case 1: both cylinders flush;1.1CD4 51^2^3Ucr1.10.85 ^0 6CD0.950.9 -1.05 -)I( Uc f/U = 0^—R-- Ucf /U - 1.4 Ucf/U = 4.2A Ucf/U = 5.9^Rn = 10 5^CASE 2 (HELICAL)Figure 20. Effect of the twin helical cylinder configuration on the momentum injection and boundary-layer control: (b) Case 2: front cylinder flush, rear cylinder raised 6.35 mm;0.85 ^0Figure 20. Effect of the twin helical cylinder configuration on the momentum injection and boundary-layer control: (c) Case 3: front cylinder flush, rear cylinder raised 12.7 mm;1 2^3Ucr /U4 5 61.05 -1•• • •0.9 -CD1.1^ ucf/u • 0^ucf/U = 1.44^0 Uc f/U • 4.20A Uc f/U • 5.87^Rn • 10 5^CASE 3 (HELICAL)0.85 ^0Figure 20. Effect of the twin helical cylinder configuration on the momentum injection and boundary-layer control: (d) Case 4: front cylinder raised 6.35 mm, rear cylinder raised 12.7 mm;)1(^Ucf/U = 0A Uc f/U = 5.9Ucf/U = 1.4 —0— Ucf /U = 4.2Rn = 10 5^CASE 4 (HELICAL)61 4 52^3Ucr /U1.1C o••^ •0.9 -1.05 - )t( Uc f/U = 0^H± Ucf/U = 1.44 0 Uc f/U = 4.20-74\- Uc f/U = 5.87^Rn = 10 5^CASE 5 (HELICAL)• .2 :11 111.05 -CD0.85 ^0 1^2^3^4^5^6UcrFigure 20. Effect of the twin helical cylinder configuration on the momentum injection and boundary-layer control: (e) Case 5: front cylinder raised 6.35 mm, rear cylinder flush;0.85 ^0 5^61 42^3Ucr /U1.11CD0.9 -1.05XUcf/U • 0 Ucf/U z 1.44 ^ Ucf/U • 4.20f\ Ucf/U z 5.87^Rn = 10 5^CASE 6 (HELICAL)Figure 20. Effect of the twin helical cylinder configuration on the momentum injection and boundary-layer control: (f) Case 6: front cylinder raised 12.7 mm, rear cylinder flush;0.8501.05 -161 4 52^3Ucr /UUcf Ucr00^01.10.95 -0.9 -IN 01:‘ I I I MI I MST.Cp)1( Uc f/U = 0^—R— Ucf /U = 1.44^Ucf/U • 4.20^ Uc f/U • 5.87^Rn • 10 5^CASE 7 (HELICAL)Figure 20. Effect of the twin helical cylinder configuration on the momentum injection and boundary-layer control: (g) Case 7: front cylinder raised 12.7 mm, rear cylinder raised 6.35 mm;0.8501.05 -11 62^3Ucr /U4 51.1coUcf Ucr,0.9 -eizE Wifel I I I • • r. •Co0.95Ucf/U = 0 Ucf/U = 1.44 0 Ucf/U = 4.20A Uc f/U = 5.87^Rn = 10 5^CASE 8 (HELICAL)Figure 20. Effect of the twin helical cylinder configuration on the momentum injection and boundary-layer control: (h) Case 8: front cylinder raised 12.7 mm, rear cylinder raised 12.7 mm;0.85 ^0 5^642^3Ucr11.1••• ''1.05 -CD)I( Ucf/U = 0 ucf/u = 1.44 ^ Ucf/U = 4.20^ Uc f/U = 5.87^Rn = 10 5^CASE 9 (HELICAL)Figure 20. Effect of the twin helical cylinder configuration on the momentum injection and boundary-layer control: (1) Case 9: front cylinder raised 6.35 mm, rear cylinder raised 6.35 mm.69cylinder(s) appears quite promising in reducing pressure drag of the tractor-trailer truck configuration through MSBC (Figure 21).Both the truck configurations were subjected to extensive flowvisualization study to get better appreciation of the flow particularly withreference to stagnation, separation reattachment and wake condition. The testswere carried out in a water channel using slit lighting with polyvinyl chlorideparticles serving as tracers as mentioned before. The test Reynolds numberbased on freestream speed and trailer height (H) was around 4x10 4 .The flow visualization pictures were obtained over a wide range ofcylinder orientations and speeds. For brevity, only a typical set of results fortwo cylinders at front face of the trailer are presented here for two differentcab geometries (Figures 22, 23).Figure 22 considers a truck configuration where the trailer projectssignificantly higher than that in Figure 23. The gap between the cab and thetrailer is also relatively large. In absence of the cylinders, the flow separatedat the top leading edge of the trailer and a large bubble covers the top face. Along wake was also observed which is partially visible. Rounding of the cornersby the cylinders, even in absence of their rotation, reduces the size of theseparation bubble. Note, even with L/c / U = 1, the flow on the top face isessentially attached. With E/c / U = 3, the flow on both top and bottoms faces ofthe trailer is rather smooth and there is a significant reduction is the size ofthe turbulent wake.0.4 ^0Figure 21. Variation of the drag coefficient CD with the speed ratio for the twin spline cylinderconfigurations: (a) Case 1: both cylinders flush;1 2^3Ucr /U4 5 6)1( Uc f/U = 0^—El— U cf/U = 1.4^Ucf/U = 2.7Uc f/U = 4.1^A Uc f/U = 6.1^Rn = 10 51.41.2 -CASE 1 (SPLINE)Cp. ff: Tel I I MI rol".0.40 1 2^3Ucr /U4 5 6CpUcf UcrO0.6 -• ma r•Y•11111111111rif•1.2 -)1( tic f/U = 0^Ucf/U = 1.4^-4<— Ucf/U = 2.70 Uc f/U z 4.1^A Uc f/U = 6.1^Rn = 10 5CASE 2 (SPLINE)1.4Figure 21. Variation of the drag coefficient CD with the speed ratio for the twin spline cylinderconfigurations: (b) Case 2: front cylinder flush, rear cylinder raised 6.35 mm;0.40 1 2^3Ucr /U4 5 6CASE 3 (SPLINE)1.4CD0.8Ucf Ucrcw10 00.6 - U • m.Y.^.T.1.2 -)I( ucf /U z 0^Ucf/U= 1.4^X Uc f/U = 2.70 Uc f/U = 4.1^A Uc f/U = 6.1^Rn = 10 5Figure 21. Variation of the drag coefficient CD with the speed ratio for the twin spline cylinderconfigurations: (c) Case 3: front cylinder flush, rear cylinder raised 12.7 mm;0.4 ^0Figure 21. Variation of the drag coefficient CD with the speed ratio for the twin spline cylinderconfigurations: (d) Case 4: front cylinder raised 6.35 mm, rear cylinder raised 12.7 mm;1 2^3Uc r /U4 5 6^ Ucf/U = 0^Ucf/U = 1.4^—><— Uc f/U = 2.7Uc f/U = 4.1^A Uc f/U = 6.1^Rn = 10 5CASE 4 (SPLINE)1.4CD-2 roY0‘1111.Y.1.2 -1 2 3Uc r /U4 5 6)I( Ucf /U = 0^-÷}-- Ucf/U = 1.4^>< Uc f/U = 2.7—le— Uc f/U = 4.1^A Uc f/U = 6.1^Rn =10 5CASE 5 (SPLINE)CDFigure 21. Variation of the drag coefficient CD with the speed ratio for the twin spline cylinderconfigurations: (e) Case 5: front cylinder raised 6.35 mm, rear cylinder flush;4 65CASE 6 (SPLINE)1.41.2 -0.8 -• •2^3Ucrc)( 1.10/U = 0 Ucf/U = 1.4^—) Uc f/U = 2.70 Uc f/U = 4.1^A Uc f/U = 6.1^Rn = 10 5Figure 21. Variation of the drag coefficient CD with the speed ratio for the twin spline cylinderconfigurations: (f) Case 6: front cylinder raised 12.7 mm, rear cylinder flush;0.4 ^0 1 2^3Ucr /U4 5 6CASE 7 (SPLINE)1.40.6U■■►•CD0.81.2 -)I( Ucf/U 0^lion) = 1.4^Ucf/U = 2.7Uc f/U = 4.1^A Uc f/U = 6.1^Rn = 10 5Figure 21. Variation of the drag coefficient CD with the speed ratio for the twin spline cylinderconfigurations: (g) Case 7: front cylinder raised 12.7 mm, rear cylinder raised 6.35 mm;5^6Figure 21. Variation of the drag coefficient CD with the speed ratio for the twin spline cylinderconfigurations: (h) Case 8: front cylinder raised 12.7 mm, rear cylinder raised 12.7 mm;1 40.4 ^0 2^3Ucr /UCASE 8 (SPLINE)1.4C D1.2 -)I(^Ucf / U a 0^Ucf / U M 1.1^UCf / U MI 2.5licf / U - 3.8 A licf / U = 5.4^Rn = 10 5U1 2^3Uc r /U4 5^6CASE 9 (SPLINE)0.6 -I 1.41.20.40Cp^ Ucf/U = 0^—1' Uc f/U - 1.1^X Uc f/U = 2.5Uc f/U = 3.8^A Uc f/U = 5.4^Rn = 10 5Figure 21. Variation of the drag coefficient CD with the speed ratio for the twin spline cylinderconfigurations: (i) Case 9: front cylinder raised 6.35 mm, rear cylinder raised 6.35 mm.79Uc/U =Uc /U =Uc/U =Figure 22. Effect of cylinder rotation on the flow past a tractor-trailer truckconfiguration.Uc/U =Uc/U =Uc /U =Uc / U =Figure 23. Effect of cylinder rotation on the flow past a tractor-trai er rucconfiguration. Note, the tractor geometry and spacing between thetractor and trailer are different here.81Figure 23 considers a truck configuration where trailer projects over thecab by a relatively small amount. Note, the character of the flow issignificantly different compared to the previous case. The size of the separationbubble on the trailer's top face is rather small (with cylinder, i.e. smoothcorners, but Ue l U = 0), however, there is a large turbulent field on the bottomface. This results in a rather wide wake. The cylinder rotation progressivelyimproves the situation as Ue / U increases, and for Uc / U = 3 there is anessentially attached flow on topo and bottom faces of trailer with a dramaticreduction in the wake-size.3.3 Rectangular Prism with FencesAs against the momentum injection to delayer boundary-layerseparation, fences located on the front face of a bluff body tend to trip theboundary-layer thus creating a low pressure region by interrupting thepressure recovery. This, in turn, can reflect as a reduction in the drag of thebody. To assess effectiveness of the fences as a drag reducing device acomprehensive wind tunnel test-program was undertaken.As pointed out before, typical fence is a thin flat plate of width (b1),height (h f), and negligible thickness. The fences were mounted normal to thefrontface. The fence dimensions are presented in a nondimensional form (d,hydraulic diameter of the projected area normal to the flow).82Figure 24 shows variation of CD with the geometry of the fence for athree-dimensional rectangular prism (H = 22.8 cm, B = 21.5 cm, L = 101.5 cm,d = 25.4 cm). Four fences were mounted on the front face of the prism to forma square about its geometric center. Note, there appears to be an optimumheight as well as width of the fences. The maximum reduction in CD obtainedwas from 1.25 to 0.85 (31%) for hf/d = 0.13 and bf/d = 0.5.3.4 Tractor-trailer Truck Model with FencesWith this encouraging trend, it was decided to explore application of thefences on the front exposed area the trailer (Figure 25). Of course, as part ofthe area is shielded by the cab, the available exposed space for mounting fenceson the trailer is relatively small. The trailer dimensions in this case were H =25 cm, B = 22.6 cm, and L = 128.4 cm with a hydraulic diameter of 31.14 cm.Based on the earlier results on the fence height and width, hid and bidwere fixed at as 0.087 and 0.683, respectively, and an optimum location of ahorizontal fence (fence 1) in the vertical direction (i.e. along y 1) was searched(Figure 26). The critical position was given by (yfi /d)cr --- 0.36 leading to aminimum CD = 1.048. When fence 1 held at its critical location, similarprocedure was applied to fence 2 to arrive at its optimum location (Figure 27).For the reference drag coefficient without fences, CD°, at 1.124, the twohorizontal fences in their optimum location resulted in a drag reduction of>K b f id = 0.501^ bf /d = 0.724 bf /d = 0.601^X bf /d = 0.701R n = 10 51.41.30.8 ^0 0.04^0.08^0.12^0.16CD1.10.9h f /dFigure 24. Effect of the fence width and height on the drag of a three-dimensional prism.Figure 25. A schematic diagram showing application of the fences on the front, exposed face of the trailerof a truck.^ 0041.04 ^0.29 0.33^0.35 0.390.370.311.12 ^1 .1C DC00 = 1.124h f /d = 0.087b f /d = 0.6831.081.06Yfi IdFigure 26. Variation of the drag coefficient with the position of horizontal fence 1.0.08 0.120^0.04Yf2 IdFigure 27. Variation of the drag coefficient with the position of horizontal fence 2 when fence 1 is fixedat its critical orientation.CDO = 1.124h f /d = 0.087b f /d = 0.683(Yfi id)cr = 0.359R n = 1.2X10 51.0351.03C D1.0251.021.015-0.08 -0.04y f /d(CD)min(Yf2 /d)crI^ IYfi /d)cr^► x f /d87around 10.6%. Introduction of vertical fences improved the situation further(Figure 28) giving, for optimum locations of the four fences, CD 0.9, areduction of 20% !It was felt that by fine-tuning of the fence dimensions furtherimprovement in their performance can be achieved. The improved fenceconfiguration and the associated drag coefficient are shown in Figure 29. Note,the horizontal fences are extended over the entire width of the trailer's frontface. The vertical fences are also longer and reach to the top edge. The fencearrangement lowered CD to around 0.85, a reduction from the reference valueby 24.6% !3.5 Tractor-trailer Truck Model with both Twin Cylinders and FencesIt was tempting to explore a hybrid combination of the momentuminjection in presence of the fences to have some appreciation as to the possiblefavourable trends. To that end the previous tractor-trailer with twin rotatingcylinders was used. The optimum configuration of the vertical fences and thecylinder orientations (spline cylinder, Case 6) were used in the study. Thehorizontal fences were avoided as they would create a turbulent flow field justin front of the leading edge cylinder. The results are presented in Figure 30.The cylinder rotation in presence of the vertical fences improves theperformance only a little. For Ucf 4 and Ucr = 0, the drag coefficient changes0.23^0.27^0.31^0.35Xf3,4 /dFigure 28. Variation of the drag coefficient with the position of twin vertical fences 3 and 4 when fences1 and 2 are fixed at their critical orientations.3 yf /d 4A •1H/dh f /d = 0.087B/d = 0.729H/d = 0.841RI = 1.2x10 5(yfl /d)cr= 0.359(Yf2 /d)cr = -0.048(x f1 , 2 /d)cr = 0.285CD = 0.861•x f /d2B/dFigure 29. Stages in fine tuning of the fence dimensions and associated drag coefficient: (a) The fenceswere extended to the edges of the trailer front face in both horizontal and vertical directions.••H/dh f Id = 0.087B/d = 0.729H/d = 0.841Fi r, = 1.2x10 5(Yfi /c)cr = 0.359(Yf2 /Wu = -0.048(xf1,2/d)cr = 0.285CD = 0.848B/d3^yf Id^41x f /d2Figure 29. Stages in fine tuning of the fence dimensions and associated drag coefficient: (b) The twinvertical fences were truncated to horizontal 2 fence position from the bottom edge of thetrailer front face.3^ yf /d^ 4• •1x f /dH/d2h f /d = 0.087B/d = 0.729H/d = 0.841R I, = 1.2x10 5(Yfi id)cr = 0.359(Yf2 id)" = -0.048(xf1,2/d)cr = 0.285CD = 0.897•B/dFigure 29. Stages in fine tuning of the fence dimensions and associated drag coefficient: (c) The twinvertical fences were truncated further to horizontal 1 fence position from the top edge of thetrailer front face.H/d•3^yf id^4h f /d = 0.087^(Yfi id)cr = 0.359B/d = 0.729^(Yf2 /d) cr = -0.048H/d = 0.841^(Xft2 /d)cr = 0.285R n = 1.2X10 5^CD = 0.9011X f /d2B/dFigure 29. Stages in fine tuning of the fence dimensions and associated drag coefficient: (d) The twinhorizontal fences had offset from both side edges of the trailer front face.1.21)I(^h f /d = 0^ Uc f/U = 3.04Uc f/U = 5.69hf/d = 0.08A U c f/U = 3.94bf /d = 0.47>< ^= 2.240 Uc f/U = 4.80xf /d = 0.29R n = 1.2x10 5CD0.8h f —01 14--Ue f Ucr0 00.6Tow side fences• wa I I w.T.1.2^2.4^3.60.40 4.8 6Ucr /UFigure 30. Variation in the drag coefficient with the cylinder speed ratio for a hybrid configurationinvolving two vertical fences and twin rotating cylinders in their optimum geometry.94from 1.12 to 0.86 (as against 0.885 for the no fence case). This corresponds toa reduction in the drag coefficient by 23%. Note, rotation of the rear cylinderimproves the situation only marginally.3.6 Tractor-trailer Truck Model with a Cylinder KitWith effectiveness of the MSBC and boundary-layer trip devicesestablished, the attention was directed towards their practical implementation.Of course, application of fences presents no problem. They are merely plateelements, judiciously oriented on the front exposed face of a trailer. Beingentirely passive in character, no power input units are involved. They provide18-24% reduction in the drag coefficient depending on the geometry of thefences and available frontal exposed area of the trailer. Obviously this is arather attractive option for drag reduction of trucks.On the other hand, the MSBC did provide a spectacular reduction in thepressure drag and one would like to exploit it in practice. Ideally, it should beintegrated with the next generation of trailer designs. However, acceptance ofany new idea usually takes time. Considering the history of the truck industry,in terms of its reluctance to change and passion for add-on devices (deflectors,shrouds, skirts, etc.), a simple approach to implementation of the MSBCconcept on existing trucks will have to be explored. There is also the questionof payload volume lost due to presence of the rotating element at the topo front95edge of the trailer. Also, installation of the power unit and transmission systemwill have to be considered.These consideration led to the development of the "cylinder-fence kit"idea. The kit, essentially an add-on device, consists of a housing containing arotating cylinder with power (motor/electric or hydraulic) and transmissionsystems. The flat upstream surface of the housing permits attachment of fencesthus resulting in a hybrid drag reducing device. The kit can be mounted readilyon the front face of the existing trailer, without modifying the top leading edgeor paying penalty in terms of the lost cargo space. It was obvious that, with themomentum injection unit no longer an integral part of the trailer, theboundary-layer control effectiveness will suffer to some extent. However, if thepenalty is not severe, the kit idea may find wide acceptance with the existingtrucks. It may also lead to a small kit manufacturing industry in Canada.The idea evolved into the development of two slightly differing kitconfigurations (Figure 7) and assessment of their effectiveness throughextensive wind tunnel tests. The main difference between the two kits is theextent to which the cylinder projects into the fluid stream. In the first case thecylinder immersion is to an extent of 12.7 mm, while in the second case thecylinder is immersed up to its center (i.e. radius, 1.25 in. = 31.8 mm). Onlysome typical results useful in establishing trends are recorded here.3.6.1 Cylinder kit 196The kit fits in the gap between the tractor and the trailer, and isadjustable in the vertical direction to arrive at an optimum position.Furthermore, it is provided with a removable cover in the front (upstreamdirection) where the fences, when used, are attached. The wind tunnel testswere aimed at assessing the effect of cover, vertical location of the kit asspecified by Hk, fence geometry and the cylinder rotation. The resultspresented in Figure 31-33.Effect of vertical orientation of the kit, in absence of fences and thecover, on the drag coefficient as affected by the MSBC is assessed in Figure32a. Four different values of Hk /d (0.12, 0.08, 0.06 and 0.04) were used in thetests. Note, for Hk /d = 0.12 and the cylinder stationary, the drag coefficient isa minimum at 1.027. This is because of the favourable alignment between thetop of the cylinder and top leading edge of the trailer. However, with themomentum injection the pattern changes and Hk /d = 0.08 configurationappears to be more favourable. For Uc / U = 4, CD 0.935, which represents areduction of 12.8% from the stationary value of 1.077. It is apparent that theefficiency of the momentum injection process has suffered primarily due to:(a) spacing between the cylinder and the trailer leading edge;(b) turbulent flow created due to absence of the cover, which exposesthe motor and belt drive system to the fluid stream.Figure 31b evaluates the kit performance with the cover in place. Note,the positive influence of the cover. Now the drag coefficient for U c / U = 4 isHk1.15NO FENCE, NO COVERC DA Hk/d = 0.12Hk/d = 0.08)1( H k/d = 0.06HE— Hk/d = 0.0413 c/d = 0Hc/d = 0Rn = 10 5 ma IMO MIMI MR.0.75 ^0 1^2^3^4^5^6Lic /UFigure 31. Variation in the drag coefficient with the cylinder speed ratio as affected by kit 1 verticalorientation: (a) absence of the front cover;1.05WITH COVER, NO FENCECDHk/d • 0.08>K Hk/d 0.06Hk/d • 0.0413c/d = 0.73Fick! = 0.41Rn = 10 50.750.65 ^0UCoverMBMT= 111111 IN.Y.ucHk4^5^61^2^3UciUFigure 31. Variation in the drag coefficient with the cylinder speed ratio as affected by kit 1 verticalorientation: (b) with the front cover.^ coor)xf/d = 0.29^R n = 1.2x10 5- Hk/d = 0.08^ Hk/d = 0.06- Hk/d = 0.0413c/d = 0.73^hf/d = 0.08Hc /d = 0.41651 40.55 ^0 2^3Lic /U0.95WITH COVER AND TWO VERTICAL FENCESHe = bf11►CD0.850.750.65THkT.^ST.Figure 32. Variation in the drag coefficient with the cylinder speed ratio and kit 1 position with twovertical fences mounted optimally on the front face of the kit.hfU CTHk1.2Bc/d z 0.73Hk/d z 0.06•AftMNff:d;A■4141114MEIlowhf/d z 0.08^Rn 1.2x105xf Id z 0.29 TWO VERTICAL FENCESAMINIv NEBAl•CD0.8 -0.6 -0.40A Hc/d z 0He/d= 0.29, bf /d z 0Hc/d z bf /d z 0.29zHc/d z 0.41, bf /d z 0 He bfX Hc/d bf /d z 0.41^U1^2^3Uc /UA4^5^6Figure 33. Variation in the drag coefficient with the cylinder speed ratio for kit 1 at the optimum verticalorientation. Note the influence of cover and fence geometry.1010.85 for Hk /d = 0.06, a reduction of 16.2%.Effect of two vertical fences placed on the cover was evaluated next asshown in Figure 32. Note, even in the absence momentum injection (E/ c / U =0), the drag coefficient for Hk Id = 0.06 reduces from 1.014 (no fence case) to 0.9(with two vertical fences), a drop of 11.2%. With cylinder rotation at Uc / U =4, the CD = 0.72 suggesting the reduction in aerodynamic resistance by 29% !Note, the drag coefficient is relative insensitive to the kit positions tested.Thus such hybrid combination of the MSBC together with trip fences in a kitform appears quite promising for the prototype application.It is apparent that larger the frontal area of the kit, more effective canbe the fences. Of course, the fence geometry itself plays an important role ineffective tripping of the boundary-layer and arresting the pressure recovery.Hence, it was decided to study influence of the exposed height (He) of the coveravailable for mounting fences and the fence height (b1). The results arepresented in Figure 33. There are five distinct cases considered. Note, for r/c / U= 0:(i) the highest drag coefficient CD ----. 1.092 corresponds to the no covercase (He ld = 0);(ii) in absence of the fences, and the kit cover exposing 3/4 of thecylinder's frontal area, the drag coefficient drops to around 1.004(He /d= 0.29, bf/d = 0);(iii) introduction of twin vertical fences on the cover configuration102corresponding to (ii) leads to a drag coefficient of 0.96 (11c l d =bfI d = 0.29);(iv) with the cover extending to the center point of the cylinder (thecylinder half exposed to the free stream) and with no fences, thedrag coefficient is .--- 1.014, higher than that in case (ii) or (iii);(v) with the twin vertical fences mounted on the cover described incase (iv), CD reached the lowest value of 0.9 in this series ofstudies.Thus, even in absence of rotation, judicious selection of the fencegeometry and orientation can lead to a significant reduction in drag of acomplex bluff body like a tractor-trailer truck. Recognizing that the referencedrag coefficient (tractor-trailer without kit, cylinders or fences) is 1.124, areduction in aerodynamic resistance by 19.92% through entirely passive meansis indeed remarkable.As can be expected, with momentum injection, the performance improvesfurther with a reduction CD by 30.6% at Uc / U = 3 (L/c / U = bf/d = 0.41).3.6.2 Cylinder kit 2The cylinder-fence kit 2, being thinner, does not completely fill the gapbetween the tractor and the trailer. Of course, the gap can be covered by aplate if desired. As pointed out before, now the cylinder projects in the free103stream by an amount equal to its radius.Figure 34 studies the effect of gap and the kit's vertical position on thedrag coefficient without and with the cylinder rotation. Five different casesrepresenting the gap condition (covered or uncovered) and the kit location(Hk /d) are considered. It is apparent that, irrespective of the cylinder rotationspeed, covering the gap reduced the drag. Note, with the gap covered and thecylinder center aligned with the trailer's top face (Hk /d = 0), the drag isminimum over the entire range of tic / U studied. However, the drag reductionobtained was significantly less than that obtained with kit 1 undercorresponding condition (Figure 31b).Figure 35 presents results for the kit performance as affected by twovertical plates in absence of cylinder rotation. Note, the gap condition and kitposition correspond to the optimum found in the previous study (Figure 34, gapcovered, Hk / d = 0). The critical position, (xf/d)cr, of the twin vertical fenceswas found to be 0.285 with (CD )min - 0.983, a reduction of around 8%(reference CD ,.--, 1.071).With the vertical fences at their critical positions, a horizontal fence wasintroduced and its optimum location established (Figure 36). It is apparent thata further reduction in the drag coefficient is relatively small. Now theminimum CD attained is around 0.952 with (yf/d)cr - 0.2, with corresponds toa drag reduction of around 11.1%. Thus the horizontal fence by itself reducesthe CD by approximately 3%.-----."'""iiiiir^ •Soo1111....- •1111^ NM1111E KIMAM.>i< Hk/d = 0.09, Gp/d = 0.08Hk/d - 0.09, Gap CoveredX Hk/d • 0.03, Gap Covered^ Hk/d = 0,^Gap CoveredA^ Hk/d • -0.04, Gap CoveredRn = 1.2x10 5 .Y.1.11.05CD0.950.850.75 ^0 1 2^3Uc iU54Figure 34. Variation in the drag coefficient with the cylinder speed ratio for kit 2 at different verticalposition. Note the effect of gap.Figure 35. The drag coefficient as affected by position of the two vertical fences on the front face of kit2, set at the optimum position, with the gap covered and stationary cylinder.0.310.27 0.29 0.331.031.021 .0 110.990.980.25CDxf /d0.18 0.20.950.16Gap covered^hf/d = 0.08^bfv/d = 0.47Hk/d = 0^(xf /d)crz bfh/d = 0.285^Uc/U = 0CD z 0.983 (bfh/d = 0)^Rn = 1.2x10 50.970.965CD0.960.955y f /c10.22 0.24Figure 36. Variation in the drag coefficient as affected by the position of a horizontal fence with twocritically-oriented vertical fences on the front face of kit 2. The gap was covered and thecylinder was held stationary.107The next logical step was to assess the effect of cylinder rotation withoutthe fences as well as with their optimum orientation (Figure 37). As can beexpected, the momentum injection improves the drag reduction, however, it issignificantly less than that obtained with kit 1. For example, at Ue / U = 3, theminimum CD corresponds to the twin vertical fence case and has a value ofaround 0.8. This represents a drag reduction of 22.5% compared to 30.6% givenby kit 1.0.70 542^3Uc11.1ff Idx /dCD0.9^ bfv/ci = 0, bfhid = 0bf v /d = 0.47, (xf id)cr = 0.29, bfh/d = 0—X— bfv/d = 0.47, (xf /d)cr • (bfh/d)/2 = 0.285, (Yf /d)cr = 0.20hf^= 0.08, Hk/d = 0, Rn = 1.2x10 50.8CylinderFigure 37. Variation in the drag coefficient with the cylinder speed ratio for kit 2 at its optimum verticalorientation, gap covered and fences in critical geometry.1094. CONCLUDING REMARKS4.1 Summary of Results4.1.1 Application of the MSBC to a 2-D wedge airfoil modelBased on a rather fundamental study of Moving Surface Boundary - layerControl (MSBC) with a two dimensional wedge airfoil model, conducted at asubcritical Reynolds number of 3x104, following general conclusions can bemade:(i) Rotation of the leading edge cylinder results in increased suctionover the nose. It is the propagation of the lower pressuredownstream that determines effectiveness of the rotation. Thisdepends mainly on the speed of rotation, surface roughness andsmoothness of transition from the cylinder to the airfoil surface.A large gap (> 3 mm) substantially decreases beneficial effect ofthe cylinder rotation.(ii) The increased momentum injection into the boundary-layer, withan increase in speed and appropriate surface roughness, delaysseparation of the flow from the upper surface resulting in higherlift and reduced drag. The existence of a critical speed is alsoevident beyond which the momentum injection through a movingsurface appears to have relatively less effect.110(iii) With the rotation of the cylinder the onset of flow separationoccurs at higher angles of attack. The upper surface flow remainsattached up to a distance downstream of the leading edge at whichpoint it separates followed by, at times, reattachment downstream.(iv) Rotation of the smooth cylinder resulted in the increase of CLmaxby 170%. The corresponding decrease in drag was about 36%.(v) Among the cylinder surfaces tested, the splined configurationproved to be the most successful in increasing lift as well asreducing drag. It raised the CLmax from 1.47 (reference case) to 4.3(spline-2 case), an increase of around 193%! The reduction in dragwas also quite impressive. In fact, the maximum CL / CD increasedfrom 1.53 to 78.93. Although the splined cylinder proved to be thebest, the results showed that an increase in roughness of thecylinder surface, in general, improves the boundary-layer control.(vi) The large CL / CD attained here through MSBC can be used toadvantage in designing next generation of high performanceairplanes.(vii) As the separation of shear layers is delayed, or even suppressed,the process of vorticity generation and its shedding in the wakewill be affected. Hence the moving surface boundary-layer controlmay prove effective in suppressing vortex induced and gallopingtype of instabilities. Investigation in this area is in progress and111appears quite promising.(viii) The concept of MSBC is essentially semi-passive in characterrequiring negligible amount of power for its implementation. Inthe present set of model tests a 1/8 H.P. (.-,--. 90 W) motor was morethan adequate to obtain L/c / U = 4.4.1.2 Application of the MSBC and fences to a 3-D truckThis study is aimed at assessing the effect of momentum injection andtripping of the boundary-layer on drag reduction of three dimensional modelsof the rectangular prism and the tractor-trailer truck. Based on the windtunnel data, following general conclusions can be made:(i) Both the concepts are quite promising in reducing the drag.(ii) Effectiveness of the momentum injection procedure diminisheswhen the rotating element is submerged in the wake.(iii) Helical surface roughness of the rotating element improvesefficiency of the MSBC only by a small amount. The maximumdrag reduction even with twin cylinders (Case 2, Figure 20) wasfound to be around 14.4%.(iv) A splined rotating cylinder injects momentum into the boundary-layer more directly presenting an exciting possibility of a furtherreduction in CD . With the twin splined cylinders (Case 6, Figure11221), the drag reduction of 52% was realized (Ucf/ U = Ucr l U --, 6)! Even with the speed ratio of 4, the CD reduction was as large as26%.(v) The moving surface boundary-layer control (MSBC) process isessentially semi-passive, i.e. it requires very little energy. For themodel tests, 1/8 H.P. motors (--, 95 W) were more than adequate toattain Uc l U = 6. For a prototype truck, a little over 1.5 H.P.would be required, which is negligible compared to 400-500 H.P.engine of a typical truck.(vi) The concept of fences to trip boundary-layer appears to be evenmore promising. For a three-dimensional prism, simulating atrailer, 31% reduction in the drag coefficient was observed. Areduction in CD by around 24.6% for a truck configuration isindeed exciting. Note, the process is entirely passive requiring noadditional energy.(vii) A hybrid combination of fences and the MSBC appears favourable.A cylinder-fence kit would make application of the concepts toexisting trucks more attractive. Among the two kits tested, kit-1proved more efficient promising the drag reduction of around 30%at Uc /U = 3.4.2 Recommendation for Future Work113A comment concerning the future plan of study would be appropriate.Effectiveness of the MSBC and fence concepts having been established, severalexciting possibilities present for further study and diverse application:(i) More precise wind tunnel tests using accurate models, simulationof relative road motion and the side wind induced yaw conditionsin supercritical Reynolds number range represent obviousextension of the project.(ii) The road tests using prototype truck with MSBC and fencesshould provide valuable field data.(iii) Development a numerical code for multielement airfoil and bluffbodies with momentum injection represents the area that receivedvirtually no attention. Of course, it is enormously challenging taskbut, if successful, should prove equally satisfying. Even acomputational tool for a two-dimensional flat plate with MSBCwould represent a major advance in the field.Similar studies with fences also need attention.(iv) As the momentum injection affects the separating boundary-layerand the associated wake, it could be used to advantage incontrolling vortex resonance and galloping type of wind inducedinstabilities often encountered in industrial aerodynamicsproblems.(v) Enormous opportunity exists for application of the MSBC to high114performance next generation of aircraft, both at the leading edgeof the wing and the control surfaces. The same is true forunderwake application with hydrofoils and rudders.(vi) The concept of MSBC can be applied at the end of threedimensional wings to counter tip vortices thus facilitating theirdispersion, minimizing the downwash, and providing improvedlift/drag characteristic. It was tried out by Prof. V.J. Modi in apreliminary fashion in 1986 and appeared promising. However,more precise study is necessary.(vii) The MSBC can be used to advantage in design of an efficientdiffuser with large diverging angle as often encountered inchemical industry.(viii) The fence concept can also be applied quite effective in industrialaerodynamics problems mentioned in (iv). They can be used toreduce drag of a wide variety of bluff bodies including buses,railway carriages, marine vehicles and others.115REFERENCES[1] Simanaitis, D., "Reduced Resistance Equals Increased Miles per Gallon,"Road and Truck, June 1980, pp. 88-90[2] McDonald, A. T., Palmer, G. M., et al., "Truck and Bus AerodynamicsInvestigated,"Automotive Engineering, Society of Automotive Engineers,Vol. 88, No. 11, November 1980, pp. 50-57.[3] Goldstein, S., Modern Developments in Fluid Mechanics, Vols. I and II,Oxford University Press, 1938.[4] Lachmann, G. V., Boundary layer and Flow Control, Vols. I and II,Pergamon Press, 1961.[5] Rosenhead, L., Laminar Boundary Layers, Oxford University Press,1966.[6] Schlichting, H., Boundary Layer Theory, McGrawHill Book Company,1968.[7] Chang, P. K., Separation of Flow, Pergamon Press, 1970.[8] Favre, A., "Contribution a l'Etude Experimentale des MouvementsHydrodynamiques a Deux Dimensions," Thesis presented to theUniversity of Paris, 1938.[9]^Alvarez-Calderon, A., and Arnold, F. R., "A Study of the AerodynamicCharacteristics of a High Lift Device Based on Rotating Cylinder Flap,"Stanford University Technical Report RCF-1, 1961.116[10] Cichy, D. R., Harris, J. W., and MacKay, J. K., "Flight Tests of aRotating Cylinder Flap on a North American Rockwell YOV-10AAircraft," NASA CR-2135, November 1972.[11] Weiberg, J. A., Giulianettij, D., Gambucci, B., and Innis, R. C., "Takeoffand Landing Performance and Noise Characteristics of a Deflected STOLAirplane with Interconnected Propellers and Rotating Cylinder Flaps,"NASA TM X-62, 320, December 1973.[12] Cook, W. L., Mickey, D. M., and Quigley, H. G., "Aerodynamics of JetFlap and Rotating Cylinder Flap STOL Concepts," AGARD FluidDynamics Panel on V/STOL Aerodynamics, Delft, Netherlands, April1974, Paper No. 10.[13] Johnson, W. S., Tennant, J. S., and Stamps, R. E., "Leading EdgeRotating Cylinder for Boundary-layer Control on Lifting Surfaces,"Journal of Hydrodynamics, Vol. 9, No. 2, April 1975, pp. 76-78.[14] Modi, V. J., Sun, J. L. C., Akutsu, T., Lake, P., McMillian, K., Swinton,P. G., and Mullins, D., "Moving Surface Boundary-layer Control forAircraft Operation at High Incidence," Journal of Aircraft, AIAA, Vol.18, No. 11, November 1981, pp. 963-968.[15] Mokhtarian, F., and Modi, V. J., "Fluid Dynamics of Airfoil with MovingSurface Boundary-layer Control," AIAA Atmospheric Flight MechanicsConference, August 1986, paper No. 86-2184-CP; also Journal of Aircraft,Vol. 25, No. 2, February 1988, pp. 163-169.117[16] Mokhtarian, F., Modi, V. J., and Yokomizo, T., "Rotating Air Scoop asAirfoil Boundary-layer Control," Journal of Aircraft, AIAA, Vol. 25, No.10, October 1988, pp. 973-975.[17] Mokhtarian, F., Modi, V. J., and Yokomizo, T., "Effect of MovingSurfaces on the Airfoil Boundary-layer Control," AIAA AtmosphericFlight Mechanics Conference, Minneopolis, Minnesota, August 1988,Paper No. AIAA-88-4303CP; also Proceedings of the Conference, Editors:R. Holdway and B. Kaufman, AIAA Publisher, pp. 660-668; also Journalof Aircraft, AIAA, Vol. 27, No. 1. January 1990, pp. 44-50.[18] Sovaran, G., Morel, T., and Mason, T. W. Jr., "Aerodynamic DragMechanisms of Bluff Bodies and Road Vehicles," Proceedings of theSymposium held at the General Motors Research Laboratories, Plenumpress, New York, 1978.[19] Koernig-Facsenfeld, F. R., Aerodynamik des kraftfahrzeugs: Verlay derMotor-Rundschau, Umshau Verlag, Frankfurt, West Germany, FirstEdition 1951, Reprinted 1980.[20] Kramer, C., and Gerhardt, H. J., "Road Vehicle Aerodynamics,"Proceedings of the 4th Colloquium on Industrial Aerodynamics, Aachen,June 1980.[21] Kurtz, D. W., "Aerodynamic Design of Electric and Hybrid Vehicles: AGuidebook," U. S. Department of Energy, Report No. 5030-471,September 1980.118[22] Bearman, P. W., "Review of Bluff Body Flows Applicable to VehicleAerodynamics," Transactions of ASME, Journal of Fluids Engineering,Vol. 102, Sept. 1980, pp. 265-274.[23] Wacker, T., "A Preliminary Study of Configuration Effects on the Dragof a Tractor-trailer Combination," M.A.Sc. Thesis, University of BritishColumbia, Vancouver, Oct. 1985.[24] Kataoka, T., China, H., Nakagawa, K., Yanagimoto, K., and Yoshida, M.,"Numerical Simulation of Road Vehicle Aerodynamics and Effect ofAerodynamic Devices," SAE International Congress and Exposition,Detroit, Michigan, U.S.A., 1991, Paper No. 91-0597.[25] Larsson, L., Broberg, L., and Janson, C., "A Zonal Method for PredictingExternal Automobile Aerodynamics," SAE International Congress andExposition, Detroit, Michigan, U.S.A., 1991, Paper No. 91-0595.[26] Modi, V. J., Shih, E., Ying, B., and Yokomizo, T., "On the DragReduction of Bluff Bodies through Momentum Injection," AIAA 8thApplied Aerodynamics Conference, Portland, Oregon, Paper No. 90-3076;also Journal of Aircraft, AIAA, in press.[27] Mason, W. T., Jr. and Beebe, P. S., "The Drag Related Flow FieldCharacteristics of Trucks and Buses," Proceedings of the Symposium onAerodynamic Drag Mechanisms of Bluff Bodies and Road Vehicles, Ed.G. Sovran, T. Morel and W. T. Mason, Jr., Plenum Press, New York,1978, pp. 77-78.
- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- UBC Theses and Dissertations /
- Boundary-layer control of bluff bodies with application...
Open Collections
UBC Theses and Dissertations
Featured Collection
UBC Theses and Dissertations
Boundary-layer control of bluff bodies with application to drag reduction of tractor-trailer truck configurations Ying, Bin 1991
pdf
Page Metadata
Item Metadata
Title | Boundary-layer control of bluff bodies with application to drag reduction of tractor-trailer truck configurations |
Creator |
Ying, Bin |
Date Issued | 1991 |
Description | Effectiveness of two fundamentally different concepts of boundary-layer control for the drag reduction of bluff bodies is studied experimentally. The methods are: (i) Moving Surface Boundary-layer Control (MSBC) involving momentum injection through one or more rotating elements (light hollow cylinders); and (ii) tripping of the boundary-layer using judiciously located fences to interrupt pressure recovery. Wind tunnel tests with a two-dimensional wedge airfoil model suggest that injection of momentum can significantly delay separation of the boundary-layer resulting in a narrow wake and the associated reduction in the pressure drag. It also leads to a substantial increase in the lift at a given angle of attack resulting in a dramatic rise in the lift to drag ratio, from 2 to 80, under optimum conditions. Effectiveness of the momentum injection process is primarily governed by the gap-size, cylinder surface roughness and the ratio of the cylinder surface velocity (Uc) to the free stream velocity (U). A three-dimensional model of a rectangular prism and a 1/12 scale model of a typical tractor-trailer truck configuration show that both the MSBC and fence approaches are promising in reducing the aerodynamic resistance. A kit configuration is proposed for ease of implementation of the concepts on new and existing trailers. Road tests with a full scale cube-truck are recommended to assess effectiveness of the boundary-layer control procedures in reducing the drag during highway conditions. |
Extent | 9773015 bytes |
Genre |
Thesis/Dissertation |
Type |
Text |
FileFormat | application/pdf |
Language | eng |
Date Available | 2008-09-16 |
Provider | Vancouver : University of British Columbia Library |
Rights | For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use. |
DOI | 10.14288/1.0081045 |
URI | http://hdl.handle.net/2429/2106 |
Degree |
Master of Applied Science - MASc |
Program |
Mechanical Engineering |
Affiliation |
Applied Science, Faculty of Mechanical Engineering, Department of |
Degree Grantor | University of British Columbia |
GraduationDate | 1992-05 |
Campus |
UBCV |
Scholarly Level | Graduate |
AggregatedSourceRepository | DSpace |
Download
- Media
- 831-ubc_1992_spring_ying_bin.pdf [ 9.32MB ]
- Metadata
- JSON: 831-1.0081045.json
- JSON-LD: 831-1.0081045-ld.json
- RDF/XML (Pretty): 831-1.0081045-rdf.xml
- RDF/JSON: 831-1.0081045-rdf.json
- Turtle: 831-1.0081045-turtle.txt
- N-Triples: 831-1.0081045-rdf-ntriples.txt
- Original Record: 831-1.0081045-source.json
- Full Text
- 831-1.0081045-fulltext.txt
- Citation
- 831-1.0081045.ris
Full Text
Cite
Citation Scheme:
Usage Statistics
Share
Embed
Customize your widget with the following options, then copy and paste the code below into the HTML
of your page to embed this item in your website.
<div id="ubcOpenCollectionsWidgetDisplay">
<script id="ubcOpenCollectionsWidget"
src="{[{embed.src}]}"
data-item="{[{embed.item}]}"
data-collection="{[{embed.collection}]}"
data-metadata="{[{embed.showMetadata}]}"
data-width="{[{embed.width}]}"
async >
</script>
</div>
Our image viewer uses the IIIF 2.0 standard.
To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0081045/manifest