steady and unsteady flow visualization inside a …phoenics/em974/projetos/temas projetos/vortex...

15th International Conference on Experimental Mechanics

ICEM15 1

PAPER REF: 3022 STEADY AND UNSTEADY FLOW VISUALIZATION INSIDE A CHANNEL WITH WALL SQUARE PROTUBERANCES AT MODERATE REYNOLDS NUMBERS Rodrigo A. Bassan1, Odenir de Almeida1

, Sérgio Said Mansur2, and Edson Del Rio Vieira2(*)

1, Faculty of Mechanical Engineering (FEMEC), Federal University of Uberlândia, Uberlândia, MG, Brasil 2Department of Mechanical Engineering (DEM), Mechanical Engineering Faculty, São Paulo State University (UNESP), Ilha Solteira, SP, Brazil (*)Email : [email protected] ABSTRACT Digital images concerning with the flow inside a channel with wall square protuberances are captured in the present work. Liquid dye injected by means of long hypodermic injection needles has been utilized to make the flow visible. Digital photos and videos have been captured, showing the features of the steady and unsteady flow in such a geometry. All tests have been performed in a low turbulence vertical hydrodynamic tunnel operating in blow-down mode at Reynolds number up to 1 000. The characteristic dimension (D) used for calculating Reynolds has been taken as the height of the square protuberances, i.e. equal to 10 mm. The channel has 146 × 30 mm of cross section and square protuberances are orthogonally positioned relatively to the flow, resulting in a blockage ratio of 1/3. Flow instabilities like detachments, recirculation, and vortex shedding have been identified and explained in detail.

INTRODUCTION Surface roughness offers a large influence in a wall-bounded flow. The fluid moving over and around the roughness protuberances modifies significantly the flow near the wall. Generally, roughness increases the drag in a turbulent boundary layer due the pressure forces actuating in the roughness elements. In some cases, special geometries of roughness cause a drag reduction. For example, the effect of alternating favorable and adverse pressure gradients generated by convex and concave curvature in a wavy wall produces in a turbulent incompressible flow an appreciable drag reduction (Abdel-Shafi and Nishikawa, 1993).

In fluid flow systems, most practical interest surfaces have texture, roughness or imperfections generated in production process or during the operation due to fouling, pitting corrosion, or surface deposits by incrustation, dust or oil impregnation. Understanding the mechanism responsible for the changes in the fluid motion for different roughness types is important to predict turbulence and drag in aerodynamic design.

Nowadays, flow over a wide variety of fully aerodynamically rough surfaces has been frequently employed in engineering in order to increase the performance of thermal and hydraulic devices. From a hydrodynamic view point, two groups of superficial rough can be identified. The first one is named K type roughness and shows an aspect like a sandpaper with granular material dispersed on it’s the surface. In this case, very little unstable verticals structures with length scale near of the height K of the roughness have been continuously produced and due to flow advection generate a tridimensional turbulence of small scale very near to wall. The second type is named D type roughness and can be made by several

Porto/Portugal, 22-27 July 2012

Editors: J.F. Silva Gomes and Mário A.P. Vaz 2

geometries of solid obstacles carefully positioned in the smooth wall. Generally, square, triangular, rectangular or semi circular protuberances are employed to produce a D type surface. D type surface are frequently observed in electrical devices and electronic circuits where the roughness or wall protuberances are formed by several types of electronic components. In general, the flow with a D type roughness wall shows stable vortices confined in the reentrance between the wall protuberances.

Transverse ribs of several different cylindrical geometries arranged in a periodic fashion are utilized in order to enhance heat transfer in several engineering applications. Generally, the studies of heat transfer enhancement of this D type surface were directed toward investigating the average heat transfer. Local variation of heat transfer determination is necessary since non-uniform cooling (or heating) may induce excessive thermal stress and reduction component life time. Local heat transfer coefficient near the protuberance is determined primarily by the flow field topology in this region. Generally, flow past a protuberance produces boundary layer separation, reattachment of the separated shear layer and boundary layer redevelopment generating local maximum heat transfer coefficients. On this sense, a detailed knowledge of the flow field around the protuberances is absolutely necessary. Numerical and experimental studies of the incompressible flow in channel with wall protuberances permit to predict accurately the streamlines.

The flow around cylindrical obstacles inside a channel positioned transversally in the wall has been intensively studied by several authors in the last decades because they are a recognized practical importance in engineering design (Bassan et al. 2011). The turbulence level is strongly affected by the protuberances modifying heat transfer and drag coefficients. The detailed knowledge of this flow field is necessary to improve design of several engineering equipments.

In the present work, a study of the flow inside a channel with square cylindrical protuberances transversally positioned on the internal wall of the channel is carried out by utilizing a vertical hydrodynamic tunnel. Flow visualization techniques have been employed through a liquid dye injection. Images of the flow for Reynolds number up to 1000 have been captured digitally to provide still photo and video images. The flow images show different verticals structures occurring in steady and transient flow.

FLOW VISUALIZATION Many human activities deal with analysis of a great amount of data. Satellites orbit planning astronomical data analysis, spacecraft design, geophysical and meteorological studies, and complex engineering projects are good example. The human brain cannot interpret adequately the information inherent to these activities, if the data is presented exclusively in a numerical format. On the other hand, one unique image can condense gigabytes of information, and make easier the interpretation of biological, medical, meteorological or physical information. The methodology to transform data into visual information - in other words, the technique which scientists utilize to generate and interpret images - is named visualization, and offers a form for seeing the unseen. All experimentalists know that once a phenomenon is visualized, a large steep has been taken toward its understanding and solving the theoretical and experimental problems involved with. Flow visualization can be defined as the art and science of obtaining a clear image of a physical flow field and the ability to capture it on chemical pellicle or electronic devices for further processing (Freymuth, 1993). In applied fluid mechanics, the optimized design of most equipment depends on the knowledge acquired on the flow features within different components. In such flows,


ICEM15 3

complicated mechanisms are almost always present whose complete nature cannot be revealed exclusively by means of punctual measurements, as it is usually performed with the help of hot-wire or laser Doppler anemometry. In many cases, this local measurements no permit see an entire flow behavior and a detailed topology of the flow near the measurement points. On the other hand, flow visualization represents an efficient tool for qualitative analysis of fluid motion, providing topological issues about flow fields. In other words, while local measurements in general supply only quantitative information – or hard data – flow visualization walks in the opposite way contributing intensely to physical interpretation of results. Flow visualization can be performed in gases and liquids, but the use of water as a fluid of work offers a possibility of employing several different techniques in order to obtain detailed flow images. In an aqueous medium, pathlines, streamlines, streaklines, or timelines can easily became visible by using solid particles, liquid dye filaments, or little gaseous bubbles suspended in the water. A suitable lighting generates good images with striking results. Many reviews in flow visualization in water have been presented in the technical literature (Clayton & Massey, 1967; Werlé, 1973, Freymuth, 1993, Mueller, 1983; Vieira & Mansur, 2004).

Flow visualization has allowed putting in evidence the complex topology of vortical structures generated in the mounted-wall protuberance vicinity. By means of flow visualization, patterns of recirculation, boundary layer detachment, vortex shedding, and other complex mechanisms inherent to such a flow are easily identified. The flow visualization technique applied in the present work is the direct injection of opaque liquid dye in non-perturbed flow by means a rake of long hypodermic needles of 0.7 mm external diameter. The dye utilized is a solution of black PVA pigments, water and ethyl alcohol. Ethyl alcohol is only applied to correct the solution density and make it very close to water’s density. If occurs density difference between the dye solution and water, in very low Reynolds number, undesirable effects of convection is visible. Strong amount of this colored dye has been injected directly in the non perturbed stream, sufficient to color the entire flow field. Subtly, the injection dye is stopped, and the clean water flow wash the entire flow field, except in the cylinder wake, because in this region the flow speed is significantly small than other regions. This procedure permits to see, for seconds, the re-circulating bubble and the wake downstream the protuberance. Details about flow visualization technique employed are available in Bassan et al. (2011) work presented in CIBEM 2011 – 10th Iberoamerican Congress of Mechanical Engineering, Porto, Portugal.

EXPERIMENTAL APPARATUS

In order to build a flow visualization facility it is necessary a flow generator device, i.e. an apparatus to make the flow visible and equipment to record the images. An extensive use of water tunnels and towing tanks is observed around the world providing good conditions for experimental external flow studies. But, many internal flow studies require the construction of experimental devices specially designed to solve a specific problem. Generally, the experiments in internal channels utilize special flow generator devices (Onur & Baydar, 1992; Wang et al., 2010). In the present research a vertical hydrodynamic tunnel with a closed test section is used to generate a low turbulence flow with a flat velocity profile and a thin boundary layer. The channel flow with wall square protuberances is adequately positioned in the center line of the test section. Fig. 1 shows the channel with internal square protuberances mounted inside the hydrodynamic tunnel test section. That mounting enables quick and easy



replacement of canal by other settings, facilitating the testing of other geometric configurations.

Flow velocity measurements inside the channel have been performed with a 55R11 fiber-film probe made by Dantec Measurement Technology, with 70 μm diameter quartz fiber coated with 2 μm nickel film and with an overall length of 3 mm. This is a straight general-purpose type sensor which permits a wide measurement range in water medium. For very small velocities (up to 0.10 m/s) several special cares could be adopted in order to reduce the convection effect around the probe, as discussed by Rivir et al., (1996). A Dantec StreamLine 90C10 frame with 3 CTA modules 90C10 permits simultaneously measurements in 3 channels. An A/D board NI-DAQmx 8.7.1 (16 bits), made by National Instrument, has been utilized in order to record the output voltage signal.

Fig.1 Channel with wall protuberances mounted inner the hydrodynamic tunnel test section.

In the present work, all tests have been carried out in a 146 × 146 × 500 mm test section hydrodynamic tunnel with octagonal cross section and 6 m of total height. Fig. 2 depicts a sketch of the vertical hydrodynamic tunnel showing the principal parts. An external subterraneous water tank (LR) with a capacity of 9.8 m3, careful protected against dust and possible contaminants, provides the water for the experiments. The construction of an external sealed tank with large dimensions shows a good solution in order to remain clear the water along the experiments avoiding frequent water changes. A membrane filter (no showed in Fig. 2) removes up to 10 µm solid particles of the water. The water pump (PP) is a KSB pump model Megachem 32-200 type of 5.5 KW of power all constructed in stainless steel to fit out in a level below of the reservoir level. The pump installation bellow the reservoir level eliminates the check valve. The pump is installed in a subterraneous power-house, external the laboratory room, careful to fit up on vibration isolated supports in order to minimize the vibration transmission to the tunnel. A 75 mm nominal diameter PVC tube with 5 mm of wall thickness positioned in the exhaust of the pump discharge the flow to the upper of the tunnel. All valves are made in stainless steel except to valve #3. All valves showed in Fig. 1 are of manual operation. An automatic sphere valve pneumatically powered and electrically driver, non-showed in Fig. 2, should be mounted in order to remote control the flow. The use of a sphere valve also permits to control the flow rate adequately. Unfortunately, sphere and butterfly valves utilized in chemical industry applications are expensive because the rigid security norms. The valve #2, is a butterfly valve installed in order to manually control the

30 mm

120 mm

30 mm

146 mm 10 mm

Test section wall

Channel with wall protuberances


ICEM15 5

inlet flow inside the tunnel. This valve type mounted after the pump in a high pressure line and operated only by one quarter of lap need to be slowly moved in order to avoid a sudden flow interruption causing undesirable hydraulic ham effect.

The tunnel stagnation section (the upper part of the tunnel) is composed by an upper reservoir (UR), an upper contraction (UC), screens (S), honeycombs (SH) and a discharge diffuser. Discharge diffuser, contraction and screens are needed in order to introduce the flow field with a minimal turbulence in the stagnation section. All walls and parts need to be not flexible and enough rigid in order to no provoke undesirable flow oscillations in the test section. The work of Vogt, (1983) discusses the problem due to residual turbulence internal the stagnation section in vertical hydrodynamic tunnel. The honeycombs with hexagonal cells of 6 mm and 280 mm of thickness are made of fine sheet of alloy 3003 aluminum. The discharge diffuser shows 614 small holes of 3 mm of diameter proportionally distributed order to delivery uniformly the flow inside the stagnation section producing a minimal perturbation, principally in continuous mode operation.

Fig.2 Vertical low turbulence hydrodynamic tunnel.

The maximum water level internally in the upper reservoir is controlled by an exhaust PVC pipe of 100 mm nominal diameter. The contraction (LC) has a short length and a contraction ratio of 1:16. The test section (TS) was made of aeronautical aluminum 4050 with windows of optical Plexiglas with 10 mm of thickness. The average velocity at the test section has been determined from the water flow rate measured by an electromagnetic flowmeter (FM). This practice of determining the average velocity in the test section measuring the downstream bulk flow rate is recommended by several authors, and currently used in many



water tunnel facilities. The non-perturbed velocity, upstream the test model has been obtained, in this work, using an AXF100G model Yokogawa electromagnetic flow meter mounted downstream the test section. An assessment of the uncertain associated to free stream velocity shown less than 4%, when compared with data obtained by hot film anemometer (Dantec CTA Streamline).

The tunnel structural support was entirely constructed of NPS 6 Schedule 40 steel pipes with 150 mm of nominal diameter and 7 mm of wall thickness for a rigid structure with minimal vibrations. Seamless schedule pipes provide a high quality tube for structural application. The tunnel structure need be to place carefully in an adequate foundation bases in order to isolate external vibration conducted by the ground.

All parts of the tunnel have been made of low water absorption composite material (polyester isophthalic resins and glass fiber) with 8 mm of thickness strengthened with a steel gage in order to provide very strong walls. In a hydrodynamic tunnel, all walls need to be sufficiently rigid because small wall oscillations can provoke boundary layer detachment and high level of turbulence. Direct water contact with metallic walls provokes the ions formation generating several chemical reactions in many parts of the tunnel and irreversible damages on expensive hot film probes. In order to minimize the chemical reactions, all the tubes and connections have been made of thermoplastic PVC (Polyvinyl chloride). In order to avoid undesirable vibrations and deformations of the tubes and to support severe operation modes, all tubes present a minimal of 3 mm of wall thickness. All valves and the pump are of stainless steel due to use of carbon steel in water contact generates severe oxidation and water contamination.

The employment of hot film anemometry need to an adequate electrical ground reference. Due to use of electric non conductor plastic composite material in many tunnel parts an adequate electrical ground must be careful made. A system to remove gaseous dissolved in the work water by suction is positioned in the discharge tube (DP) after the valve of flow control (valve V#4) avoiding small bubble formation in the exhaust pipe. Air bubbles in exhaust pipe need to be drained because produce undesirable flow rate oscillations. Boundary layer thickness effect on velocity profile is controlled by divergent walls in the test section showing an adequate solution. Low velocity aerodynamic tunnel also utilizes currently divergent walls in test section in order to remain constant the velocity profile.

The flow image has been illuminated by cold fluorescent lamps placed in line with the camera and shielded by white velvet-like translucent paper to provide a uniformly diffuse bright background against which the dye patterns were photographed. Fluorescent lamps with high color temperature but minimal heat emission have been utilized permitting sharp and well defined images. The use of Rosco color illuminating filter Cinegel#3308 converts daylight fluorescent lamps to 5 500 K and a diffuser Cinegel#3007, a slight filter with less density softens edge and provides a good illumination for pictures and video capture.

The water tunnel is operated by gravitational action, and can be used either in continuous or blow-down mode. Blow-down mode have been used in this work, due to its lower turbulence level, although in this mode, the free stream mean velocity decreases noticeably with the water level inside the upper reservoir. To account for that, it has been estimated that, for a period up to 15 seconds, the effects of decreasing free stream mean velocity are overshadowed by turbulence.

Despite the fact that the flow around bluff bodies has been studied for over 100 years, this problem is still under intensive experimental and numerical investigation. In this work, a


ICEM15 7

preliminary study of the flow internally a channel with square cylindrical protuberances has been carried out utilizing flow visualization tools.

In the past, flow visualization has demonstrated a strong influence in fluid mechanics development, on an historical and scientific basis. Today, in an engineering context, the use of images is widely spread in a large variety of applications. More specifically, flow visualization techniques associated to image processing contributing in a decisive way to the solution of practical engineering problems.

In the last decades, sophisticated flow visualization techniques have been created or improved. Nevertheless, many techniques employed today are the same as a century ago, for instance, the direct injection of liquid dye into the free flow by means of needles. This technique is cheap, easy to implement and commonly used in hydrodynamic environment. Every individual image capture process requires a proper illumination and many different light sources and techniques can be found in the literature, but their successful utilization is highly based on the researcher personal experience. Image capturing of flow visualization generally is not a trivial task, but it can be appropriately accomplished with the help of few basic equipments. A DSLR photographic full frame 35 mm camera equipped with a luminous macro lens (preferably a 100 mm model) is sufficient to obtain high quality pictures and full HD video.

Two flow visualization techniques have been employed in this work. The first, called dye wash technique consists on adequate injection of opaque liquid dye into the non-disturbed flow field upstream to the solid body by means of a needle. Then, dye injection is suddenly stopped, the needle is removed, and the clean water stream washes the flow field, except the regions where the flow speed is relatively small, as in the cylinder boundary layer and wake. This procedure enables to visualize, for some seconds, the vortex street downstream the cylinder. A second flow visualization technique employed is a traditional liquid dye injection by means a long hypodermic needle - 0.7 mm O.D. - adequately positioned upstream the test model. In this situation, an opaque liquid dye filet is continually injected in the free flow permitting the streakline visualization.

RESULTS The channel depicted in Fig. 1, have two protuberances (side L = 10 mm) on the bottom wall with a relative distance between them equal to three times the side of the protuberance. In this situation, aspect ratio is 14.6 and channel blockage ratio is 33 %. Results include several still photos and videos of the vortex wake produced by the square ribs. The non-dimensional parameter Reynolds number (Re) is defined by Eq. (1)

μρ LVRe = (1)

where (ρ) is the fluid density, (L) is the height of the protuberance, (V) is non-perturbed flow velocity and (μ) is the viscosity.

Figure 3 depicts a steady flow inside the channel in at different Reynolds numbers. Fig. 3(a) shows (Re ≅ 4) the flow at a relatively low Reynolds number where viscous forces are dominating (near creeping flow). It is possible to see a very little detachment with the dye streakline, skirting almost perfectly, solid obstacles. In Figure 3(b), by increasing the Reynolds number (Re ≅ 11), the detachment can be seen clearly in the rear of the protuberance edges. Also, it is possible to visualize recirculation in the space between the



two protuberances and immediately after the second protuberance. These recirculation bubbles show a geometry which remains unchanged over time. The flow is attached in the top of the two protuberances. Boundary layer reattachment occurs in the bottom channel wall very near the second protuberance. In the image of Fig 3(c), for the Reynolds number equal to 19, it is possible to see the bubble recirculation remaining unchanged over the time. The boundary layer reattachment occurs downstream the second protuberance too far away. For higher Reynolds numbers the boundary layer reattachment moves away from the second protuberance.

In Fig. 3(d) – Re = 44, is possible to visualize small amounts of mass injected into the space between the two protuberances in a pulsatile mode. The mass injection occurs in a very proximity upstream the second protuberance. Immediately, after stopping the mass injection, an equivalent mass is ejected out. Mass ejection and injection is occurring alternately but only very small instabilities have been observed in the streakline near the free flow. At Reynolds numbers equal to 88 – Fig. 3(e) – due to mass injection an increase of momentum is observed internally in the recirculating bubble causing a stable bubble. The flow in the top side of the protuberances remains attached. In the Fig. 3(f and g) the flow topology remains practically unaltered and the recirculation bubble shows a very stable configuration. However, in the fig. 3(h) for Reynolds 150, the flow shows a weak detachment followed by an immediate reattachment on top of the first protuberance. This detachment promotes small instabilities in the streamline in the boundary of the recirculation bubble increasing the frequency of the mass injection and ejection process. But, in the top of the second protuberance the flow remain attached. By increasing the Reynolds number - Figure 3 (g) and (h) we see that the streamline existent in the top of the recirculation bubble between the two protuberances loses their alignment with the top faces of the protuberances starting the process of galloping structures. In Fig.3 (i), for Re = 348, the detachment in the front edge of the first protuberance increases significantly and no more reattachment is observed on the upper surface of the first protrusion. On the upper surface of the second protuberance is possible to observe a small turbulent galloping structure. The wake produced downstream the second protuberance shows a vortex structure with a defined vortex frequency. In Fig. 3(j), for the Re = 419, there is a sharp change in the structure of the recirculation bubble between the protuberances and we can see a intense mass exchange taking place in the region that occurs in the recirculating bubble after the second protuberance which can be better observed in the image of Figure 3 (k) with the aid of a fillet dye properly positioned. Fig. 3(k) - Re = 543 - also allows the visualization of a counter rotating recirculating region near the posterior surface of the first protuberance. In Fig. 3(l), (m) and (n) it can be seen that the instability shock against the front edge of the second protuberance producing intense energy and mass exchange with recirculating existing in the region between the two protuberances. Finally, in Fig. 3(o) we see that the detachment occurring in the front edge of the first protuberance creates shear layers colliding against the front surface of the second protuberance introducing strong exchange of mass and energy into the recirculating bubble existing between the protuberances. Inside the region between the protuberances we can also observe the existence of two distinct regions, the first near the second protuberance where the effects of exchange of mass and energy are intense, generating high relative velocities and the second region, close to the first protuberance, where exchanges are smaller and the relative speeds are low. Upstream of the first protuberance, another separation bubble occurs as a result of the strong adverse pressure gradient. This small clockwise recirculation can be clear identified in Fig. 3 (l) (m) and (n). Finally, Fig 4 depicts a temporal sequence of images showing an accelerating transient flow in a channel with two protuberances from Re = 26 at the time equal to zero up to Re = 136 after 33 seconds.


ICEM15 9

(a) Re ≅ 4 (b) Re ≅ 11 (c) Re ≅ 19 (d) Re ≅ 44 (e) Re ≅ 88

(f) Re ≅ 101 (g) Re ≅ 135 (h) Re ≅ 150 (i) Re ≅ 348 (j) Re ≅ 419



(k) Re ≅ 543 (l) Re ≅ 558 (m) Re ≅ 660 (n) Re ≅ 722 (o) Re ≅ 850

Fig.3 Steady flow visualized images internal the channel with two square ribs.

(a)t=0s (b)t=3s (c)t=6s (d)t=9s (e)t=12s (f)t=15s


ICEM15 1

(g)t=18s (h)t=21s (i)t=24s (j)t=27s (k)t=30s (l)t=33s

Fig.4 Unsteady flow visualized images internal the channel with two square ribs. Temporal sequence of images showing a transient flow in a channel with two ribs from Re = 26 at

t = 0s to Re = 136 at t = 33s.

CONCLUSION Rough surfaces are intensively utilized in engineering in order to enhance heat transfer, albeit at the expense of a possible increasing in the drag force. Several roughness geometries, like riblets, may be used so as to decrease the drag or to delay transition. Obviously, roughness can modify the performance of airfoils, wings, turbomachinery and predicting its effect is very important in such applications. The intense activity of experimental research verified in this field has been based frequently on results obtained by different flow visualization techniques, which have contributed substantially to the understanding of the phenomena related to vortex dynamics. An additional advantage of flow visualization is that one can obtain qualitative and quantitative data from the flow images, without introducing physical disturbance to the flow field.

In the present work flow visualized images captured in a channel with square cylindrical wall protuberances have been depicted and the flow structures have been detailed explained in Reynolds number up to 1 000. The tests were conducted in a stable and unstable flow. Analyzing the stable flow, we observed the development of structures as the Reynolds number was increased. A recirculating structure is observed inside the rectangular cavity formed between the two protuberances. Additionally, Kelvin-Helmoltz instabilities start to develop around Re = 500. Downstream of the second cylindrical body these Kelvin-Helmoltz instabilities are generated over Re = 350. For the unstable flow, by accelerating from Re = 26 up to 136, there is the development of a vortex inside the cavity as well as downstream of the second body. For a given time inside the cavity, there is the appearing of two recirculating structures, where the top structure rotates counterclockwise and the lower one rotates clockwise as depicted in Figure 4 (d). This image was captured 9 seconds after the valve opening. As time evolves it is possible to see that the lower structure is engulfing the first upper structure until reaching almost the size of the cavity between the ribs. A stretched



structure is formed after the second protuberance mainly due to the proximity of the channel wall. Apparently a single and stretched recirculating bubble becomes stable.

The flow visualization technique used in this work provided clear insights about the flow dynamics of a channel with wall square protuberances at moderate Reynolds numbers.

ACKNOWLEDGMENTS The authors are grateful to the financial support provided by FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo) and also thankful for CAPES/PROCAD, FUNDUNESP and FEPISA for supporting the construction and acquisition of the experimental apparatus used in this study. Thanks also to Prof. Emanuel Rocha Woiski for proofreading the manuscripts and opportune collaborations in several opportunities. The experiments were conducted using the facilities of the Flow Visualization Laboratory of the Mechanical Engineering Department at UNESP, Ilha Solteira.

REFERENCES Abdel-Shafi, NY, Nishikawa, N. Numerical Analysis of Flows over Walls with Protuberances. Journal of Wind Engineering and Industrial Aerodynamics, 1993, 47, p. 275-281.

Bassan, RA, Mansur, SS, Vieira, ERDR, Vieira, EDR. Flow Structure Around a Square Section Cylindrical Protuberance Mounted on a Smooth Flat Plate, proceedings of the DINCON 2011 – X Conferência Brasileira de Dinâmica, Controle e Aplicações, Águas de Lindóia, Brasil.

Bassan, RA, Mansur, SS, Vieira, EDR. Rebuilding and Testing a Vertical Hydrodynamic Tunnel, proceedings of COBEM 2011 – 21st Brazilian Congress of Mechanical Engineering, Natal, Brasil.

Bassan, RA, Mansur, SS, Vieira, EDR. Experimental Flow Visualization Internal Channels with Wall Protuberance (in Portuguese), proceedings of CIBEM 2011 – 10º Congresso Iberoamericano de Engenharia Mecânica, Porto, Portugual.

Clayton, BR, Massey, BS, Flow Visualization in Water: A review of Techniques, Journal of Scientific Instrumentation, 1967, 44, p.2-11.

Freymuth, P, Flow Visualization in Fluid Mechanics, Review Scientific Intruments, 1993, 64, p.1-18. Merzkirch, W, Flow Visualization, ed.2, Academic Press, Orlando, 1987.

Onur, HS and Baydar, E, Laminar Channel flow Over a Square Step, International Journal Engineering Science, 1992, 9, p.1109-1116.

Mueller, T., Flow Visualization by Direct Injection, in: Fluid Mechanics Measurements, 1983, Goldstein, R.J. (ed.), cap. 7, p.307-375.

Rivir, RB, Chyu, M. and Maciejewski, PM, Turbulence and Scale Measurements in a Square Channel with Transverse Square Ribs, International Journal of Rotating Machinery, 1996, 2, p.209-218.


ICEM15 1

Vieira, EDR, Mansur, SS, Visualização Experimental de Escoamentos, Coleção Cadernos de Turbulência, vol.4, Möller, SV, Silvestrini, JH (eds.), p.33-71, Associação Brasileira de Engenharia e Ciências Mecânicas - ABCM, ISBN 85-85769-19-X.

Vogt, GL, Water Tunnel Construction for Continuous Mode Flow Visualization, AIAA-83-0657, 1983, 10 p.

Wang, L; Salewski, M and Sundén, B, Turbulent Flow in a Ribbed Channel: Flow Structures in the Vicinity of a Rib”, Experimental Thermal and Fluid Science, 2010, 34, p.165-176.

Werlé, H, Hydrodynamic Flow Visualization”, Annual Review of Fluid Mechanics, 1973, 5, p.361-382.


ICEM15 1

PAPER REF: 3036 FLOW AROUND A SQUARE CYLINDER POSITIONED NEAR A SMOOTH FLAT PLATE Rodrigo A. Bassan1, Odenir de Almeida1

, Sérgio Said Mansur2, Edson Rodrigo Del Rio Vieira2 and Edson Del Rio Vieira2(*) 1 Faculty of Mechanical Engineering (FEMEC), Federal University of Uberlândia, Uberlândia, MG, Brasil 2 Department of Mechanical Engineering (DEM), Engineering Faculty, São Paulo State University -

UNESP, Ilha Solteira, SP, Brazil (*)Email: [email protected] ABSTRACT Vortex shedding phenomenon produced by a square cylinder placed close to a smooth flat plate is experimentally studied by means of flow visualization techniques and hot-film anemometry. Qualitative and quantitative information about the flow field has been obtained for Reynolds numbers up to 1,000. Vortex shedding images in several Reynolds number have been captured and the non dimensional vortex shedding frequency has been obtained as a function of the Reynolds number.

INTRODUCTION Flow around cylinders positioned in a free stream has been extensively studied for over a century due to wide practical applications. The flow around a single square cylinder transversally mounted on a smooth flat plate has also been intensely investigated because it has a recognized importance in engineering applications. Tubes placed near heat transfer walls, vortex generator to improve the cooling in electronic devices, wires near walls, chimneys near tall buildings, pipelines near the ground or a sea or river bed are some examples of that. In such a situation, the flow field over the plate is strongly affected by the presence of the solid obstacle.

Another situation of engineering interest occurs when a cylinder is placed near a solid wall. According to Wang and Tan (2008), this flow type depends mainly on the following three parameters: the Reynolds number (Re), the boundary layer thickness (δ) and the height gap (G) between the cylinder bottom and the wall surface. But, the influence of other parameters on the flow should be observed. If the cylinder is partially immersed into the boundary layer of the flat plate, the boundary layer thickness (δ) is an important parameter, principally in high Reynolds numbers. Length scale of the wall turbulence (lw) is a function of the wall shear velocity (uτ) and offers hard influence on the flow field (Martinuzzi and Tropea, 1993). Nevertheless, the wall shear velocity (uτ) changes slightly over a wide range of Reynolds numbers in a fully developed channel flow or in a turbulent boundary layer (Kin and Lee, 2001). Obviously, the upstream flow conditions – upstream velocity profile and free stream turbulence – affect significantly the flow topology around the cylinder.

Different experimental tools have been employed to study this kind of problem. Qualitative information obtained by flow visualization has allowed putting in evidence the complex topology of vortical structures generated in the protuberance vicinity. Measurements using hot-wire anemometry (HWA) or laser Doppler velocimetry (LDV) permit to obtain quantitative data about velocity gradients and turbulent properties of the flow field. Combination of these



two techniques is a very good way to investigate complex flows, as demonstrated by (Freymuth et al., 1983).

Most studies of the flow past cylinders placed near a wall have so far been carried out experimentally for circular cylinder. The experimental study of the flow past a square cylinder placed near a smooth plane wall at Re = 22,000 – based on the side length of the square cylinder (D) –, was made by Bosch et al. (1996) using a flow visualization techniques and LDV. The cylinder was placed in a relative distance (G/D) from the wall equals to 0.75 and turbulence level in the oncoming water flow was 3 %. The flow topology was found to be very similar to that in the flow past two side-by-side cylinders.

Shi et al. (2010) studied the influence of the relative gap in the process of vortex suppression in Reynolds equal to 13,200 using a subsonic open circuit wind tunnel and HWA. They observed a complete vortex shedding elimination using G/D = 0.25.

Experiments performed out by Martinuzzi et al. (2003) at a suction open circuit boundary layer wind tunnel in Re = 18,900 using LDV and HWA showing below the relative gap height 0.4 periodic activity cannot be observed on the cylinder wake.

Numerical three dimensional simulations were performed out by Mahir (2009) in order to investigate the flow structure in the wake of a square cylinder placed near a plane wall by implicit finite difference method to solve the Navier Stokes equations. In this work the gap ratio (G/D) was varied from 0.2 to 4 for Reynolds number of 175 up to 250. The mean drag and lift coefficients obtained slightly decreases at the large gap ratios and these values differ from that of an isolated cylinder in a free stream case. Other numerical simulation in high Reynolds number are the work of Liou et al. (2002) at Re = 2.2×104, Lohász et al. (2005) at Re = 4×104 and

Unfortunately, most of the studies using a square cylinder placed near a flat wall were carried out in relative high Reynolds numbers. Hydrodynamic flow visualization and hot-fim anemometry have been applied in the present study of a flow around a square cylinder positioned in a smooth flat plate in relative gap equals to 0.5. Results, showing flow visualized images and the vortex shedding frequency, have been obtained for Reynolds number, based in the height of the square cylinder, up to 1,000.

FLOW VISUALIZATION TOOLS Flow visualization is perhaps the most useful experimental tool in obtaining, by a fast and cheap way, qualitative information on flow structure. Many experimental flow visualization techniques are described in the technical literature, only to exemplify, Yang (1989), Merzkirch (1987), Goldstein (1976) and Freymuth (1993), among several other references. In a broad sense, flow visualization is more easily performed in aqueous than in air medium, since there are many ways of making water visible by means of different color inks, gaseous bubbles or solid dust in suspension. Werlé (1973) and (1974) details several hydrodynamics facilities and flow visualization techniques applicable in water in low velocities.

Generally, water tunnels have smaller test section dimensions when compared with aerodynamic tunnels. Therefore, in the hydrodynamic tunnel case, small bodies tested with moderate blockage ratios, implies relative low Reynolds numbers. In spite of the water kinematics viscosity might be less than 15 times the air kinematics viscosity, in the same temperature, the Reynolds numbers obtained in hydrodynamics tunnels are smaller than the Reynolds obtained in aerodynamics tunnels. Relatively lower Reynolds numbers stands a


ICEM15 3

serious restriction to the use of water tunnels in substitution to aerodynamic tunnels. Erickson (1980), motivated by the possibility of vortex wake flow visualization in hydrodynamics medium, debates, in details, this problem of low Reynolds tests in water tunnels and the care in the results interpretation.

In this paper, in order to obtain the preliminary results of the effort work for knowledge of flow over square cylinders positioned near flat smooth plates, two flow visualization techniques utilizing an opaque liquid dye were employed. In the first method traditional direct injection technique has been used. In this technique a long hypodermic needle (E.D. = 0,5 mm) is adequately positioned upstream the test model and a black opaque liquid dye is injected in the free flow, permitting the visualization of streaklines.

The second flow visualization method utilized in this work, named dye wash technique, employed liquid dye, Large amount of a dense black dye is injected close of the test model, enough to tint the entire flow field. Suddenly, the dye injection is stopped; the hypodermic needle is taken off the flow, in order to minimize flow perturbations. The clean water flow washes the entire flow field, except in the cylinder wake, since in this region the flow speed is significantly smaller than in other regions. This procedure allows, during some seconds, for the wake downstream the cylinder to be watched. According to Lindquist (2010) and Bassan et al. (2011b), dye wash technique is strongly recommended visualization of the recirculation zone near the leeward face of the solid obstacle.

EXPERIMENTAL APPARATUS AND METHOD All experiments have been carried out in a low turbulence hydrodynamic tunnel with a closed vertical test section. As described by Bassan et al. (2011a), the tunnel is operated by gravitational action in blow-down mode producing a low turbulence level (less than 0,2%). The experiments have been performed for Reynolds numbers up to 1,000 in a hydrodynamic tunnel with a vertical test section of 146 × 146 × 500 mm. The volumetric flow rate in the tunnel has been determined utilizing a Yokogawa electromagnetic flowmeter. The uncertainty in the free stream velocity is less than 4%, in more adverse case, producing a maximum uncertainty less than 5% in the Reynolds number (based in the square height, i.e. 10 mm). Fig. 1 shows the non perturbed free stream velocity and the turbulence level produced in the test section.

0,0 0,2 0,4 0,6 0,8 1,0 1,20,0

0,1

0,2

0,3

3,60

3,65

Turb

ulen

ce le

vel

%

Vel

ocity

[ cm

/ s

]

time [ s]

Velocity Mean velocity Relative turbulence level Mean relative turbulence level

Fig. 1. Free stream velocity and turbulence level in the hydrodynamic tunnel test section.



Fig. 2 shows the square cylinder positioned in a smooth flat plate. The non perturbed free stream is U∞ and the cylinder side length (D) is equal to 10 mm. The square cylinder is positioned at 55 mm from the edge of plate and the gap length (G) is equal to 5 mm.

Fig. 2. Square cylinder placed near on a smooth flat plate with a gap.

All images have been captured using a D 90 Nikon DSLR camera equipped with a special Nikkor macro lens with 60 mm and f/1:4. The very expensive medical Nikkor macro lens was designed for application in full frame (24 ×36 mm) chemical 35 mm roll film cameras and adequately adapted for a half frame (23.6×15.8 mm) digital camera resulted in an excellent optical device for capture close up images.

Cold illumination by means of fluorescent lamps with high color temperature, but minimal heat emission, has been adapted in the tunnel allowing sharp and well defined images. A Rosco color illuminating filter Cinegel#3308 converts daylight fluorescent lamps to 5,500 K, while a diffuser Cinegel#3007, a slight filter with less density softens edge, provides a good illumination for still and video image capture.

Velocity measurements have been performed with a 55R11 fiber-film probe made by Dantec Measurement Technology, with 70 μm diameter quartz fiber coated with 2 μm nickel film and with an overall length of 3 mm. This is a straight general-purpose type sensor which permits a wide measurement range in water medium. For very small velocities (up to 0.10 m/s) several special cares could be adopted in order to reduce the convection effect around the probe. Indeed, a hot-wire probe immersed in recirculation zones can produce a high level of thermal convection interfering in the measurements, like discussed by Lindquist et al. (2010). A Dantec StreamLine 90C10 frame with 3 CTA modules 90C10 permits simultaneously measurements in 3 channels. An A/D board NI-DAQmx 8.7.1 (16 bits), made by National Instrument, has been utilized in order to record the output voltage signal. Single element hot-film measurement downstream the protuberance were employed to obtain temporal flow velocity fluctuations. Data acquired by hot-film probe have been processed to obtain a frequency spectrum with a FFT - Fast Fourier Transform.

RESULTS

Flow visualized images are depicted in Fig. 3 for low Reynolds numbers. In Fig. 3(a) for Reynolds equals to 3, obtained in very low velocity, shows a very stable shear layer downstream the square cylinder. The velocity field at the bottom side of the cylinder is significantly different of the velocity field in the cylinder upper side generating a non-

55 mm

10 mm

Gap 5 mm

U∞


ICEM15 5

symmetry shear layer. In Fig. 3(b) for a higher Reynolds number equals to 9, a very small recirculating bubble can be observed on the downstream the cylinder at the beginning of the shear layer. The size of this recirculation bubble rises for higher Reynolds as showed in the Fig. 3(c). For very low Reynolds numbers, near one, the shear layer starts very close to the bottom face – Fig. 3 (a). For more elevated Reynolds number, the shear layer start tends to move to the upper face of the square cylinder – Fig. 3(d). At Reynolds equals to 48 – Fig. 3(e) – the detachment occurs on the top edge. For Reynolds less than 48 – Fig. 3 (e) – here is a steady wake and only one point of detachment. However, for Reynolds equals to 58, two detachment points are observed on the rear face of the test body. A detachment occurs downstream the test body in the flat plate showing a severe adverse pressure gradient and the wake turns unsteady.

Fig. 4 shows, for Reynolds 58, a temporal sequence of images of a close up view of the asymmetric and unsteady wake formation.

In Fig. 5 is possible to visualize the development of two large asymmetric recirculating bubbles. The growth of the recirculating bubbles increases wake instability.

Finally, to higher Reynolds number is observed the formation and shedding of the vortex, as shown in Fig. 5.

(a) Re ≈ 3 (b) Re ≈ 9

(c) Re ≈ 20 (d) Re ≈ 40

(e) Re ≈ 48 (f) Re ≈ 58

Fig. 3. Flow in very low Reynolds number.



(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

(j) (k) (l)

(m) (n) (o)

Fig. 4. Temporal evolution of the asymmetric and unsteady wake at Re ≈ 58.

(a) Re ≈ 96 (b) Re ≈ 145

Fig. 5. Unsteady and asymmetric recirculating bubbles.


ICEM15 7

(a) Re ≈ 272

(b) Re ≈ 557

(c) Re ≈ 804

(d) Re ≈ 993

Fig. 6. Vortex generation and shedding.

CONCLUSION Experimental flow visualization plays a key role in understanding of complex flow phenomena. That important experimental tool has been utilized in this work in order to obtain the first results of the study of flow around a square cylinder positioned near a flat smooth plate The flow around a square cylinder placed in the vicinity of a solid plane wall can be shows a great complexity involving the development of various shear layers, including those separated from the upper and lower sides of the cylinder, as well as the boundary layer. The



presence of a wall creates an asymmetric condition in the velocity and pressure fields around the cylinder. Also is possible to see a non uniform velocity profile over the wall in front of the cylinder. This non uniformity generates a shear layer with an asymmetric influence on the test body. Probably, this non symmetric influence results in a non zero lift force when the height gap (G) is small. The plane wall imposes a severe non rotational restriction to the cylinder wake resulting in a vortex shedding and wake development different from those of the alone square cylinder far the wall.

A lot of video tapes showing the vortex wake have been captured showing a defined vortex shedding frequency. But, a detailed video image exam reveals several instabilities in the vortex frequency. Possibly, the vortex wake shows a weak stability of the vortex frequency also named by weak periodicity. Quantitative data measurements of the vortex wake frequency are needed in order to obtain the data correlation showing the stability of vortex frequency behavior. Further tests using an HWA should be performed in order to verify the vortex shedding frequency stability.

There are some limitations in the present work, such as relatively low Reynolds numbers of the tests. Low Reynolds numbers are due to reduced model dimensions implying in severe difficulties in order to visualize fine flow details. Construction of a new hydrodynamic tunnel facility with a large test section, able to host large test model, is absolutely need.

ACKNOWLEDGMENTS The experiments have been, in part, financed by FAPESP. Recognition is likewise due to Fundunesp and PROPP/Unesp.

REFERENCES Bassan, RA, Mansur, SS, Vieira, EDR. Rebuilding and testing a vertical hydrodynamic tunnel, Proceedings of the 21st Brazilian Congress of Mechanical Engineering – COBEM, 2011(a), Natal, Brazil.

Bassan, RA, Mansur, SS, Vieira, EDR. Visualização experimental de escoamentos no interior de canais munidos de protuberância parietal, Proceedings of the 10º Congresso Iberoamericano de Engenharia Mecânica – CIBEM, 2011(b), Porto, Portugual.

Bosch, G, Kappler, M and Rodi, W, Experiments on the flow past a square cylinder placed near a wall, Experimental Thermal and Fluid Science, 1996, 13, p.292-305.

Erickson, GE, Vortex Flow Visualization, AFWAL-TR-80-3143, 1980.

Freymuth P, Bank W and Palmer M. Flow visualization and hot-wire anemometry”, TSI Quartely, 1983, 9 (4), p.11-14.

Freymuth, P, Flow visualization in fluid mechanics, Revue Science Instruments, 1993, 64, p.1-18.

Goldstein, RJ (ed.), Fluid mechanics measurements, Hemisphere Publishing Co., 1976.

Kim, HB and Lee, SJ, Time-resolved velocity field measurements of separated flow in front a vertical fence, Experiments in Fluids, 2001, 31, p. 249-257.


ICEM15 9

Lindquist, C, Vieira, EDR and Mansur, SS, Flow around square cylinders in several attack angles, Proceedings of the 13th Brazilian Congress of Thermal Sciences and Engineering – ENCIT, 2010, Uberlândia, Brazil.

Liou, T-M, Chen, S-H and Hwang, P-W, Large eddy simulation of turbulent wake behind a square cylinder with a nearby wall, Journal of Fluids Engineering, 2002, 124, p. 81-90.

Lohász, MM, Rambaud, P and Benocci, C, Flow features in a fully developed ribbed duct flow as a result of LES, Enginnering Turbulence Modeling and Experiments, Enginnering Turbulence Modeling and Experiments, Rodi, W (Ed.), 2005, 6, p.267-276.

Mahir, N, Three-dimensional flow around a square cylinder near a wall, Ocean Engineering, 2009, 36, p.357-367.

Martinuzzi, R and Tropea, C, The flow around surface-monted, prismatic obstacles placed in a fully developed channel flow, Journal of Fluids Engineering, 1993, 115, p. 85-92.

Martinuzzi, RJ, Bailey, SCC and Kopp, GA, Influence of a wall proximity on vortex shedding from a square cylinder, Experiments in Fluids, 2003, 34, p. 585-596.

Merzkirch, W, Flow visualization, 2nd ed., Academic Press, Orlando, 1987.

Shi, LL, Liu, YZ and Sung, HJ, On the wake with and without vortex shedding suppression behind a two-dimensional square cylinder in proximity to a plane wall, Journal of Wind Engineering and Industrial Aerodynamics, 2010, 98, p. 492-503.

Wang, XK and Tan, SK, Comparation of the flow patterns in the near wake of a circular cylinder and a square cylinder placed near a plane wall”, Ocean Engineering, 2008, 35, p.458-472.

Werlé, H, Hydrodynamic flow visualization, Annual Review of fluid Mechanics, 1973, 5, p.361−382.

Werlé, H, Le Tunnel Hydrodynamic au Service de la Recherche Aérospatiale, ONERA Publication, 1974, 156, p.86.

Yang, W.-J. (ed.), Handbook of Flow Visualization, Taylor & Francis, Ann Arbor, 1998.

steady and unsteady flow visualization inside a …phoenics/em974/projetos/temas projetos/vortex...

Documents