• unlimited access with print and download
    $ 37 00
  • read full document, no print or download, expires after 72 hours
    $ 4 99
More info
Unlimited access including download and printing, plus availability for reading and annotating in your in your Udini library.
  • Access to this article in your Udini library for 72 hours from purchase.
  • The article will not be available for download or print.
  • Upgrade to the full version of this document at a reduced price.
  • Your trial access payment is credited when purchasing the full version.
Buy
Continue searching

Experimental investigations on complex vortex flows using advanced flow diagnostic techniques

ProQuest Dissertations and Theses, 2009
Dissertation
Author: Zifeng Yang

TABLE OF CONTENTS

LIST OF TABLES .................................... ................................................... . VI

LIST OF FIGURES ................................... ................................................. V II

CHAPTER 1. INTRODUCTION TO THE THESIS ........ ..................... 1

CHAPTER 2. CHARACTERISTICS OF TORNADO-LIKE WINDS .. 4

2.1.

I NTRODUCTION TO C HARACTERISTICS OF T ORNADOES ......................... 4

2.2.

ISU

T ORNADO S IMULATOR AND E XPERIMENTAL S ETUP ....................... 7

2.3.

C HARACTERISTICS OF T ORNADO - LIKE F LOW ...................................... 10

CHAPTER 3. FLOW AROUND A BUILDING MODEL IN TORNADO-LIKE WINDS ................................. ......................................... 20

3.1.

I NTRODUCTION TO W IND L OADS AND F LOW S TRUCTURES OVER A B UILDING M ODEL IN T ORNADO - LIKE W IND ............................................... 20

3.2.

F LOW AROUND A H IGH - RISE B UILDING M ODEL IN T ORNADO - LIKE W INDS .................................................. ................................................... ..... 22

3.2.1. Wind Loads and PIV Measurements on a High-rise Building Model in Straight-line Winds .................................................................................. 22

3.2.2. Wind Loads acting on a High-rise Building Model in Tornado-like Winds ....................................................................................................... 29

3.2.3. Flow Structure around a High-rise Building Model in Tornado-like Winds ....................................................................................................... 34

3.3.

F LOW AROUND A L OW - RISE G ABLE - ROOFED B UILDING M ODEL IN T ORNADO - LIKE W INDS .................................................. ............................... 41

3.3.1. Wind Loads and PIV Measurements in Straight-line Winds ................. 41

iv

3.3.2. Wind Loads and Pressure acting on a Gable-roofed Building Model in Tornado-like Winds ................................................................................. 48

3.3.3. Flow Structure around a Gable-roofed Building Model in Tornado-like Winds ................................................................................. 55

CHAPTER 4. EXPERIMENTAL STUDY OF TRUBINE BLADE TRAILING EDGE COOLING ............................. ...................................... 65

4.1.

I NTRODUCTION TO T RAILING E DGE C OOLING ..................................... 65

4.2.

P RESSURE S ENSITIVE P AINT T ECHNIQUE ............................................. 70

4.2.1. Principal Physics of Pressure Sensitive Paint ......................................... 70

4.2.2. PSP Calibration....................................................................................... 73

4.3.

T HE C OOLING E FFECT M EASUREMENT BY U SING PSP ........................ 77

4.3.1. Experimental Apparatus ......................................................................... 77

4.3.2. Results and Discussion ........................................................................... 83

4.3.3. Error Analysis on the Cooling Effectiveness ......................................... 93

4.4.

PIV

M EASUREMENTS ON THE T RAILING E DGE C OOLING F LOW .......... 95

4.4.1. Experimental Apparatus ......................................................................... 95

4.4.2. Experimental Uncertainties .................................................................... 99

4.4.3. Results and Discussion ........................................................................... 99

CHAPTER 5. CONCLUSIONS AND FUTURE WORK ............ .......... 120

5.1.

T HE C HARACTERISTICS OF A T ORNADO - LIKE V ORTEX ...................... 120

5.2.

F LOW - STRUCTURE I NTERACTION FOR A B UILDING M ODEL IN A T ORNADO - LIKE V ORTEX .................................................. .......................... 121

5.3.

F ILM C OOLING E FFECT ON A T RAILING E DGE ................................... 122

5.4.

T HE C HARACTERISTICS OF THE T RAILING E DGE C OOLING F LOW ...... 123

5.5.

F UTURE W ORK .................................................. ................................ 124

APPENDIX A. THE DESIGN OF CALIBRATION CELL ........ ........... 125

v

APPENDIX B. THE DESIGN OF A TRAILING EDGE MODEL . .... 129

APPENDIX C. SYCHRONIZATION SETUP FOR PIV .......... ............. 131

APPENDIX D. THE PRINCIPLE OF STEREOSCOPIC PIV ..... ........ 133

REFERENCE ......................................... ................................................... .. 136

ACKNOWLEGEMENTS .................................... ...................................... 145

vi

LIST OF TABLES T ABLE C-1.

T HE TIMING SETTING FOR PIV MEASUREMENT ............................................................... 132

vii

LIST OF FIGURES F IGURE 2.1.

A TYPICAL STRUCTURE OF A TORNADO .............................................................................. 5

F IGURE 2.2.

S CHEMATIC AND PICTURE OF THE ISU TORNADO SIMULATOR ........................................... 8

F IGURE 2.3.

E XPERIMENTAL SETUP FOR PIV MEASUREMENTS .............................................................. 9

F IGURE 2.4.

T YPICAL INSTANTANEOUS STREAMLINE PATTERNS IN THE LOW EST MEASURED HORIZONTAL PLANE (Z=4 MM ) ..................................................................................................... 12

F IGURE 2.5.

D ISTRIBUTION OF INSTANTANEOUS VORTEX CENTERS AT THE LOWEST HORIZONTAL PLANE (Z=15 MM ) ......................................................................................................................... 12

F IGURE 2.6.

T HE FLOW CHARACTERISTICS OF THE TORNADO - LIKE VORTEX ....................................... 13

F IGURE 2.7.

V ELOCITY FLOW FIELD WITH STREAMLINES AND VORTICITY F IELD IN HORIZONTAL PLANES AT VARIOUS ELEVATIONS (Z=4,

65 AND 150 MM ABOVE THE GROUND ) ......................... 14

F IGURE 2.8.

T HE FLOW CHARACTERISTICS OF THE TORNADO - LIKE VORTEX ....................................... 16

F IGURE 2.9.

T HE SIMULATED TORNADO - LIKE VORTEX VS .

M ULHALL AND S PENCER TORNADOES ..... 18

F IGURE 2.10.

A PLANAR VIEW OF A TORNADO - LIKE VORTEX IN EACH HORIZONTAL PLANE ............... 19

F IGURE 3.1.

T EST MODEL AND DEFINITION OF PARAMETERS ............................................................... 23

F IGURE 3.2.

A PICTURE OF THE ROUGHNESS SETUP .............................................................................. 24

F IGURE 3.3.

V ELOCITY AND TURBULENCE INTENSITY PROFILES ......................................................... 24

F IGURE 3.4.

A PICTURE OF THE TEST MODEL AND JR3 LOAD CELL ...................................................... 27

F IGURE 3.5.

W IND LOADS ACTING ON THE TEST MODEL VS . ORIENTATION ANGLES ........................... 27

F IGURE 3.6.

PIV MEASUREMENTS AROUND THE TEST MODEL AT Z=70 MM IN STRAIGT - LINE WIND .... 28

F IGURE 3.7.

W IND LOADS VS . THE DISTANCE BETWEEN THE TEST MODEL AND THE TORNAD O - LIKE VORTEX ........................................................................................................................................ 30

F IGURE 3.8.

T HE MEASURED WIND LOADS VS . ORIENTATION ANGLES ................................................. 32

F IGURE 3.9.

S CHEMATIC OF THE FLOW STREAMS APPROACHING THE TEST M ODEL IN STRAIGHT - LINE AND TORNADO - LIKE WINDS ......................................................................................................... 33

F IGURE 3.10.

F LOW FIELD AROUND THE TEST MODEL VS . THE DISTANCE BETWEEN THE CENTERS OF THE TEST MODEL AND THE TORNADO - LIKED VORTEX ................................................................. 35

F IGURE 3.11.

F LOW FIELD AROUND THE TEST MODEL VS . ORIENTATION ANGLE AT R/R 0 =1 ............... 38

F IGURE 3.12.

F LOW FIELD AROUND THE TEST MODEL AT DIFFERENT ELEVAT ION LEVELS AT R/R 0 =1 40

F IGURE 3.13.

T HREE - DIMENSIONAL VIEW OF THE WAKE VORTEX STRUCTURES AROU ND THE MODEL AT R/R 0 =1 WITH OA=0 DEG ......................................................................................................... 41

viii

F IGURE 3.14.

T EST MODEL AND DEFINITION OF PARAMETERS ............................................................. 43

F IGURE 3.15.

P RESSURE TAP DISTRIBUTION ......................................................................................... 43

F IGURE 3.16.

PIV EXPERIMENTAL SETUP ............................................................................................. 44

F IGURE 3.17.

F ORCE AND MOMENT COEFFICIENTS VS . ORIENTATION ANGLES .................................... 45

F IGURE 3.18.

PIV MEASUREMENTS ON A VERTICAL PLANE (Y=0 MM ) IN STRAIGT - LINE WIND ............ 46

F IGURE 3.19.

PIV MEASUREMENTS ON A HORIZONTAL PLANE (Z=15 MM ) IN STRAIGHT - LINE WIND .. 47

F IGURE 3.20.

W IND LOADS VS . THE DISTANCE BETWEEN THE TEST MODEL AND THE TORNAD O - LIKE VORTEX ........................................................................................................................................ 49

F IGURE 3.21.

M EASURED TYPICAL PRESSURE DISTRIBUTIONS AROUND THE T EST MODEL .................. 53

F IGURE 3.22.

W IND LOADS VS . ORIENTATION ANGLES ........................................................................ 55

F IGURE 3.23.

F LOW FIELD AROUND THE TEST MODEL VS . THE DISTANCE OF BETWEEN THE CENTERS OF THE TEST MODEL AND THE TORNADO - LIKED VORTEX ............................................................ 60

F IGURE 3.24.

F LOW FIELD AROUND THE TEST MODEL VS . THE ORIENTATION ANGLES ........................ 61

F IGURE 3.25.

T HREE - DIMENSIONAL FLOW FIELD WITH STREAMLINES AND SWIRLIN G STRENGTH CONTOURS IN Y AND Z DIRECTIONS ............................................................................................ 63

F IGURE 4.1.

S CHEMATIC DRAWING OF A TYPICAL TURBINE BLADE WITH A CUTBACK TRAILING EDGE .............................................................................................................................................................. 66

F IGURE 4.2.

M ODEL OF PRESSURE SENSITIVE PAINT ............................................................................ 71

F IGURE 4.3.

T HE PICTURE AND SCHEMATIC OF THE PSP CALIBRATION SETUP .................................... 74

F IGURE 4.4.

I NTENSITY CONTOURS OBTAINED FROM CALIBRATION : A .

I NTENSITY FROM REFERENCE IMAGE ; B .

I NTENSITY FROM VACUUM IMAGES P/P REF =0.23; C .

I NTENSITY RATIO I REF /I;

D .

E RROR L EVEL % ...................................................................................................................... 75

F IGURE 4.5.

PSP CALIBRATION CURVE ................................................................................................. 76

F IGURE 4.6.

E RROR LEVEL OF THE INTENSITY RATIO VS . PRESSURE RATIO ......................................... 76

F IGURE 4.7.

PSP

E XPERIMENTAL S ETUP .............................................................................................. 77

F IGURE 4.8.

D IMENSIONS OF STUDIED TRAILING EDGE COOLING MODEL ............................................ 80

F IGURE 4.9.

S CHEMATIC OF THE EXPERIMENTAL SETUP ...................................................................... 81

F IGURE 4.10.

T HE PICTURE OF THE EXPERIMENTAL SETUP .................................................................. 82

F IGURE 4.11.

A DIABATIC FILM COOLING EFFECTIVENESS ON THE TRAILING EDGE FOR DIFFERENT BLOW RATIOS M ........................................................................................................................... 84

F IGURE 4.12.

C OOLING EFFECTIVENESS PROFILES FOR THE CASE WITHOUT LANDS ............................ 86

F IGURE 4.13.

S LOT GEOMETRY FOR THE FILM COOLING OF A SURFACE IN M UKHERJEE ’ S MODEL ...... 87

ix

F IGURE 4.14.

S PANWISE AVERAGED C OOLING EFFECTIVENESS PROFILES FOR THE CASE WITHOUT LANDS .......................................................................................................................................... 88

F IGURE 4.15.

A DIABATIC FILM COOLING EFFECTIVENESS (C.E.) ON THE TRAILING EDGE CUTBACK FOR DIFFERENT BLOW RATIOS M ................................................................................................. 89

F IGURE 4.16.

C OOLING EFFECTIVENESS PROFILES FOR THE CASE WITH LAN DS ................................... 90

F IGURE 4.17.

S PANWISE AVERAGED COOLING EFFECTIVENESS PROFILES FOR THE CASE WITH LANDS .......................................................................................................................................... 91

F IGURE 4.18.

C OMPARISON OF COOLING EFFECTIVENESS PROFILES .................................................... 93

F IGURE 4.19.

S CHEMATIC OF THE CAMERA VIEWS ............................................................................... 94

F IGURE 4.20.

T HE SCHEMATIC OF THE PIV SETUP FOR FLOW MEASUREMENT ..................................... 98

F IGURE 4.21.

PIV MEASUREMENT RESULTS AT BLOW RATIO OF M

=0.43 ......................................... 101

F IGURE 4.22.

PIV MEASUREMENT RESULTS AT BLOW RATIO OF M

=0.64 ......................................... 102

F IGURE 4.23.

PIV MEASUREMENT RESULTS AT BLOW RATIO OF M

=0.76 ......................................... 102

F IGURE 4.24.

PIV MEASUREMENT RESULTS AT BLOW RATIO OF M

=1.1 ........................................... 103

F IGURE 4.25.

PIV MEASUREMENT RESULTS AT BLOW RATIO OF M

=1.6 ........................................... 103

F IGURE 4.26.

S TREAMWISE VELOCITY PROFILES ON THE CENTER VERTICAL PLANE WITH VARIOUS BLOW RATIOS AT DIFFERENT DISTANCES AFTER THE EXITS ...................................................... 104

F IGURE 4.27.

T HREE - DIMENSIONAL VELOCITY VECTOR AND AXIAL VELOCITY CONT OUR AT BLOW RATIO OF M=0.43 AT FIVE DIFFERENT DISTANCES AFTER THE SLOT EXIT : A . WITHOUT LANDS B . WITH LANDS ........................................................................................................................... 107

F IGURE 4.28.

T HREE - DIMENSIONAL VELOCITY VECTOR AND AXIAL VELOCITY CONT OUR AT BLOW RATIO OF M=0.64 AT FIVE DIFFERENT DISTANCES AFTER THE SLOT EXIT : A . WITHOUT LANDS B . WITH LANDS ........................................................................................................................... 108

F IGURE 4.29.

T HREE - DIMENSIONAL VELOCITY VECTOR AND AXIAL VELOCITY CONT OUR AT BLOW RATIO OF M=0.76 AT FIVE DIFFERENT DISTANCES AFTER THE SLOT EXIT : A . WITHOUT LANDS B . WITH LANDS ........................................................................................................................... 109

F IGURE 4.30.

T HREE - DIMENSIONAL VELOCITY VECTOR AND AXIAL VELOCITY CONT OUR AT BLOW RATIO OF M=1.1 AT FIVE DIFFERENT DISTANCES AFTER THE SLOT EXIT : A . WITHOUT LANDS B . WITH LANDS ........................................................................................................................... 110

F IGURE 4.31.

T HREE - DIMENSIONAL VELOCITY VECTOR AND AXIAL VELOCITY CONT OUR AT BLOW RATIO OF M=1.6 AT FIVE DIFFERENT DISTANCES AFTER THE SLOT EXIT : A . WITHOUT LANDS B . WITH LANDS ........................................................................................................................... 111

x

F IGURE 4.32.

A VERAGED STREAMWISE VORTICITY CONTOUR WITH VARIED BL OW RATIOS AT THREE DIFFERENT DISTANCES AFTER THE EXITS FOR THE C ASES WITH LANDS ......................... 113

F IGURE 4.33.

F LOW FIELD AT DIFFERENT CROSS PLANES FOR M=0.43 AND M=1.6 .......................... 114

F IGURE 4.34.

F LOW FIELD COMPARISONS AT DIFFERENT DISTANCE DOWNSTR EAM .......................... 116

F IGURE 4.35.

F LOW FIELD COMPARISONS AT DIFFERENT DISTANCE DOWNSTR EAM .......................... 117

F IGURE 4.36.

F LOW STRUCTURES IN THE TRAILING EDGE OF TURBINE BLAD E .................................. 118

F IGURE 4.37.

A VERAGED 3-D TURBULENT KINETIC ENERGY AT X/H=8 AT BLOW RATIO OF M=1.6 . 119

F IGURE B-1.

A SCHEMATIC OF THE TRAILING EDGE OF A TURBINE BLADE ........................................ 129

F IGURE B-2.

T HE UPPER PART OF THE TRAILING EDGE MODEL .......................................................... 129

F IGURE B-3.

T HE LOWER PLATE OF THE TRAILING EDGE MODEL ....................................................... 130

F IGURE B-4.

T HE LAND OF THE TRAILING EDGE MODEL .................................................................... 130

F IGURE C-1.

T HE SYNCHRONIZATION SETUP FOR LASER AND CAMERA ............................................ 131

F IGURE D-1.

S CHEIMPFLÜG CONFIGURATION FOR STEREOSCOPIC PIV ............................................. 133

F IGURE D-2.

I MAGE PERSPECTIVE EFFECT DUE TO CAMERA TILT IN S CHEIMPFLÜG ARRANGEMENT 135

1

CHAPTER 1. INTRODUCTION TO THE THESIS Everyone has noted at some time swirling leaves on a windy autum n day. Circle eddies on the surface of a river can be observed when one plays ducks and drakes. What is the attraction of this phenomenon? Is there an unknown force behind the fast sw irling movement of matter? Until 19 th century, humans have thought that supernatural forces were the causes of vortices. Again and again the danger and mystery of vortical moti ons, such as tornadoes, have excited fantasy and superstition. Today, however, no student of scie nce believes in supernatural causes on the cryptic meaning of vortices, but both danger and mystery have remained characteristic of many vortex flows. Hurricanes and t ornadoes still occur all around the world, and many people die when these disasters happen. Obviously vortices play much more powerful roles in nature than merely whirling leaves or eddie s in a river suggest. Thus it is not surprising that the vortex concept is of central importanc e in the history of science and philosophy. Modern science has extended human understanding of vortices and their role in nature or aerodynamic field in two ways: It has clarified the basic si gnificance of vortical motion in mechanics, and it has extended the spatial horizon of humans so that they can observe both large-scale and micro-scale vortices and interpret them. Today on e can follow the movements of hurricanes on satellite photographs and can visualize vor tical flow structure in micro-scale by using advanced diagnostic techniques, such as Part icle Image Velocimetry (PIV). Therefore, the vortex flow problems are not that mysterious today. Using advanced diagnostic techniques, two complex vortex flow topics were addressed in the present thesis. First, the characteristics of flow struct ures around building models in a

2

tornado-like vortex were studied by using a high-resolution Particle Image Velocimetry (PIV) system. Second, the 3D vortical structures of the film cooling flows in the trailing edge of a turbine blade were studied by using a stereoscopic PIV system. The flow measurement results were correlated to the film cooling effectiveness me asured by the application of a relatively new technique, pressure sensitive paint (PSP). In the study of tornado-like vortex, by using the world-largest tornado simulator of Iowa State University, a comprehensive PIV study on the flow struct ures around a high-rise building model and a gable-roofed building model, as well as the surface pressure measurements and force measurements, was conducted to elucidate the underlying physics. The ultimate objective of the present study is to quantify the su rface winds generated by tornadoes and flow-structure interactions between tornadoes and built e nvironments to assess wind-induced damage with the purpose of mitigating damage and improving public safety. The characteristics of tornado like flow will be demonstrated and di scussed in Chapter 2. The flow features around a building model and flow-structure interacti ons will be discussed in Chapter 3. In the study of vortex flow in the trailing edge of a turbine blade, detailed distributions of film cooling effectiveness measurements were obtained in the cut back region of trailing edge by using the PSP technique. Before the application of PSP, a self-designe d calibration facility was developed to study the characteristics and to complete the ca libration of the PSP. Corresponding to the film cooling effectiveness measurements, an exper imental study was conducted to quantify the characteristics of coolant flows in t he cutback region at the trailing edge of a turbine blade. A high-resolution stereoscopic PIV system were used to conduct detailed flow measurements to quantitatively visualize the evoluti on of the unsteady vortex

3

and turbulent flow structures in coolant jet streams, and to quantify t he mixing process between the wall jet cooling streams and main streams. The f ilm cooling effectiveness and flow characteristics in the cutback region at the trailing e dge of a turbine blade will be discussed in Chapter 4.

4

CHAPTER 2. CHARACTERISTICS OF TORNADO-LIKE WINDS

2.1. Introduction to Characteristics of Tornadoe s Tornadoes are violently rotating columns of air and are considered n ature’s most violent storms. Tornadoes may occur wherever conditions favor the development of strong thunderstorms. Essential conditions for such storms are the presence of cool, dry air at middle levels in the troposphere, overlying a layer of moist, condit ionally unstable air near the surface of the Earth. Conditional instability occurs when a s aturated air parcel continues to rise once set in motion. When conditionally unstable air rises, it becomes warmer owing to the condensation of water vapor. As the water condenses, heat is rele ased, further warming the air and fueling its rise. This convective action (i.e., the circulation of air as a result of heat transfer) produces the huge clouds commonly associated with thunderstorm s and tornadoes. Convection can be initiated when the sun heats a localized area of the ground, destabilizing the near-surface air. Most tornadoes are formed when a strong updraft, such as those des cribed above, acts to concentrate atmospheric rotation, or spin, into a swirling column of ai r, as shown in Figure 2.1. Spin is a natural occurrence in air because horizontal winds alm ost always experience both an increase in speed and a veering in direction with increasi ng height above the ground. The increase of wind speed with height (called vertical speed sh ear) produces “crosswise spin,” that is rotation about a horizontal axis crosswise to the dire ction of wind flow. When air containing crosswise spin flows into an updraft, the spin is dra wn upward, producing rotation about a vertical axis. The veering of wind direction with he ight (vertical direction shear) is another source of horizontal spin, this time oriented in the same direction as the

5

wind flow and known as “streamwise spin.” When air containing streamw ise spin is drawn into an updraft, it too is tilted upward and rotates about a vertical a xis. Although crosswise spin and streamwise spin are oriented at right angles to each other, both rotations exist in the horizontal plane, and both types have been revealed by Doppler radar obser vations to contribute to the evolution of a rotating updraft.

Figure 2.1. A typical structure of a tornado

In an average year, 800 ~ 1000 tornadoes occur in the U.S. alone, and cause about 80 deaths (on average), over 1500 injuries, and $850 million worth of property da mage [1]. Tornados are inherently destructive and complicated in appearance, incl uding variety of shapes (such as thin ropes, cylindrical columns, funnels etc.), sizes a nd configurations under a wide variety of circumstances. All have a central core ( i.e. tornado eye) of organized and concentrated vorticity. Therefore, the flow field in a tornado is m uch different from the straight-line, boundary-layer wind. Tornado flows have been extensivel y studied by conducting field observations [2-4], numerical simulation and experimental simulation using

6

a vortex chamber or tornado simulator. Many laboratory simulator de signs have been based on the pioneering work of Ward [5]. For laboratory tornado simulators, t he common flow parameter has historically been the swirl ratio. Essentially a measure of the relative amount of angular to radial momentum in the vortex, the swirl ratio was expressed by Ward [5] and Church et al. [6] as:

Г

(1) where Г is the circulation, , r 1 is the radius of the domain, Q is the inflow rate, V θ is the maximum tangential velocity, V r is radial velocity at r 1 and a is the aspect ratio defined as:

(2) where h is the inflow depth. The swirl ratio is a measure of the amount of rotational ener gy in the vortex relative to the convective energy in the vortex and correl ates well with vortex structure. Typical value of swirl ratio ranges from 0.1~6.

Subsequent efforts—based on the Ward model—at Purdue University [6], the University of Oklahoma [7] and that of Davies-Jones [8] employed various means to improve the similarity between laboratory simulations and full-scale torn ado events. These laboratory simulations were aimed at obtaining a greater understanding of the tornado vortex itself. However, numerical simulation has overtaken physical simulation as t he tool of choice for tornado vortex studies—both because of cost and because of versatility. W hile both laboratory and numerical simulation efforts have revealed a great deal about tornado structure [9], the problem of numerically simulating a domain large enough to accommodate a tornado vortex with ground-based structures remains computationally impract ical.

7

From the lessons learned from testing several simulator concepts and from some insights into tornadogenesis, a new approach was developed and tested [1]. The new design employs a “rotaing forced downdraft” to loosely match observations from natura l tornadoes. A central fan produces an updraft, and the incoming flow is directed downward t hrough an annular duct. Vanes in the annular duct impart rotation to the flow which leads to the technique being named a “rotating forced downdraft”. The rotating downdraft exits t he duct near the ground plane and diverges with a sizeable portion of the flow moving inward ben eath the fan. This flow feeds the updraft as a circular symmetric inflow. The vorti city present in the inflow is stretched beneath the updraft, forming a tornado that travels a long the ground plane as the entire fan/downdraft-producing mechanism translates. In the present study, a high resolution PIV system was utilize d to conduct whole field measurements to quantify the characteristics of the tornado-li ke vortex generated by the ISU tornado simulator before the test building models were mounted on the ground plane. 2.2. ISU Tornado Simulator and Experimental Setu p Figure 2.2 shows the schematic and a photo depicting the flow circ uit and dimensions of the ISU tornado simulator used in the present study. A circular duct of 5.49m in diameter and 3.35m in height is suspended from a heavy duty overhead crane. A 1.83m di ameter fan (maximum flow rate is 59.0 m 3 /s, 125,000 cfm) is mounted concentrically inside the circular duct to generate a strong updraft. The flow from the fan is redire cted downward in a 0.30m wide annular duct to simulate the rear flank downdraft (RFD) encir clement found in natural tornadoes [10]. Swirling is imparted to the airflow in the duct by adjusting the angle of the vanes at the top of the tornado simulator. The downdraft air diverges upon hitting the ground

8

with most of the flow moving inward toward the fan. The fan updraft st retches the low-level vorticity into a tornado-like vortex. A unique feature of the ISU tor nado simulator is that the tornado-like vortex can travel along the ground plane as the entire fan/downdraft-producing mechanism translates. This translation, along with the fact that there is a adjustable clearance between the translating duct and the ground plane, allows a wide ran ge of building models to be placed in the path of the tornado-like vortex for testing. The IS U tornado simulator can generate a tornado-like vortex with a maximum diameter of 1.2 m and maximum tangential velocity of 14.5 m/s. The maximum swirl ratio achieved is 1.14, and the translation speed of the tornado-like vortex can reach up to 0.8m/s. The vortex height can var y from 1.2m to 2.4m by adjusting the ground plane upward or downward. Further information a bout the design, construction, performance and the quantitative comparisons of the t ornado-like vortex generated by the ISU tornado simulator with tornadoes found in na ture can be found in Haan et al. [1].

a.

Schematic diagram b. Picture of the ISU tornado simulator Figure 2.2. Schematic and picture of the ISU tornado simulator In the present study, the ground floor was fixed at 0.457m below the exit of the outer duct , and the fan speed was fixed at 20Hz (1/3 rd of the maximum speed). The radius of the

9

tornado-like vortex core, R O, was found to be 0.165m ( R O = 0.165m), where the maximum tangential speed ( 0.7 = O V m/s) was observed. Based on the definition of Church et al. [6], the swirl ratio of the tornado-like vortex (i.e., the measure of rotating momentum t o the radial inflow momentum) used in the present study was about 0.1, and the aspect r atio of the tornado-like vortex was about 3.6.

Figure 2.3. Experimental setup for PIV measurements

As shown in Fig. 2.3, a digital Particle Image Velocimetry (PIV) system was used to conduct detailed flow field measurements to quantify the evolution of the unsteady vortex and turbulent flow structures around the transparent test model. The fl ow was seeded with ~ 1

m oil droplets by using a droplet generator. Illumination was provi ded by a double-pulsed Nd:YAG laser (NewWave Gemini 200) adjusted to the second harmonic a nd emitting two pulses of 200 mJ at a wavelength of 532 nm with a repetition rate of 10 Hz. The laser beam

10

was shaped into a sheet by a series of spherical and cylindri cal lenses. The sheet was then adjusted to the plane of interest with a mirror. The thickness of t he laser sheet in the measurement region was about 1.0mm. A high resolution 12-bit CCD camer a (Pixelfly, CookeCorp) was used for PIV image acquisition with the axis of the camera perpendicular to the laser sheet. The CCD camera and the double-pulsed Nd:YAG las ers were connected to a workstation (host computer) via a Digital Delay Generator (Ber keley Nucleonics, Model 565), which controlled the timing of the laser illumination and the im age acquisition. More detail about the times series setup can be found in Appendix C. Instanta neous PIV velocity vectors were obtained from a frame to frame cross-correlation technique involving successive frames of patterns of particle images in an inter rogation window of 32 × 32 pixels. An effective overlap of 50% of the interrogation windows was employe d in PIV image processing. The time-averaged velocity ( V U , ) distributions were obtained from a cinema sequence of 498 frames of measured instantaneous velocity fields. The measurement uncertainty level for the velocity vectors was estimated to be within 1.0%. 2.3. Characteristics of Tornado-like Flow During the experiments, PIV measurements were first conducted i n the flow field without building model to quantify the characteristics of the tornado-like vor tex generated by the ISU tornado simulator. The cinema sequence of instantaneous PIV measurement results revealed clearly that tornado-like vortices are highly turbulent with vary ing vortex sizes and vortex center locations from one frame to another. Fig. 2.4 demonstrates a t ypical instantaneous flow field with streamlines in the lowest measured horizontal pla ne (Z=4mm). Images were taken from below the generated tornado, but all flow field images a re shown from above for

11

convenience. A single counter-clockwise vortex is apparently shown in the streamlines. Even though the streamline are not smooth spiral circles, the vortex cent er was found to be located at the origin center (X=0, Y=0). Vortex wandering is the seemingly random motion around the mean posit ion of the vortex center, which was firstly mentioned in Lund and Snow [11]. This feature causes difficulty in determining the vortex center. This is an important issue because the vortex center is the reference for computation of all the physical quantities and int erpretation of vortex structure. To some extent, vortex wandering may be an inherent feature of a vortex core, cl osely related to turbulence in the incoming flow and the flow separation. Church and Snow [ 6] modified their vortex chamber configuration and made efforts to minimize vor tex wandering, but they found it impossible to avoid completely. For point measurement techniques , such as LDV, big problems in setting up the measurement system and interpreting the obtained results occur due to vortex wandering. Fortunately, the vortex center can be tracked using PIV technique and the mean vortex center can be obtained by ensemble-ave raging sufficiently large amounts of data samples. Here vortex wandering was observed in the lowest horizontal plane as shown in Figure 2.5. The instantaneous vortex centers were sca ttering distributed without evident orientation priority. Almost all vortex centers recor ded from the 498 PIV images were randomly distributed in a circle with a radius of 2R 0 /3, where R 0 is the tornado core radius where the maximum tangential velocity is found. Since t he vortex center is always travelling in a certain range, the most appropriate a pproach to determine the location of the vortex center and corresponding flow fields is ensemble-avera ging the flow fields, provided enough samples are collected. The vortex center did not vary m uch when it was

12

compared for 300 PIV images and 498 images. Therefore 498 images can be said enough for the averaged measurement results.

Figure 2.4. Typical instantaneous streamline patterns in the lowest measured horizontal plane (Z=4mm)

Figure 2.5. Distribution of instantaneous vortex centers at the lowes t horizontal plane (Z=15mm) X ( m m ) Y ( m m ) -100 -50 0 50 100 -100 -50 0 50 100

13

Figure 2.6. The flow characteristics of the tornado-like vortex The time-averaged PIV measurement results are presented to cha racterize the time- averaged behavior of the tornado-like vortex. Figure 2.6 shows the time -averaged PIV measurement results (i.e., velocity distributions and the correspondi ng streamlines) to reveal the 3-D flow structures of the tornado-like vortex. In the figure, axi symmetric flow patterns in the form of a well-defined single clockwise vortex can be see n clearly in the horizontal planes. The streamlines in the vertical plane passing the time -averaged vortex center reveal clearly that flow streams near the ground and far away from the vortex center would move towards the vortex core and turn upward abruptly before reaching the vortex center. It indicates a radial and upward vertical flow appearing in the region outside the vortex core, as expected. An interesting flow feature is seen in the vortex core region, whe re flow is found to be a downdraft jet impacting the ground. As the downdraft jet approaches the ground in the

Full document contains 156 pages