LIBRARY

MiChiQan State
University

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/Datoouopss-p. 15

AN EXPERIMENTAL STUDY OF THE

AERODYNAMIC SHROUD IN AXIAL VENTILATION FAN SYSTEMS
By

Douglas R. Neal

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

MASTER OF SCIENCE

Department of Mechanical Engineering

2002

ABSTRACT

AN EXPERIMENTAL STUDY OF THE
AERODYNAMIC SHROUD IN AXIAL VENTILATION FAN SYSTEMS

By

Douglas R. Neal

An experimental investigation was performed on a building ventilation fan with an installed aero-
dynamic shroud. The fan system used for this study was characterized by its high volume flow
rate, low pressure rise and downstream conical diffuser. A new definition of efficiency is pre-
sented that differs than the currently accepted industry standard. Arguments are presented to sup-
port this definition, and its use not only in this type of fan, but also other turbomachinery
applications. An energy balance is derived that includes this definition of efficiency and the spe-
cific loss mechanisms of turbomachinery. This formulation clearly shows how the efficiency can

be increased by reducing these loss mechanisms.

Integral measurements of pressure rise vs. flow rate and efficiency are presented for a number of
different configurations, blade designs, diffusers and aerodynamic shroud conditions. These data
show that the aerodynamic shroud is able to increase the flow rate by over 35% while simulta-
_ neously increasing the efficiency by over 15%. Furthermore, it is shown that the choice of blade
design is critical to these improvements and that a larger diffuser cone can be used. Particle Image
Velocimetry (PIV) measurements are presented to show how the aerodynamic shroud affects the
flow field downstream of the fan. These data Ashow that the aerodynamic shroud is able to reduce
specific loss mechanisms that were identified in the energy balance and also create a more uni-
form exit profile at the exit of the diffuser cone. Finally, a new fan system is proposed that incor-

porates the modifications that were made and proven successful in this study.

ACKNOWLEDGEMENTS

I would like to thank Dr. John Foss, my advisor, friend and much more of the past several
years. His guidance, persistence and dedication to his students is tremendous. The quality
of this manuscript is largely due to his careful review. Additionally, I would like to thank
the efforts of Dr. Ahmed Naguib, Dr. Charles Petty and Dr. Tom Shih for their thoughtful
comments during the review process. The engineering faculty at M SU create a solid aca-

demic research environment.

I would also like to thank my colleagues in the Turbulent Shear Flows Laboratory (TSFL):
Richard Prevost, Scott Treat, Kyle Bade and A1 Lawrenz. Without Al Lawrenz, I often
wonder how the laboratory would continue to function. A special thanks to Scott Morris,
who I shared an office in the TSFL with for several years. Scott provided hours of engag—
ing conversation that have really stimulated my interest in scientific research, and he
knows the value of strong black coffee in the success of research. Thanks for making thou-

sands of pots of coffee. I look forward to continued interactions down the road.

Finally, I want to thank members of my family, including my fellow MSU alumnus and
uncle, James T. Neal (MS. 1959). I will enjoy having a publication that shares a space in
the MSU library alongside his thesis. Lastly, my mother and father, who have always
encouraged me and taught me the valuable lesson that the most important things in life are
not bought and sold, but earned and maintained. They continue to support my academic

(and other life) endeavors.

iii

TABLE OF CONTENTS

 

 

 

 

 

 

List of Figures- - - ix
List of Tables - -- - xv
Nomenclature - - -- -- xvi
1.0 Introduction ...... 1
1.1 Problem Statement and Background Information .................................................. 1

1.2 The Aerodynamic Shroud Concept - Previous Work ............................................. 4

2.0 Fan Analysis - - - - 8
3.0 Equipment and Facility Calibrations- 17
3.1 The Axial Fan Research and Development (AFRD) Facility .............................. 17

3.1.1 Overview ................................................................................................. 17

3.1.2 Flow Rate Measurement System ............................................................. 20

3.2 Calibrations .......................................................................................................... 23

3.3

3.4

3.5

3.2.1 Calculation of the Discharge C oefficient for the

AFRD Calibration Nozzle ....................................................................... 23
3.2.1.1 3/16 Scale Model Study ........................................................................ 24
3.2.1.2 Hot-Wire Study ..................................................................................... 26
3.2.2 Calibration of the Flow Rate Measurement System .............................. 28
3.2.3 Shroud Flow Rate - Elbow Meter System .............................................. 32
3.2.4 Power Meter Calibration ......................................................................... 34
Data Acquisition and Processing System ............................................................. 36
Pressure Transducers ............................................................................................ 37
Optical Encoder .................................................................................................... 37

4.0

5.0

6.0

 

 

 

3.6 Uncertainty Considerations .................................................................................. 38

3.6.1 Repeatability ............................................................................................ 38

3.6.2 Accuracy .................................................................................................. 39

3.6.2.1 Accuracy of Flow Rate Measurements ................................................. 40

3.6.2.2 Accuracy of the Efficiency Measurements ........................................... 41

Experimental Configuration -- -- - -_ 42

4.1 Baseline Configuration: A Standard 0.61 m Production Assembly ...................... 42

4.2 Modified Configurations ...................................................................................... 43

4.2.1 Different Fan Blade Designs ................................................................... 43

4.2.2 Installed Centerbody Device ................................................................... 43

Experimental Techniques--.-- -- 49

5. l Integral Measurements ......................................................................................... 49

5.2 Particle Image Velocimetry (PIV) Measurements ............................................... 52

5.2.1 PIV Measurement Locations ................................................................... 55

Results and Discussion - Integral Measurements 57

6.1 Baseline Measurements - 4-Blade Fan with 7 Deg. Diffuser Cone ..................... 57
6.1.1 Baseline Measurements (Production Assembly) -

"Shroud Off" Condition .......................................................................... 57

6.1.2 Production Assembly - "Shroud On" Condition ..................................... 59

6.2 The 4-B1ade Fan with an 11 Deg. Diffuser Cone ................................................. 59

6.2.1 The 4-Blade Fan and 11 Deg. Diffuser - "Shroud Off" Condition ......... 61

6.2.2 The 4-Blade Fan and 11 Deg. Diffuser - "Shroud On" Condition .......... 61

6.3 The 4-Blade Fan and 11 Deg. Diffuser Cone at Variable Shroud Gaps .............. 62

6.3.1 The 4—Blade Fan and 11 Deg. Diffuser at a

7.0

Medium Shroud Gap (g = 5.0 mm) ......................................................... 62

6.3.2 The 4-Blade Fan and 11 Deg. Diffuser at a

 

Large Shroud Gap (g = 7.0 mm) ............................................................ 64

6.4 Comparison of the S-Blade and the 4-Blade Fans ................................................ 65
6.4.1 The 4-Blade and S-Blade F ans - "Shroud-Off" Condition ...................... 66

6.4.2 The 4-Blade and 5-Blade Fan - "Shroud-On" Condition ........................ 67

6.5 A New Blade Design - Comparisons of3. 5 and 8-Bladc Fans ........................... 69
6.5.1 Comparisons of 3, 5 and 8-Blade Fans - "Shroud Off" Condition .......... 69

6.5.2 Comparisons of 3, 5 and 8-Blade Fans - "Shroud On" Condition .......... 70

6.6 The S-Blade Fan with an Installed C enterbody Device ....................................... 71
6.6.1 The Cylindrical Centerbody .................................................................... 72

6.6.1.1 The Cylindrical Centerbody - "Shroud Off" Condition ........................ 72

6.6.1.2 The Cylindrical Centerbody - "Shroud On" Condition ......................... 73

Results and Discussion - Velocity Measurements - 91
7.1 PIV Measurements on the Production 4-Blade Fan Configuration ..................... 91
7.1.1 Near Hub Region (4-Blade Fan with 11 Degree Diffuser Cone) ............ 92

7.1.1.1 Near Hub Region - "Shroud Off“ Condition ......................................... 92

7.1.1.2 Near Hub Region - Low-Pressure Shroud Condition
(Apshroud = 498 Pa) ............................................................................. 93
7.1.1.3 Near Hub Region - High-Pressure Shroud Condition
(APshroud = 996 Pa) ............................................................................ 93
7.1.2 Diffuser Wall Interior Region
(4-Blade Fan with 11 Degree Diffuser Cone) ......................................... 94
7.1.3 Exit Plane ofthe Diffuser Cone (7 Degree Cone) ................................... 97

7.1.3.1 Center Region ofthe Diffuser Cone Exit Plane (7 Degree Cone) ........ 98

vi

7.1.3.2 Near Wall Region of the Diffuser Cone Exit Plane
(7 Degree Cone) ................................................................................. 100
7.1.3.3 Overlap Region of the Diffuser Cone Exit Plane ('7 Degree Cone) 101

7.1.3.4 Composite View of the Diffuser Cone Exit Plane (7 Degree Cone) ..102

7.1.4 Exit Plane of the Diffuser Cone for the 4-Blade Fan with
11 Degree Cone .................................................................................... 103
7.1.4.1 Center Region of the Diffuser Cone Exit Plane (11 Degree Cone) 103
7.1.4.2 Near Wall Region of the Diffuser Cone Exit Plane
(11 Degree Cone) ............................................................................... 105
7.1.4.3 Overlap Region of the Diffuser Cone Exit Plane (11 Degree Cone) .. 106
7.1.4.4 Composite View of the Diffuser Cone Exit Plane (11 Degree Cone) 107
7.2 PIV Measurements on the 5—Blade Fan Configuration ...................................... 107
7.2.1 Exit Plane of the Diffuser Cone (11 Degree Cone) ............................... 108
7.2.1.1 Center Region ofthe Diffuser Cone Exit Plane
(11 Degree Cone) ............................................................................... 108
7.2.1.2 Near Wall Region of the Diffuser Cone Exit Plane
( 11 Degree Cone) ............................................................................... 110
7.2.1.3 Overlap Region of the Diffuser Cone Exit Plane
(11 Degree Cone) ................................................................................ 1 11
7.2.1.4 Composite View of the Diffuser Cone Exit Plane
(11 Degree Cone) ............................................................................... 112
7.3 Exit Plane Velocity Profiles and Mass Flow Rate Distributions ........................ 113
7.3.1 Exit Plane Velocity Profiles: 4-Blade Fan with
7 Degree Diffuser Cone ........................................................................ 114
7.3.2 Exit Plane Velocity Profiles: 4-Blade Fan with

1 1 Degree Diffuser Cone ...................................................................... 1 l6

vii

7.3.3 Exit Plane Velocity Profiles: 5-Blade Fan with

 

11 Degree Diffuser Cone ...................................................................... 116

8.0 Conclusions, a Proposed New Fan System, and Future Work 150
8.1 Conclusions ........................................................................................................ 150

8.2 Proposed New Fan System ................................................................................. 153

8.3 Future Work ....................................................................................................... 154
References -- -- - -- - - -- 156

 

viii

 

LIST OF FIGURES

FIGURE 1.1 Greenhouse Ventilation Fan Systems .......................................................... 6
FIGURE 1.2 Close-up View of 3 Greenhouse Ventilation Fan ........................................ 7
FIGURE 1.3 Livestock Houses with Tunnel Ventilation ................................................. 7
FIGURE 1.4 The Aerodynamic Shroud ............................................................................ 7
FIGURE 2.1 Control Volume for a Typical Axial Fan ................................................... 10
FIGURE 3.1 AF RD Facility with Test Fan Assembly .................................................... 19
FIGURE 3.2 Schematic Representation of Flow Rate Metering Device ........................ 20
FIGURE 3.3 Details of the "Proving Ring" Force Transducer and Bridge Circuitry ..... 22
FIGURE 3.4 Schematic Representation of the AF RD Facility with

Calibration Nozzle ..................................................................................... 23
FIGURE 3.5 3/ 16 Scale Model of the AF RD on the Sonic Nozzle Test Stand .............. 25
FIGURE 3.6 Discharge Coefficient Values for the 3/16 Scale Mode

l and Hot-wire Study on the F ull-Scale Calibration Nozzle ...................... 27
FIGURE 3.7 Close—up View Showing Agreement Between the Model Study

and Hot-Wire Study ................................................................................... 28
FIGURE 3.8 Calibration Data and Fit for the North Metering System .......................... 31
FIGURE 3.9 Calibration Data and Fit for the South Metering System .......................... 31
FIGURE 3.10 Data Fit for the Shroud Elbow Meter Calibration ..................................... 33
FIGURE 3.11 Calibration of the Elbow Flow Meter Using the Ford Haus Facility ......... 33
FIGURE 3.12 Overview of the Shroud Delivery with Elbow Flow Meter ....................... 34
FIGURE 3.13 Schematic of the Current and Voltage Circuit Used

to Measure Fan Power ............................................................................... 36
FIGURE 3.14 Circuit Schematic for AFRD Optical Encoder .......................................... 38
FIGURE 4.1 Schematic View of the Production 0.61m Ventilation Fan and

Diffuser Cone ............................................................................................. 44
FIGURE 4.2 Schematic of Standard Production Assembly ............................................ 44
FIGURE 4.3 Production 4-Blade Propeller ..................................................................... 45
FIGURE 4.4 Prototype 5-Blade Propeller ....................................................................... 45
FIGURE 4.5 Prototype 8-B1ade Propeller ....................................................................... 46
FIGURE 4.6 Prototype 3-Blade Propeller ....................................................................... 46
FIGURE 4.7 Schematic View of the Production 0.61m Ventilation Fan and Diffuser

Cone with Installed Centerbody ................................................................. 47
FIGURE 4.8 Close-up View of Centerbody and 5-Blade Hub ....................................... 48
FIGURE 5.1 PIV Interrogation Region ........................................................................... 53
FIGURE 5.2 Location of PIV Measurement Regions ..................................................... 56
FIGURE 6.1 Baseline (Shroud Off) Performance Measurements for the Production 4—

Blade Fan with 7 Degree Diffuser. The Multiple Tests Show the
Repeatability of the Measurements ............................................................ 76

FIGURE 6.2 Baseline (Shroud Off) Efficiency Measurements for the Production 4-
Blade Fan with 7 Degree Diffuser. The Multiple Tests Show the
Repeatability of the Measurements ............................................................ 76

FIGURE 6.3 Performance Data for the 4-Blade Fan with 7 Degree Diffuser at Various
Shroud Pressures ........................................................................................ 77

FIGURE 6.4 Efficiency Data for Data for the 4-Blade Fan with 7 Degree Diffuser at
Various Shroud Pressures .......................................................................... 77

FIGURE 6.5 Geometric Variables for the Diffuser Cone (as described in Table 4) ...... 78
FIGURE 6.6 Performance Curves for the 4—Blade Fan for 7 vs. 11 Degree Diffuser

(Shroud Off) ............................................................................................... 78
FIGURE 6.7 Efficiency Curves for the 4-Blade Fan for 7 vs. 11 Degree Diffuser

(Shroud Off) ............................................................................................... 79
FIGURE 6.8 Performance Data for the 4-Blade Fan with 11 Degree Diffuser - Low

Shroud Pressure (APshmud= 249 Pa) .......................................................... 79
FIGURE 6.9 Efficiency Data for the 4-Blade Fan with 11 Degree Diffuser - Low

Shroud Pressure (APshrouf 249 Pa) .......................................................... 80
FIGURE 6.10 Performance Data for the 4-Blade Fan with 11 Degree Diffuser at

a High Shroud Pressure (APshmud= 498 Pa) .............................................. 80
FIGURE 6.11 Efficiency Data for the 4-Blade Fan with 11 Degree Diffuser -

High Shroud Pressure (APShrouf 498 Pa) ................................................. 81
FIGURE 6.12 Performance Data for the 4-Blade Fan with 11 Degree Diffuser -

Medium Gap Height (g z 5.0 mm) ............................................................ 81
FIGURE 6.13 Efficiency Data for the 4-Blade Fan with 11 Degree Diffuser -

Medium Gap Height (g z 5.0 mm) ............................................................ 82
FIGURE 6.14 Performance Data for the 4-Blade Fan with 11 Degree Diffuser -

Large Gap Height (g z 7.0 mm) ................................................................. 82
FIGURE 6.15 Efficiency Data for the 4-B1ade Fan with 11 Degree Diffuser -

Large Gap Height (g z 7.0 mm) ............................................................... 83
FIGURE 6.16 Performance Data for the 4-B1ade and 5-Blade Fans ................................ 83
FIGURE 6.17 Efficiency Data for the 4-Blade and 5-Blade Fans .................................... 84
FIGURE 6.18 Performance Curves for 5-Blade Fan "Shroud On" Condition vs.

Baseline Fan (4-Blade) .............................................................................. 84
FIGURE 6.19 Efficiency Curves for 5-Blade Fan "Shroud On" Condition vs. Baseline

Fan (4-Blade) ............................................................................................. 85

FIGURE 6.20 Performance Curves for 4-Blade fan ("Shroud Off") vs. 5—B1ade

("Shroud Off“) vs. 3-B1ade ("Shroud Off") vs. 8-Blade ("Shroud Off") ...85
FIGURE 6.21 Efficiency Curves for 4-Blade fan ("Shroud Off") vs. 5-Blade

("Shroud Off") vs. 3-Blade ("Shroud Off") vs. 8-Blade ("Shroud Off") ...86
FIGURE 6.22 Performance Curves for 4—B1ade fan ("Shroud off") vs. 5-Blade

("Shroud On") vs. 3-B1ade ("Shroud On") vs. 8-B1ade ("Shroud On") ..... 86

 

h

 

FIGURE 6.23 Efficiency Curves for 4-Blade fan ("Shroud off") vs. 5-Blade ("Shroud
On") vs. 3-Blade ("Shroud On") vs. 8-Blade ("Shroud On") ..................... 87

FIGURE 6.24 Comparison of Performance for 5-Blade with and without Installed
Centerbody at the "Shroud Off" Condition. The Baseline

Production Case (4-Blade, "Shroud Off") is shown as a reference ........... 87

FIGURE 6.25 Comparison of Efficiency for 5-Blade with and without Installed
Centerbody at the "Shroud Off" Condition. The Baseline

Production Case (4-Blade, "Shroud Off") is shown as a reference ........... 88

FIGURE 6.26 Comparison of Performance for S-Blade with and without Installed
Centerbody at the "Shroud On" Condition. The Baseline

Production Case (4-Blade, "Shroud Off") is shown as a reference ........... 88

FIGURE 6.27 Comparison of Efficiency for 5-Blade with and without Installed
Centerbody at the "Shroud On" Condition. The Baseline

Production Case (4-Blade, "Shroud Off") is shown as a reference ........... 89
FIGURE 6.28 Performance Data for the S-Blade Fan with Installed Centerbody at

Various Shroud Pressures .......................................................................... 89
FIGURE 6.29 Efficiency Data for the 5-Blade Fan with Installed Centerbody at

Various Shroud Pressures .......................................................................... 90

FIGURE 7.1 Schematic Defining the "Near Hub Region" PIV Measurement Plane ...120
FIGURE 7.2 The 4-Blade Fan with 11 Deg. Cone - "Shroud Off" Condition -

Near Hub Region ..................................................................................... 120
FIGURE 7.3 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition

(APshroud: 498 Pa) Near Hub Region ...................................................... 121
FIGURE 7.4 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition

(APshrouf 996 Pa) Near Hub Region ...................................................... 121
FIGURE 7.5 Setup for Interior Wall PIV Measurements ............................................. 122
FIGURE 7.6 The 4-Blade Fan with 11 Degree Cone — "Shroud Off" Condition,

Cone Wall Interior Region ....................................................................... 122
FIGURE 7.7 The 4-Blade Fan with 11 Degree Cone - "Shroud On" Condition

(APshrouf 498 Pa), Cone Wall Interior Region ...................................... 123
FIGURE 7.8 The 4-B1ade Fan with 11 Degree Cone - "Shroud On" Condition

(APShrouf 996 Pa), Cone Wall Interior Region ...................................... 123
FIGURE 7.9 Setup for Exit Plane PIV Measurements ................................................. 124
FIGURE 7.10 Schematic Defining the "Center Region" for the Exit Plane PIV ............ 124
FIGURE 7.11 The 4-Blade Fan with 7 Deg. Cone — "Shroud Off" Condition

Center Region .......................................................................................... 125
FIGURE 7.12 The 4-Blade Fan with 7 Deg. Cone - ”Shroud On" Condition

(APShmud: 498 Pa) Center Region ........................................................... 125
FIGURE 7.13 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition

(APshmud= 996 Pa) Center Region ........................................................... 126

xi

FIGURE 7.14 Schematic Defining the "Near Wall" for the Exit Plane PIV .................. 126
FIGURE 7.15 The 4-Blade Fan with 7 Deg. Cone - "Shroud Off" Condition

Near Wall Region .................................................................................... 127
FIGURE 7.16 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition

(APshmud= 498 Pa) Near Wall Region ..................................................... 127
FIGURE 7.17 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition

(APshroud= 996 Pa) Near Wall Region .................................................... 128
FIGURE 7.18 Schematic Defining the "Overlap" for the Exit Plane PIV ...................... 128
FIGURE 7.19 The 4-Blade Fan with 7 Deg. Cone - "Shroud Off" Condition

Overlap Region ....................................................................................... 129
FIGURE 7.20 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition

(APshmud= 498 Pa) Overlap Region ......................................................... 129
FIGURE 7.21 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition

(APshroud: 996 Pa) Overlap Region ......................................................... 130
FIGURE 7.22 Composite View of the 4-Blade Fan with 7 Deg. Cone - "Shroud Off"

Condition .................................................................................................. 130
FIGURE 7.23 Composite View of the 4-Blade Fan with 7 Deg. Cone - "Shroud On"

Condition (APshmud= 498 Pa) .................................................................. 131
FIGURE 7.24 Composite View of the 4-Blade Fan with 7 Deg. Cone - "Shroud On"

Condition (APshmud= 996 Pa) .................................................................. 131
FIGURE 7.25 The 4-Blade Fan with 11 Deg. Cone - "Shroud Off" Condition Center

Region ...................................................................................................... 132
FIGURE 7.26 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition

(APshroudz 498 Pa) Center Region ........................................................... 132
FIGURE 7.27 The 4-B1ade Fan with 11 Deg. Cone - "Shroud On" Condition

(APshrouf 996 Pa) Center Region .......................................................... 133
FIGURE 7.28 The 4-Blade Fan with 11 Deg. Cone - "Shroud Off" Condition

Near Wall Region .................................................................................... 133
FIGURE 7.29 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition

(APshmud= 498 Pa) Near Wall Region ..................................................... 134
FIGURE 7.30 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition

(APshroudz 996 Pa) Near Wall Region .................................................... 134
FIGURE 7.31 The 4-B1ade Fan with 11 Deg. Cone - "Shroud Off" Condition

Overlap Region ........................................................................................ 135
FIGURE 7.32 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition

(APshmud= 498 Pa) Overlap Region ......................................................... 135
FIGURE 7.33 The 4-B1ade Fan with 11 Deg. Cone - "Shroud On" Condition

(APshroud: 996 Pa) Overlap Region ......................................................... 136

FIGURE 7.34 Composite View of the 4-B1ade Fan with 11 Deg. Cone -

xii

"Shroud Off" Condition .......................................................................... 136
FIGURE 7.35 Composite View of the 4-Blade Fan with 11 Deg. Cone -

"Shroud On" Condition (APshmud= 498 Pa) ............................................ 137
FIGURE 7.36 Composite View of the 4-Blade Fan with 11 Deg. Cone -

"Shroud On" Condition (APshroud: 996 Pa) ............................................ 137
FIGURE 7.37 The 5-Blade Fan with 11 Degree Cone - "Shroud Off" Condition,

AP = 25 Pa - Center Region ..................................................................... 138
FIGURE 7.38 The 5—Blade Fan with 11 Degree Cone - "Shroud Off" Condition,

AP = 50 Pa - Center Region ..................................................................... 138
FIGURE 7.39 The 5-Blade Fan with 11 Degree Cone - "Shroud On" Condition

(APshroud: 373 Pa), AP = 25 Pa - Center Region ..................................... 139
FIGURE 7.40 The 5-Blade Fan with 11 Degree Cone - "Shroud On" Condition

(APshmud= 373 Pa), AP = 50 Pa - Center Region ..................................... 139
FIGURE 7.41 The S-Blade Fan with 11 Degree Cone - "Shroud Off" Condition,

AP = 25 Pa - Near Wall Region ............................................................... 140
FIGURE 7.42 The 5-Blade Fan with 11 Degree Cone - "Shroud Off" Condition,

AP = 50 Pa - Near Wall Region ............................................................... 140
FIGURE 7.43 The 5-Blade Fan with 11 Degree Cone - "Shroud On" Condition

(APshmud= 373 Pa), AP = 25 Pa - Near Wall Region ............................... 141
FIGURE 7.44 The 5-Blade Fan with 11 Degree Cone - "Shroud On" Condition

(APshroud: 373 Pa), AP = 50 Pa - Near Wall Region ............................... 141
FIGURE 7.45 The 5-Blade Fan with 11 Degree Cone - "Shroud Off" Condition,

AP = 25 Pa - Overlap Region ................................................................... 142
FIGURE 7.46 The 5-Blade Fan with 11 Degree Cone - "Shroud Off" Condition,

AP = 50 Pa - Overlap Region ................................................................... 142
FIGURE 7.47 The 5—Blade Fan with 11 Degree Cone — "Shroud On" Condition

(APShmud= 373 Pa), AP = 25 Pa - Overlap Region ................................... 143
FIGURE 7.48 The 5-Blade Fan with 1 1 Degree Cone - "Shroud On" Condition

(APshroud= 373 Pa), AP = 50 Pa - Overlap Region ................................... 143
FIGURE 7.49 Composite View of the 4-Blade Fan with 11 Deg. Cone -

"Shroud Off" Condition and AP = 25 Pa ................................................. 144
FIGURE 7.50 Composite View of the 4-Blade Fan with 11 Deg. Cone -

"Shroud Off" Condition and AP = 50 Pa ................................................. 144
FIGURE 7.51 Composite View of the 4-Blade Fan with 11 Deg. Cone -

"Shroud On" Condition (APshmud= 373 Pa) and AP = 25 Pa ................... 145
FIGURE 7.52 Composite View of the 4-Blade Fan with 11 Deg. Cone -

"Shroud On" Condition (APshrouf 373 Pa) and AP = 50 Pa ................. 145

FIGURE 7.53 Exit Velocity Profiles: 4-Blade Fan with 7 Degree Cone, AP = 25 Pa....146
FIGURE 7.54 Mass Flow Rate Distribution: 4-B1ade Fan with 7 Degree Cone,
AP = 25 Pa ................................................................................................ 146

xiii

 

FIGURE 7.55 Exit Velocity Profiles: 4-Blade Fan with 11 Degree Cone, AP = 25 Pa..147
FIGURE 7.56 Mass Flow Rate Distribution: 4-Blade Fan with

11 Degree Cone, AP = 25 Pa .................................................................... 147
FIGURE 7.57 Exit Velocity Profiles: 5-Blade Fan with 11 Degree Cone, AP = 25 Pa..148
FIGURE 7.58 Mass Flow Rate Distribution: 5-Blade Fan with

11 Degree Cone, AP = 25 Pa .................................................................... 148
FIGURE 7.59 Exit Velocity Profiles: 5-Blade Fan with

11 Degree Cone, AP = 50 Pa .................................................................... 149
FIGURE 7.60 Mass Flow Rate Distributions: 5—Blade Fan with

11 Degree Cone, AP = 50 Pa .................................................................... 149
FIGURE 8.1 The Proposed New Fan Design ............................................................... 155

xiv

TABLE 1.
TABLE 2.

TABLE 3.
TABLE 4.

TABLE 4.

LIST OF TABLES

Flow Rate and Efficiency Data from Four Independent Tests .................. 39
Mean, Standard Deviation, and % Uncertainty from

the Data in Table 1. .................................................................................... 39
Comparison of Results Obtained in the AF RD vs. Published Data ........... 57
A Comparison of the Salient Features of the Tested Diffuser Cones

(see Figure 6.5 for a description of the variables: a, b, and L) .................. 61
A Comparison of the Salient Features of the Tested Diffuser Cones

(see Figure 6.5 for a description of the variables: a, b, and L) .................. 61

XV

NOMENCLATURE

Roman:

A Area

Cd Discharge coefficient

E DC Voltage (VDC)

Q Volumetric flow rate

Q Heat transfer rate

p Pressure

g) Power (Watts)

Re Reynolds number

tit Mass flow rate

V Velocity vector

V Velocity magnitude

U Spatially averaged velocity

g Gap height for aerodynamic shroud

i Alternating current (AC)

e AC voltage (VAC)

q Volume flow rate in cfm (ft3/min)

t Time

r Radial direction

x' Horizontal coordinate for PIV measurements in diffuser wall interior region

y' Vertical coordinate for PIV measurements in diffuser wall interior region
axial direction

Greek:

or Kinetic energy flux correction factor

11* System efficiency

1] Static efficiency

na Efficiency definition currently used in the agricultural ventilation fan industry

v Kinematic viscosity

xvi

p Density of air

Radial component

z Axial component

Subscripts:

A Auxiliary flow

I Instantaneous

N North side metering nozzle

S South side metering nozzle

ci Calibration inlet

i Inlet to control volume of Figure 2.1 on page 10.
e Exit to control volume of Figure 2.1 on page 10.
p Primary (fan flow)

xvii

1.0 Introduction

1.1 Problem Statement and Background Information

The exchange of fresh air and proper distribution of ventilating air within greenhouses
profoundly affects the health of the plants that are raised in these environments. Heat, dust
and other pollutants must be routinely removed from the air inside the greenhouse. Excess
heat is generated through a process of trapped thermal energy that is commonly known as
the greenhouse effect, see Hanan (1998). In this process, shortwave radiation in the form
of visible light is transmitted into the greenhouse through the walls, which are typically
made of glass, acrylic panels or plastic films. This sunlight will contact the interior sur-
faces, such as the foliage or internal structures and it heats these surfaces. Much of this
absorbed energy is in turn re-radiated as longwave infrared radiation. Since materials such
as glass, acrylic or plastic are opaque to longwave radiation, this energy remains trapped
inside the greenhouse. Solar energy can quickly elevate a closed greenhouse to tempera-
tures that are well in excess of the optimum for plant health. Most commercial green-
houses rely on ventilation fans to discharge this warm air and replace it with cool, outside
air. Additionally, proper circulation of air in commercial greenhouses is critical to prevent-
ing mold and mildew, which can destroy crops. Most mechanically ventilated greenhouses
use an exhaust, or negative pressure system (Wheeler and Both, 2002). The ventilation
fans are mounted on an exterior wall to exhaust air from the greenhouse, as shown in

Figures 1.1 and 1.2.

Fresh air ventilation and air circulation within livestock production buildings also affects
the health and growth rate of confined animals. This is especially true during periods of

high ambient temperatures when thermal stress resulting from the elevated temperatures

within a livestock enclosure negatively alters the rates of production, growth, conception
and survival for almost all livestock species (Mangold, et al., (1967); Fuquay, (1981);
Nienaber, et al., (1985) and (1987)). These effects are especially pronounced in the pro-
duction of swine and poultry where the animals are confined at relatively high density. In
the case of dairy cattle, declines in feed intake and milk production are common indicators
of heat stress, especially in high producing cows (Johnson, (1965); Yousef and Johnson,

(1966)).

Increased air velocity over the body surface of a dairy cow and the resultant increased con-
vective heat and mass transfer can offset some of the negative effects of hot weather
(Brody, et al., (1954). Traditionally this has been accomplished by maximizing the open
area in sidewalls in naturally-ventilated barns. This technique is widely recommended to
increase summer air exchange and to improve convective and evaporative cooling (Bick-
ert, (1988)). However, with the high levels of milk production possible today with
improved breeding, feeding, management and facilities, the cows’ need to dissipate meta-
bolic heat is substantially increased (Tyrrell, et a1. (1988); Tillotson and Bickert, (1994)).
Using fans to increase air velocity in the vicinity of the cow during hot weather has been

found to be an important asset (Chastain, (1991)).

It has been quantitatively shown that a form of "tunnel ventilation," in which an axial

velocity of order 5.5 km/hr (1.5 m/s) of evaporativel y cooled air passing over the chickens
in a poultry enclosure will significantly enhance their egg production (layers) and weight
gain (broilers). This cooling airflow is induced by placing a bank of exhaust fans at the end

of a (typically) 122 m long building, as shown in Figure 1.3; see Overhults and Gates.

(1993). The importance of ventilation in swine buildings has been documented by
Mangold, et a1. (1967) and Nienaber, et al. (1987). However, implementing a tunnel venti-
lation system will substantially increase the producer’s costs. Overhults and Gates (1997)
reported the following regarding the energy requirements for operating tunnel ventilated

broiler houses:

- "Electrical energy in this study was only about one-fourth of the total energy require-
ment but was about one-half of the total energy cost. Thus, the decision to tunnel
ventilate (or perhaps even to mechanically ventilate) the broiler house will notice-
ably increase the grower's cost."

Previously, Overhults and Gates (1993) noted the following:

- "Summer flocks required 2 times as much electricity as winter flocks, reflecting
increased use of the tunnel ventilation system for cooling."

- Note: Summer flocks refers to broilers raised to market age (41-49 days) during
summer months, winter flocks are raised to market age during winter months.

- Electricity use during winter months averaged 1 12 kWh/1000 birds. During summer
months the usage was 222 kWh/1000 birds.

The documented evidence, cited above, clearly states the importance of adequate ventila-
tion and air circulation in the operation of greenhouse and livestock buildings. This evi-
dence also indicates the high costs associated with operating such ventilation systems.
Developments to ventilation systems that would reduce the operating costs associated with
greenhouse and livestock facilities would be of great benefit to US. producers. However,
during the last decade, improvements to agricultural fans have been limited to passive
devices such as better inlet shrouds and the addition of exit flow diffusers. Although cur-
rent fans provide adequate airflow, they fail to achieve the maximum performance and
efficiency that can be achieved by active control techniques. Given the narrow profit mar-

gins in agriculture and significant cost of electricity to operate ventilation fans, it is appro-

priate to consider active techniques to improve efficiency and to obtain lower operating

costs.

1.2 The Aerodynamic Shroud Concept - Previous Work

The aerodynamic shroud, see Figure 1.4, was originally developed for automotive cooling
fans where the relative motion of the engine with respect to the shroud requires a large gap
between the fan tip and the outer shroud. This tip clearance region is characterized by
energy dissipation effects (i.e., losses) that are associated with the pressure-to-suction sur-
face flows of the propeller blade. These "tip losses" can be a dramatic detriment to the per-
formance and efficiency of an axial fan, and they are estimated to be 20-40% of the total

losses, as reported by Lakshminarayana ( 1996) and Bleier (1998) amongst others.

Morris (1997) conducted the initial study of the aerodynamic shroud application on the

engine cooling fan for a 1i ght-duty truck. This fan had a diameter of 457.0 mm, and the tip

clearance was 25.0 mm. Because of the high system resistanceI for this configuration,
which includes the grill, air conditioning condenser, radiator, and downstream blockage

elements such as the engine, the pressure rise of this fan was relatively high (250 - 1000

Pa). The flow rates were moderate, ranging between 0.1 m3/s and 0.9 m3/s. Morris found
that the aerodynamic shroud was able to enhance the performance at lower flow rates, but
that it degraded the performance at higher flow rates. Similarly, he showed that a large sys-
tem resistance would cause the aerodynamic shroud to lessen the efficiency; however, a
small system resistance would allow the aerodynamic shroud to improve the system effi-

ciency. Since the larger system resistance values involved a radial outflow from the fan, it

 

1. The operating point is the intersection of the performance curve with a parabola that defines the system
resistance.

was realized that a contoured, and not a sudden expansion outlet passage, could take
advantage of the aerodynamic shroud contribution. The performance benefits of these

downstream geometry adjustments are documented in [26].

Morris (1997) also conducted a brief study which showed that a fan with a larger tip clear-
ance, when used in conjunction with the aerodynamic shroud, could produce better perfor-
mance results than the same fan with a near zero tip clearance. Again, these results show
that both performance and efficiency are enhanced at higher flow rates, and degraded at
lower flow rates. These data indicate that fan applications where a small tip clearance is

allowed could benefit from a larger tip clearance with the aerodynamic shroud.

The latter observation of Morris (1997) created the primary motivation for the work pre-
sented here, since agricultural building ventilation fans have high volume flow rates (1 .0-
3.5 m3/s) and modest pressure rise (OD-75.0 Pa) conditions. This idea was initially investi-

gated by Neal and Foss (1999), and it is examined in further detail in the present thesis.

 

 

 

 

. 5‘ +5 .

FIGURE 1.2 Close-up View ofa Greenhouse Ventilation Fan

hhhhhhhhhh i

  

FIGURE 1.3 Livestock Houses with Tunnel Ventilation

Inlet Flow Direction ‘
V
Jet(iap(g) i
E JetThickness I , .
l Test Fan _ 7' 12hr ~
' .L "l" 7- ‘TipClearance .
3 Th T .4 Dhub “p I ‘ I .7 ‘ it“,
' ‘ ’ '. 7 ‘ 7; i‘
a) Elevation View b) Plan View (‘/4 Section)

FIGURE 1.4 The Aerodynamic Shroud

2.0 Fan Analysis

An analysis using the integral energy equation is presented in this section in order to pro-
vide a complete statement of the loss mechanisms in turbomachinery. Although numerous
discussions on loss mechanisms and efficiencies exist in the present literature, detailed
analyses using an energy balance are few. Lakshminarayana (1996) derived an integral
energy balance for a centrifugal compressor. His result, shown in equation 2.1 is valid for
viscous, compressible flows and includes all losses associated with viscous flow over the
blades.

1:473 = ho] — I102 (2.1)
This equation, while useful, does not provide insight into the specific loss mechanisms
associated with turbomachinery, nor does it account for a possible net kinetic energy flux.
Another formulation of the energy equation is thus proposed that will provide a better
understanding of these losses. The present derivation will begin with the generalized inte-

gral energy equation as proposed by Potter and Foss (1982):

2
Q— Ws— Wshear— ' _ “ll—216:]; v[v— + gz + u]pdV+J; (— + g; + u 4‘5ij MIA (2. 2)
Equation 2.2 describes the instantaneous balance of the terms in this control volume equa-
tion. If the equation were integrated over an integer number of cycles (T) of the fan and if
each term was then divided by T, the unsteady term would be zero. The remaining terms
would describe each quantity averaged over T cycles. The losses in the above equation
will be defined as the sum of all the terms representing unusable forms of energy and can

be written as:

losses = —Q+I {IpV-fsz (2.3)

CV

Applying equation 2.2 to the control volume shown in Figure 2.1, the following observa-
tions can be made. First, the control volume is stationary and there is no shear work
exerted on the surface of the control volume. Also, the velocity at the inlet to the control
volume can be assumed to be zero; this would represent the interior domain of a green-

house or livestock facility. Equation 2.2 then reduces to:

b : pe—pi+a<v

 

.>2
m p 2 + ZghL (2.4)

In equation 2.4, the term <Ve> is the spatially-averaged exit velocity. The or term in front

of this term accounts for the non-uniform velocity profile and allows use of this spatially
averaged velocity. The value of or is greater than or equal to unity, with the latter case
being for a uniform velocity distribution. The pressure term could also have a coefficient

if there were a correlation between the local velocity V(r,6) and p, such that:

lp‘v- :3 = v<V> <p> (2.5)

This formulation will assume that this situation does not occur and the pressure across the

inlet and outlet will be considered to be uniform. The W5 term represents the available

shaft work that is delivered to the rotor; it is related to the input power (the true energy

cost) by the motor efficiency:

1) = —' (2.6)

The losses described in equation 2.3 are now re-written into the product of gravity and
various head losses. These losses can be further defined by their contributing mechanisms.

Denoting the losses by their loss coefficient, k, as defined by

L.
2

ghL = k (2.7)

    

/
r
I
I
z
I
a
I
I
I
I

O .
1n e I

I

1

I

FIGURE 2.1 Control Volume for a Typical Axial Fan

allows for a convenient formulation by which the various losses can be defined with
respect to their contribution to the energy budget. A detailed description of these contrib—
uting losses is described in the following paragraphs. Only the losses that are relevant to
low-speed ventilation fans and their flow fields (i.e. incompressible, no shock waves, etc.)
will be considered here. More general information on turbomachinery losses can be found

in Lakshminarayana (1996) or Wallis ( 1983).

The losses that are relevant to the control volume shown in Figure 2.1 are:

1. Blade profile loss. These viscous losses are primarily caused by the development of
blade boundary layers across the entire span of the rotor and the corresponding
wakes behind the passing blade. These losses would also include separated flow on
the suction side of the blade, which occurs primarily at high pressure rise (low flow

rate) conditions.

2. Shroud and other passage loss. These losses result from the boundary layers that
form in the shroud and upstream housing regions, as well as the downstream diffuser
cone. Separated flow in the diffuser cone will provide an additional loss; therefore,
diffuser cones are typically limited to 7 degrees or less. This number is based on
empirical data by the manufacturers and is also confirmed in studies discussed by
Wallis (1983). A well designed fan housing will typically have very little separated
flow in the upstream and downstream region and flow separation is a small contribu-

tor to the overall losses.

3. Tip clearance losses. The spacing between the rotating blade section and the station-
ary outer shroud varies significantly from application to application, but this is a
source of significant loss in all turbomachinery applications. This type of loss results
from the need to have a finite distance between the outermost tip of the blade and the
surrounding shroud. The pressure (downstream) side of the blade will interact with
the suction side (upstream) of the blade and cause a leakage flow to occur in the axial
gap region. Since the leakage flow and main flow in the axial gap region are often at
different angles, the mixing of these two dissimilar flows creates further losses as a
result of the strong shearing in this region. Additionally, a "tip vortex" can also occur
in this region. This resulting vortex can create substantial losses by viscous dissipa-
tion of the associated turbulent motions, along with the losses associated with inter-
actions of the tip vortex with the main flow (e. g. entrainment of the main flow by this
tip vortex). The magnitude of these losses varies widely depending on the applica-
tion, but Lakshminarayana (1996) estimates that clearance losses account for 20-

35% of the total losses.

4. Secondary flow losses. Losses from secondary flows are the result of the mixing and

dissipation of energy in the fan system. The loss mechanisms are similar to those

11

described under tip clearance losses. These losses are often lumped together with the
shroud and other passage losses (annulus and hub-wall) boundary layer losses and
described as "endwall losses". They are described here separately since significant
secondary flow losses occur away from the shroud and passage losses and these

effects can be addressed.

Additional losses in a more general discussion would include those from the formation of
shock waves for gas flows and cavitation bubbles for liquid flows. These types of losses
(blade profile, shroud, tip clearance and secondary flow) will be referenced in the energy

equation analysis of 2.4 and 2.5 as kbl, ksh, ktC and km, respectively. Their usage in the

energy equation analysis requires that the associated velocities be identified. The blade

profile losses scale with the velocity of the fan blade, therefore the fan tip speed,

U n,. = 9212 (2.8)
is used. The losses in the tip clearance region, while distinctly different than those in from
the blade profile, also scale with the fan tip speed. The losses in fan shroud and other flow
passages would scale with the average velocity throughout the fan. This can be calculated
by taking the total flow rate through the fan and dividing it by a representative area in the

shroud or other internal flow passages. An approximate average area can be used for this

calculation and the resulting velocity is termed Vsh. The secondary flow losses, including
those in the hub-wall, would also scale with Vsh. This is supported by the fact that second-

ary flow losses are often categorized with shroud and passage losses. By incorporating

these "itemized" k values, the energy equation is written:

 

. 2 - 2 2
W3 :7 . —p- (V..> U . V_.-/, U - Vsh
_ : _‘_’ + a + k _”I’_ k __ + k ‘_“1’_. k _ 2
m p 2 bl 2 + sh 2 I( 2 + sec 2 ( 9)

12

The introduction of two well-established (Osborne ( 1966)) non-dimensional ratios can
provide useful information. The flow coefficient, (1), is a parameter that represents the

effect of the mass flow and the blade speed. It is written as:

V..
6 = 4" (2.10)
U

tip

The second non-dimensional number is the pressure rise coefficient, ulp, which establishes

the pressure rise capability of a particular fan geometry at a particular operating speed. It

is written as:

WP — T31; (2.11)
ipUlip

The energy balance can be written as:

. 2 2 2 2 2
W p<v> DU), pvsll PU,» stI
as = (pg — pl.) + 01—2" + kM—Lz’ + [cm—2 + k,(,—-’—2’ + km—2 ' (2.12)

Since 2.12 has the units of pressure, the entire expression can be made non-dimensional
by dividing each term by the product of the density and the square of the tip velocity:

2
V)

= Wp + 01[<—2—:|+ km + [fr/192 + kn- + kscc¢2 (2'13)

lip

W5
QGp U3). )
Another choice for non-dimensionalization could have been the pressure rise across the
fan, Ap = pc-pi. This is a rational choice for many turbomachinery applications; however,
for this specific application, there are complications with using this. This idea is further

explained by examining the conventional definition of static efficiency in turbomachinery:

.
'
I l'
C
.
.
..
,
-
.r‘
T-
..

. ‘ht barely.
‘04 :1
if: .3
«.4, _.

Hr U L.
'W
h Inn-L

i; Quill-3H3“ -7 '6 aiififi
'mswewimt

t
U" IL
5}
‘.

..
25."

Km

' 00%
.1 1.2”

no.1

 

115 = .Q_A£ (2.14)

60

This definition of efficiency describes the non-dimensional ratio of desired quantities,
flow rate and pressure rise, versus the input cost of obtaining these, which is the input
power. While this definition is commonly used in many turbomachinery applications, it is
not particularly instructive for other applications, such as the agricultural ventilation fan
industry. This is because the pressure rise values for these applications are very low and
the exit plane kinetic energy (that is not represented in 2.14) is a desired output since it is
related to the flow rate. It should be evident from equation 2.14 that if the pressure rise
were nominally zero, then it would correspond to an efficiency of zero. Clearly this is
problematic for fan systems with very low restrictions, or for fan systems with exceed-
ingly high restrictions, since it is either solely the flow rate or the pressure rise that is

required. The former situation is the case with the application studied here.

In general, the term "efficiency" is used to refer to the ratio of desired output over required
input, i.e. the benefits over the costs. Typically this is expressed in dimensionless form and
the efficiency expressed in this form will be between 0 and 1. Since the definition
described in 2.14 does not accurately describe the ratio of the desired output over the
required input for a ventilation fan, the agricultural fan industry has created their own def-

inition of efficiency which is described by:

Tia = 2 (2.15)

This ratio is not dimensionless, but instead has the units of volumetric flow rate over
power, which is commonly expressed as cfm/watt, or cubic feet per minute per watt, see

Ford et a1. (1999). In the interest of making this communication applicable to the intended

l4

community, the usefulness of this is noted; however this definition of efficiency will not

be used to present the results in this thesis.

Another definition of efficiency, which will be developed for discussion purposes here,
but not actively used as the definition in this thesis. is the total efficiency. This is denoted

as:

This definition attempts to additionally credit the fan with its ability to add kinetic energy
to the air, and as Bleier (1998) notes, represents the real fan efficiency. The difficulty with
this definition is that it is more difficult to calculate, since the total pressure requires that

the velocity contribution of the total pressure be known explicitly.

A definition of efficiency that incorporates the kinetic energy imparted by the blade, along
with the flow rate and the pressure rise capabilities of the fan is:
l U 2
Q ° 3p tip

11* = (2.17)
50

 

This equation better describes the relationship between the desired output of the fan and
the required input to the fan, and gives credit for the kinetic energy imparted by the blade.

Incorporating this definition into equation 2.13 yields the following equation:
11* = ' (2.18)
2 Ash 2
‘11,, + q) [ae(A—) + ks/r + k.\‘(’(‘] + kbl + klr‘

(’

 

Taking the definition for the system load curve, which is a parabola, as:

15

2

WP : A’syslemq) (219)

In this definition, Asyslem is dimensionless but it represents the sum total of all the restric-
tions through the fan system. For an agricultural ventilation fan this could include
upstream shutters, motor supports and downstream safety guards. These individual losses
would be added in either a series or a parallel arrangement, as described by Potter and

Foss ( 1982), amongst others. A fan system, where lcwslcm = 0 represents a fan operating in

open air, between identical upstream and downstream pressures, such as an internal circu-
lator fan within a greenhouse. This is commonly referred to as "free delivery". These fans
have high volume flow rates, but they generate essentially no pressure rise. A fan system,

where Asystem = 00 would be indicative of a fan operating at what is commonly known as

"shut-off" pressure, or more specifically, a fan that is pressurizing the air in an enclosed

box.

Using the definition for the system load curve, equation 2.18 becomes:

. = 1 (2.20)

2 Ash 2
(i) [Uzi—f) + ksh + ksec‘ + Asysrmn] + kbl + kl(‘

e

 

The objective of any attempt at fan optimization can now be focused on the task of mini-
mizing the denominator of equation 2.20. All the losses associated with the fan system,
including those encountered by adding necessary safety mechanisms are included. This
communication will explain how the aerodynamic shroud, when incorporated into a well-

designed system, can successfully obtain a minimal denominator in equation 2.20.

3.0 Equipment and Facility Calibrations
3.1 The Axial Fan Research and Development (AF RD) Facility

The AFRD was originally designed and built through the primary efforts of SC. Morris in
1994 for the purpose of studying automotive cooling fans. At this time, the facility was

referred to as the Automotive Cooling Fan Research and Development (ACFRD) facility.
The details of this original design can be found in Morris et a1. (1997). When further via-
ble applications were found by Neal and Bardouwalla (1998), the facility was renamed the

AFRD, accounting for the wider range of axial fans that can be tested.

3.1.1 Overview

The AFRD facility, shown in Figure 3.1, was used to acquire data for all of the fan tests.
The fan assembly, including diffuser cone, was placed on the top surface of the facility
(A). The shaft (L), which is used for driving automotive fans, was used as a mounting base
for the centerbody. This feature will be discussed in detail later. The pressure rise across
the fan was monitored by a pressure transducer (0 - 134 Pa) that measured the differential
pressure from the atmosphere to the upper receiver using a pressure tap located at (B). Air
is moved from the lab into the upper receiver and then through a set of nozzles (E), where
it enters the lower receiver (H) and the flow metering devices (F & G). The prime mover
(K, Chicago Blower Co. S Q-36.5) draws air from the lower receiver and exhausts it back
into the laboratory. The flow rate through the system can be controlled through a throttle

plate (J) that is located at the exit of the prime mover.

The AFRD facility was used to acquire integral performance (Ap vs. q) measurements and
efficiency (11 vs. Ap) of the fan and cone assembly. The AFRD is able to measure mass

flow rates induced by the fan by using a set of metering nozzles that are shown in the sche-

l7

matic representation of the AFRD facility (see Figure 3.1) and then in greater detail in
Figure 3.2. These metering nozzles are part of a unique flow measurement system that
uses a moment—of-momentum flux principle. As the airflow through the facility, the net
moment-of—momentum flux increases and the resistive force that is required to support the
turning vane is increased. As shown in Figure 3.2. the output is the response of the strain
gage assemblies that are attached to the mechanical supports for the turning vanes. There
are two nominally identical measurement systems that operate in parallel in the AFRD.
This design allows for a symmetric delivery of air from the upper receiver to the lower
receiver. Morris, et al. (2001) have shown this to be an effective technique for measuring

the total flow rate through the AFRD.

Since the AFRD utilizes this unique moment-of-momentum flux device, it does not
require long settling chambers to properly set the upstream boundary conditions for the
flow measuring device, as are typically needed in flow measuring nozzles. This allows the
AFRD to be vertically positioned into a relatively small area, thus not requiring the large
floor space usually required by similar fan testing facilities. It should be noted that an
additional strength of this facility (and similarly the flow measurement system) is the
capability to measure a wide range of fans, from very small diameter fans that operate at
very high pressure rises to large diameter fans that operate at a very small pressure rise
condition. This flexibility requires that the moment-of—momentum flux device can be
accurately calibrated against a reliable flow rate. This calibration is described in detail in

Sections 3.2.1 and 3.2.2.

The AFRD facility also has the unique capability to allow for the collection detailed mea-
surements downstream of a fan though the use of the traverse system, as shown by (C) in
Figure 3.1. This traverse can be maneuvered in each of three dimensions, r, 0, 2, which
define a polar coordinate system. These directions correspond to the radial, tangential and
axial components of velocity that are typically used in fan literature. The traverse was used

in this study to make the hot-wire measurements for the Cd study that are described in

Section 3.2.1.

y

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E
G
F
H
A TEST FAN ASSEMBLY G FORCE TRANSDUCER
B PRESSURE TAP H LOWER RECEIVER
C TRAVERSE I INLET TO PRIIVIE MOVER
D UPPER RECEIVER J THROTTLE PLATE
E NOZZLE K PRIME MOVER
F TURNING VANE L SHAFT

FIGURE 3.1 AFRD Facility with Test Fan Assembly

19

3.1.2 Flow Rate Measurement System

An integral measurement technique was implemented using a large turning passage to
redirect the air through a 90° angle. Figure 3.2 shows a detailed view of one metering
device. Details of the apparatus noted in the figure will be identified parenthetically by a

capital letter. The mass air flow through the system creates a net moment-of—momentum

flux about the pivot support (R).

 

   
 

 

 

 

 

K
/ s
/ K Inlet Filter
\ L Elliptical Nozzle
L M Turning Vane
N Counter Balance
R \ Control Surface N P Force Transducer
‘ / R Kmfe Edge P1vot
‘\ P S Upper Receiver Floor
\ T Floor of Lower Receiver
w

 

 

 

 

Es

FIGURE 3.2 Schematic Representation of Flow Rate Metering Device
This was balanced by the moment from the force transducer (P). Taking the sum of the
moments about (R) (counter-clockwise is positive):

ZMR = (r,-><f,.>+ lpirx‘v‘W-mdA = 0 (3.1)
CS

where the control surface (CS) spans the nozzle exit to the exit of the turning passage (see

Figure 3.2). The vector ; represents the radius from the pivot point (R) to the location on

the control surface. The derivation of equation 3.1 can be found in Potter and Foss (1982)

20

and Munson et. al (1996). The inflow to the control surface (CS) contributes to the

. . -‘ .x s A . .
counter—clockwrse moment—of—momentum flux (s1nce V- n < 0 and r X V 18 clockwrse).

The exit flow also contributes to the net counter-clockwise moment-of—momentum flux

. A .3 s -\ . .
(s1nce V- n > 0 and r x V 18 counter clockwrse).

The integral can be made dimensionless using a characteristic length (L) and velocity (U):

x A 3 2 ; Y; i7 A (IA
(r,_.><FP) = -pL U ”EXT/MU )P (3.2)
CS

The density (p) is considered to be constant: this is compatible with the low speed (Mach
no.<0.2) flow through the entire system. The integral can be replaced by a constant (c,) if
the dimensionless integral of equation 3.2 is not a function of the flow rate through the

system (i.e., the Reynolds number) nor the upstream velocity profile. The magnitude of
. . . 2
the force measurement can be written in terms of the mass flux(m ~ pL U) or

n't = cszzU where c2 depends upon the selection of U and L. Also, the quantity

(;F x FF) can be expressed as (£‘Fp) with the result that

 

C] 7 a
F = 11'1" = km' 3.3
P [pchf] ( )

In equation 3.3, the coefficient k has dimensions (length/mass) and m represents the

desired mass air flow rate. Once calibrated. the m value can be obtained given the mea-

sured force Fp and the coefficient ‘k’. An evaluation of k will be discussed in

Section 3.2.2.

21

As noted, the relationship in equation 3.3 is dependent on the insensitivity of the coeffi-
cient ‘k’ to changes to the velocity profile. This specific design has several attributes
which cause this condition to be met. First, the acceleration of the fluid due to the stream-
line curvature from the inlet of the control surface to the exit caused a greater momentum
flux at the exit. The accelerated exit profile is relatively insensitive to the upstream condi-
tions. This, along with the larger radius of the exit flow to the pivot point, makes the exit

flow the dominant contribution to the integral of equation 3.2.

The force, shown above in equation 3.3, is measured as the output response of the strain
gage assemblies that are attached to the mechanical support for the turning vanes. These
strain gage assemblies are shown in greater detail in Figure 3.3. Further details on the per-
formance and accuracy of this technique can be found in Morris, Neal, Foss, and Cloud

(2001).

'12 Volts

 

E(output)

 

Force Load

FIGURE 3.3 Details of the "Proving Ring" Force Transducer and Bridge Circuitry

22

3.2 Calibrations

3.2.1 Calculation of the Discharge Coefficient for the AFRD Calibration Nozzle
The metering nozzles were calibrated using an elliptical shaped contraction with a 3/2 axis
ratio. This inlet nozzle provided a 6.3/1 contraction ratio in flow area. The discharge coef-

ficient, Cd, for this device was determined using two separate methods. This section

describes the procedure by which the calibration inlet, itself, was calibrated.

  
  

Calibration
Nozzle

 

Outflow Upper
Receiver

 

 

 

 

 

Lower
Receiver

 

 

 

 

 

 

 

 

FIGURE 3.4 Schematic Representation of the AFRD Facility with Calibration Nozzle

Given that the mass flow rate through the calibration nozzle mm- is equal to the sum of the

flow rates through the two metering nozzles: m and 1h 5’ a knowledge of "'16,- per-

m m
N

mits the calibration of the metering nozzles. The mass flow rate through the calibration

inlet can be described as

23

 

m = pAm-CdA/gmmm — P,,(,(.(,,-,,(,,.) (3.4)
Each of the parameters on the right-hand side of equation 3.4 are easily measured except
the discharge coefficient, Cd. Since accurate assessment of Cd is important for accurate
measurements of flow rate, two methods were used to determine the Cd value for the pla-

nar contraction. These are described in the following sections.

3.2.1.1 3/16 Scale Model Study

One method for determining the Cd value was to build a 3/16 scale model of the AFRD

facility. The salient features of the AFRD were utilized in the model; namely, the calibra-
tion inlet, the outer surface of the AFRD and the interior of the upper receiver were accu-

rately recreated as 3/16 scale. Since the Cd value is only a function of the boundary

conditions and the Reynolds number:

U..l
Re. = “v‘” , (3.5)

 

the prototype Cd value can be determined for the Reynolds number values of the model

study.

A Testek model 1 1359 sonic nozzle test stand was used to generate a known mass flow
rate for the model study. The sonic test stand provides a known mass flow rate that ranges
from 3.5 kg/hr to 800 kg/hr. The desired flow rate is established by opening and closing an
array of nine nozzles. The nozzle areas define a binary progression, as described in

equation 3.6.

3.5kg/hr

at STP (3.6)
<V>

II
A" — 2402 where/10 —

24

The open/closed condition for each nozzle is established by a solenoid activated valve.
The chamber directly downstream of the nozzles is evacuated to a pressure sufficiently
low enough to impose a sonic (M=1) at the throat of the open nozzles. The 3/16 AFRD
model was placed onto the sonic nozzle test stand such that test stand removed air from the
lower receiver of the model. Therefore, the sonic nozzle test stand acted as the prime

mover for the 3/16 AFRD model. This configuration is shown in Figure 3.5.

OUTLET

3/16 scm: AFRD
CALIBRATION UNIT

SOLENOID
VALVES

 

NOZZLE
PLATE

FIGURE 3.5 3/16 Scale Model of the AFRD on the Sonic Nozzle Test Stand

The discharge coefficient was calculated from the known mass flow rate of the model
study and the other terms on the right hand side of equation 3.4. A plot of the calculated

Cd values vs. Reynolds number is shown in Figure 3.6. This figure shows the Cd value

25

increasing monotonically with Reynolds number before levelling out at a value near 0.980.
The high values, which are quite near the ideal value of 1.0, can be attributed to the well-
designed contraction. This contraction minimizes frictional losses by thinning the

approach boundary layers through flow acceleration.

3.2.1.2 Hot-Wire Study

After completing the study with the 3/16 scale model of the AFRD, an additional study
was conducted on the full-scale contraction. These additional measurements confirmed the
measurements made with the 3/ 16 scale model and provided further information for accu-

rate Cd values at higher Reynolds numbers. For this study, hot-wire measurements were

made in the near wall region of the exit of the planar contraction for both the length (L)

and width (w) sections. These data provided estimates of the displacement thickness, 5

associated with each wall’s boundary layer. The discharge coefficient can then be esti-

mated as:
Cd-wL = wL—28LL—25ivw (3.7)
01'
C _ 1 25‘}, 25:. 38
(1 - - W ‘T ( . )

Figure 3.6 shows a composite plot of the model study and the hot-wire study. The range of
flow rates that were needed for this study correspond to Reynolds numbers that range from
~150,000 to ~550,000. Figure 3.7 shows a close-up view of the overlapping region
between the 3/16 model study and the hot-wire study. The agreement between these two

regions further reinforces the accuracy of the Cd values. particularly since different

26

approaches were used in each study. From these data, a Cd value of 0.980 was confirmed

for the present study, which required flow rate calibrations that ranged from 1.50 - 4.75
kg/s. This flow rate requirement corresponded to Reynolds numbers that ranged from
~150,000 to ~550,000 for the calibration inlet. Over this range of Reynolds numbers, the

Cd values varied no more than 0.5%, as shown in Figure 3.6. Therefore, a constant of

value of 0.980 was used in equation 3.4.

AFRD Discharge Coefficient Calibration

 

 

 

 

 

 

 

 

 

1.00 ;—_ _____ __ 4 _
/ '
Legend
0.80 - 0.980
I ..__. Cd (3/16 Model Study)
C ..
d A... CD (Hot-wire Study)
060 l—
0.40 —
l-
0.20 —
000 I I 1 i 1 1 l l l l 1 l 1 I 1 i 1 l 1 l l 1 l i 1 1 1 l l I I
0 100000 200000 300000 400000 500000 600000 700000 800000

Red"

FIGURE 3.6 Discharge Coefficient Values for the 3/16 Scale Model and Hot-wire Study
on the Full-Scale Calibration Nozzle

27

AFRD Discharge Coefficient Calibration (Close-up View)

 

1.10

 

Legend

1.05 ~—- . n Cd (3116 Model Study)

, . ._ (THINK-wire Study)

 

0.95 ~—

0.90 r—
085 r—

080 l—

 

 

 

0.75111l111lllilim1llllllmil

75000 95000 1 15000 135000 155000 175000 195000

Re (In

FIGURE 3.7 Close-up View Showing Agreement Between the Model Study and Hot-Wire
Study

3.2.2 Calibration of the Flow Rate Measurement System

The metering nozzle calibration was accomplished by the addition of a planar contraction
on the top of the AFRD normally occupied by the test fan. The auxiliary fan, shown in
Figure 3.4, established an air flow that moved air from the laboratory, through this calibra-
tion inlet and into the upper receiver - plenum. From the upper receiver, the air moves
through the metering nozzles and into a lower receiver. The air is then returned to the lab-
oratory through the centrifugal fan. A throttle plate at the exit of the auxiliary fan was used
to control the flow rates. A range of known flow rates was generated (Section 3.1) and the

resulting voltage outputs from the strain gages were recorded for the calibration process.

The rectangular opening (outlet dimensions were 0.75 m x 0.25 m) made it possible to

directly test the insensitivity of the metering nozzle to the upstream flow distribution. This

28

was specifically tested by rotating the nozzle 90 degrees around its center axis (from the
position shown in Figure 3.4). No systematic change in the k value was observed for these

distinctly different inflow conditions to the metering nozzles.

Equation (3.3) can be written explicitly in terms of the mass flow rate give the linear rela-

tionship of the measured voltage (E) with the applied force:

m = k"../E— b (3.9)

The coefficients b and k" were empirically determined using a linear ordinary least
squares fit of the calibration data. The resulting transfer function for the two flow meters

WCI‘CZ

rizN 3.5348 EN—0.490 (kg/s) (3.10)

and

11°15 4.3506 Es — 0.507 (kg/s) (3.11)
where ES and EN represent the time mean voltages measured from the transducers on the
north and south sides of the facility, respectively. These curves show a linear fit between
the square root of the measured strain gage voltage and the resulting flow rate. These data
confirm the insensitivity of the net moment-of-momentum flux to the inflow condition
since calibration data for each of the two orientations were combined to generate the final
curve fit. This result was expected since the inlet screens/filters distributed the flow and

since an integral effect was inherent in the measurement. The calibration data and the

curve fit of these data are shown in Figures 3.8 and 3.9.

29

The final values for total mass flow rate were determined by taking the average of the two
measured values from the two metering nozzles. Total mass flow rate (m) was then
divided by the representative density, yielding a volume flow rate that is reported in the
performance curves of Section 6.0 and in the definition of n*, as described in

equation 2.17. This practice is also consistent with the current standards of the ventilation
fan industry. Using these devices, flow rates that range from 1.20 — 4.50 m3/s can be reli-
ably measured. This range of flow rates is well-suited for 24 inch livestock ventilation
fans, which typically have flow rates between 1.20 - 3.10 m3/s. as detailed in Ford et a1.

(1999). This limited range allows for good resolution in the flow rate measurements.

The quality of the data fit was determined using the standard deviation of the difference

between the calibration data and equations 3.10 and 3.1 1:

 

1 . 2
Gerror : JN - I Z (ml’rt'tlrtfrwl — mat‘tual) (312)

where the summation is taken from all the calibration data. The deviation for the ‘north’
metering system was 0.023 kg/sec, or 0.5% of the measurement range. The deviation for
the ‘south’ metering system was 0.019 kg/sec, or 0.4% of the measurement range. These
numbers support the replacement of the integral appearing in equation 3.2 with the con-
stant of equation 3.3. They also indicate the relative insensitivity of the technique to the 90

degree rotation of the rectangular calibration nozzle.

 

5.00

 
 

f
4.50 )—
400 ._

\ m = 3.5348*(En-0.49)0'5

   

Mass Flow Rate (kg/s)

f
)—
l’

2.50 i—

200 t—

1.50 ~ I

 

 

 

 

(En-0.490)"5

FIGURE 3.8 Calibration Data and Fit for the North Metering System

 

 

 

 

 

5.00
’5?
u 4.50
.M
V
3 4.00
N
m
3 350
a
_
PH
3 3.00
S
2.50
21X) IIIII l l I I I ll I I i I I I I I I I l I II
0.5 0.6 0.7 0.3 0.0 1.0 1.1 1.2
0.5
(Ea-0.5069)

FIGURE 3.9 Calibration Data and Fit for the South Metering System

3.2.3 Shroud Flow Rate - Elbow Meter System

31

The mass air flow through the aerodynamic shroud Coanda nozzle was determined by an
"elbow meter". A calibration of the elbow meter was executed using the Ford Haus facility
as shown in Figure 3.1 1. A calibrated venturi nozzle was used to measure the flow rate
through the system. A PC based AID system and two Validyne pressure transducers were
used to measure the pressure differences of both the venturi nozzles and the elbow meter.
As shown in Figure 3.1 1, the elbow meter reading is based upon the outside/inside pres-

sure difference: (Pout—Pm), which -- from the Euler - "n" equation

.3

9L) : PJL
an R (3.13)

is expected to be proportional to the mass flow rate squared. For a known density, this

becomes qa = f(z), where z = {p air - voltage of the pressure transducer} . The expres-

sion

qa = 2 - (— 4.57 —1.98z + 125.3921/2—1826021/3 +124.17z'/‘) (3.14)
provides an acceptable fit to these data. Note the factor of 2 used in equation 3.14. This is
used because the elbow meter is located downstream of a split in the delivery system, as
shown in Figure 3.12. The equality of the flow in the two branches was checked using a

method similar to that used by Morris (1 1).

 

Elbow Meter Calibration

 

 

 

 

250 "r T i l i l : fl I r T
”‘ Q = 4.57094 .982817.+125.3857z”2—182.60217.”’+124.17322”4 /
200 p- . __ . , L
r .
A
E 150 __ _5
Hut
0 -
v 3 j
o) 100 z==AV x pair [volts x kg/m ]
50 ._. _,
‘ I
O I 1 l L i I I A I 1
0.0 2.0 4.0 6.0 8.0 l0.0 12.0

FIGURE 3.10 Data Fit for the Shroud Elbow Meter Calibration

Spencer Blower

 

Elbow Meter

Control Valve
/

Venturi Meter

ll/

  
    

 

 

 

 

 

 

 

 

'J J

, , Q—
. Flow Conditioners :
I I Laboratory Air I '
d a «I. l

 

FIGURE 3.11 Calibration of the Elbow Flow Meter Using the Ford Haus Facility

33

 

(H (T

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

FIGURE 3.12 Overview of the Shroud Delivery with Elbow Flow Meter

3.2.4 Power Meter Calibration

Calculating the efficiency of a ventilation fan requires an accurate measurement of the
power consumed by the fan’s motor. This was accomplished by measuring the instanta-
neous voltage across the motor (e,) and the instantaneous current (1,) delivered to the

motor. The average power defined by equation 3.15 is calculated in its discretized from

through the relationship (3.16).

 

T.
— l . . _)
go—T2_Tl [(10) c(t))dt (3.15)
T.
N
”:15 Ztiren (3.16)

.:l

34

OI’

10= 7-E'cost (3.17)
Since the instantaneous voltage and current (and not the RMS values of the right hand side
of equation 3.17) are measured, the phase relationship (I) is not required for the evaluation
of p. The value of N and the increment in j of (3.16) are selected to ensure that an even
number of cycles are represented in the summation. The instantaneous voltage and current
were measured by using a voltage divider and a current sensor, respectively. Each of these
devices output a voltage that was measured using Analog-to-Digital (AID) hardware. A
detailed schematic of this custom circuit is shown in a model CLN-25 closed loop Hall
effect current sensor from F.W. Bell Technologies was used for these measurements. This
sensor is capable of measuring either AC or DC currents on a continuous time basis and
provides electrical isolation between the current carrying conductor and the output of the
sensor. The manufacturer’s specifications indicate that the sensor’s output is linear to bet-
ter than :0.2% with a response time of under 1 us. For calibration, the current sensor and
voltage divider were calibrated against a reliable source, the Hewlett-Packard 6632A sys-
tem DC power supply. Using this reference, the following transfer functions were estab-
fished.

System Amperage = 2.5228(ECS) - 0.004 (3.18)
System Voltage = 20.063(EVD) + 0.0021 (3.19)

Where Ecs represents the voltage output of the current sensor and Eve represents the volt—

age output of the voltage divider.

35

+1

 

 

 

 

 

 

11 CLN-zs [2 W 6 .
7 3 V005: f(l)
3:20012

+15

‘ -15
w p / LT 1028

to 115 VAC £33,,
test fan ‘ 2 \7<\/

out
4

-15

 

 

5
J / LT 1028
7
4
-15

 

w

 

 

 

 

f\

 

 

 

2AA

 

AAA

 

 

 

1, Power Supply
Ground

FIGURE 3.13 Schematic of the Current and Voltage Circuit Used to Measure Fan Power

3.3 Data Acquisition and Processing System

The integral data and hot-wire measurements were acquired using an Analogic Fast-16 A/
D (Analog-to-Digital) board that was placed in an Intel x86 computer. Measurements that
needed to be sampled simultaneously were collected using an 8 channel "sample and hold”
card that "instantaneously" stores data from 8 channels before writing to the storage, with-
out a channel-to—channel delay in scanning. The boards were run in a differential mode,
with a range of :10 volts. This range resulted in a least significant bit (LSB) of ~0.3 mv.
This value was confirmed by connecting a short circuit to the channel inputs and noting

the "bit flipping" around E = 0.

36

Collected raw data were transferred to a DEC Alpha (xXP- l 50 computer. Custom FOR-
TRAN programs, which were written in part by the author, were used to process the data.

Post-processing programs that were used include Matlab, Excel, PSI-Plot and TecPlot.

3.4 Pressure Transducers

Several different types of differential pressure transducers were used, depending on the

required range of the measurement. For the APkm, an MKS Baratron model 310CD-0001

head unit and an MKS Baratron 170M-6C signal conditioner were used. The 310CD-0001

has a range of 0- l 34 Pa. For the APShmud, an MKS Baratron 398HD-00010 head unit with

an MKS Baratron type 270B signal conditioner was used. This unit has a range of 0-1340
Pa. Finally, the pressure drop across the elbow meter that monitored the shroud flow rate
was measured using a Validyne DP15-20 head unit with a CD15 carrier demodulator. The
reported accuracy of the DP15-20 is :0.25%, which includes effects from non-linearity,

hysteresis, and non-repeatability.

3.5 Optical Encoder

The rotational speed of the fan was measured using a custom designed and built optical
encoder circuit. The schematic of this circuit is shown in Figure 3.14. Infrared light is gen-
erated from a diode device that sends light directed towards the hub of the fan (labeled A-
K in Figure 3.14). A small piece of reflective tape was placed on the hub that reflected the
light back to the collector-emitter (of the transistor) side of the circuit (labeled C-E in
Figure 3.14). The output signal of this device is 0 - 5 volt square wave that was measured
by the AID board. The fan rotational speed was calculated from this signal. Using the mea-

sured average rotational speed, the fan tip for any particular test can be calculated. The fan

37

. ‘wr.

tip speed is necessary for the calculation of the system efficiency, h*, as defined in

equation 2.17.

+5V

  
  
 

Emitted infra-red light
15 K0

 

 

 

w
+5V 150 Q A / Output to AID board
K E
_1_ Tl.—

FIGURE 3.14 Circuit Schematic for AFRD Optical Encoder

3.6 Uncertainty Considerations

The reliability of measurements made in the AFRD facility is important for both the rela-
tive comparisons made in the laboratory, but also for comparing results generated in the
AFRD with those from other facilities. This can be divided into two separate measures of

uncertainty: repeatability and accuracy.

3.6.1 Repeatability

The uncertainty of repeated (test-to-test) measurements in the AFRD facility are shown in
Figure 6.1 on page 76 for flow rate (performance) measurements and on Figure 6.2 on
page 76 for efficiency measurements. References [l] and [9], provide criteria for accept-
able tolerances when making comparative measurements. Ford et al. (1999) list that a
fan’s airflow and efficiency must fall within i 6% of the listed values to maintain certifica-

tion. This criteria is established as the basis for individual fan companies to "challenge"

38

the published (Ford et al. ( 1999)) values of a competitor’s fan. AMCA (Air Movement and
Control Association, Inc.) provide a slightly different criteria in [3]. They list that the air-
flow must be within i 3-4% over the entire range of the performance curve, and within i
5% for the power efficiency. The four individual tests that are shown in Figures 6.1

and 6.2 were used as a measure of the repeatability of the data collected in the AFRD

facility.

TABLE 1. Flow Rate and Efficiency Data from Four Independent Tests

 

 

 

 

 

 

 

 

 

 

 

 

 

Test # Interpolated Q (m3/s) 71*
Pressure (Pa)

Test 1 25.0 2.79 9.56

Test 2 25.0 2.79 9.73

Test 3 25.0 2.78 9.46

Test 5 25.0 2.76 9.25

 

 

TABLE 2. Mean, Standard Deviation and Uncertainty from the Data in Table 1.

 

 

 

 

 

 

 

 

 

 

 

% Uncertainty
Variable Mean Standard Dev. (Sc/5c) x 100
Q (m3/S) 2.779 0.0144 0.52
T]* 9.500 0.2006 2.11

 

 

The data in Table 2 show the repeatability to be 0.52% for volumetric flow rate (Q) mea-
surements and 2.1 1% for efficiency (11*) measurements. These values fall within the

guidelines established by [l] and [9].

3.6.2 Accuracy

39

The accuracy of measurements was determined by assessing the individual uncertainties
associated with each of the primary variables that were measured. A comparison of mea-
surements collected in the AFRD vs. published literature as found in Ford et a1. (1999)
will be shown later in Section 6.0. This comparison is not exact, since the configurations
are slightly different; however, these differences are noted and estimates of their effect on
volumetric flow rate (Q) and efficiency (11*) are noted in Section 6.0. This section
addresses the accuracy of the measurements with respect to the techniques and equipment

that were used.

3.6.2.1 Accuracy of Flow Rate Measurements

The accuracy of the volumetric flow rate measurements (Q) depends primarily on the
accuracy of the discharge coefficient that is used for the calibration nozzle. Section 3.2.1
details the two separate approaches that were used to determine the discharge coefficient
for the planar contraction nozzle. Figures 3.6 and 3.7 show the agreement between the two
techniques. The accuracy of the scale model approach can be analytically determined by
recalling the following relationship:

Cd = UQA (3.20)

 

The numerator, Q, is measured by the Testek model 1 1359 sonic nozzle test stand, which
has a rated (NIST traceable) accuracy of 0.5% of a mass flow reading, as reported by [20].

The uncertainty of the Cd value can be determined by:

 

 

acetenttrawl)“

40

From this equation, the uncertainty of the measured Cd value as described in Section 3.2.1

was estimated to be $0.006. This value also agrees well with the stated repeatability for

flow rate measurements, as shown in Table 2.

3.6.2.2 Accuracy of the Efficiency Measurements

The accuracy of the efficiency (11*) measurements depends primarily on the accuracy of
the electrical power measurement technique, which is described in Section 3.2.4. The
accuracy was evaluated by comparing the output of the current and voltage measuring
device (described in Section 3.2.4) with that of a known reference current and voltage.
These reference values were measured using a Hewlett-Packard 3457A multi—meter. Using

the following relationship for the uncertainty of the power measurements,

7

‘—

510 = mg?) - 51f + [(8853) - 56],}1/ (3.22)

and then taking the power consumed by a representative operating condition of AP = 25

 

Pa, the uncertainty of the power measurement was estimated to be 2 0.22%. This number,
which was calculated without considering the manufacturer’s specifications, is slightly
higher than the rated uncertainty of the Hall effect current sensor (see Section 3.2.4), indi-
cating that the current measurement is the dominant source of uncertainty in the power

measurement.

4]

4.0 Experimental Configuration

Data were collected in the AFRD Facility with the fan and diffuser cone situated as shown
in Figure 4.1. This vertical-flow configuration differs from the fan’s typical (wall-mounted
and horizontal-flow) configuration. This configuration accommodated the AFRD Facil-
ity’s calibration inlet. Further detail can be seen in Figure 4.1, which shows the schematic
layout of the fan and cone combination. The auxiliary air supply that pressurized the
shroud was delivered through PVC piping and into the four corners of the aerodynamic

shroud’s housing.

4.1 Baseline Configuration: A Standard 0.61m Production Assembly

The initial set of measurements were taken with an Aerotech standard four blade fan that

had a 7-degree stock diffuser cone. This combination of components is very similar to that

found on the Advantage© style fan, which is shown in Figure 4.2. This line of fans is
Aerotech’s hi ghest-quality, and subsequently, best-performing fan. Figure 4.2 shows that
there is a set of inlet shutters upstream of the fan plane and a safety guard downstream of
the fan plane. Neither of these accessories were included in the assembly that was tested
for this study. It is recognized that these will provide additional system resistances that are
not included in the tested assembly: however, there is an additional system resistance from
the piping for the shroud delivery system that is included in the tested assembly, for both
the "shroud off" and "shroud on” test conditions. It was determined that the pressure rise
created from this additional system resistance was comparable to that of the missing com-
ponents. This point is further discussed and quantitative data are provided to reinforce this

assertion in Section 6.1.1.

42

- 3.1“; ouq

4.2 Modified Configurations

4.2.1 Different Fan Blade Designs

A new blade design was considered after fully evaluating the baseline configuration in
both a "shroud-off" and a ”shroud—on" configuration. A flow visualization and subsequent
particle image velocimetry study were conducted, both of which provided initial insight
into the limitations of the production 4-blade fan. This new blade design incorporated the
salient features that were thought to be an improvement from these limitations. The 5-
blade version of this newer fan design is shown in Figure 4.4. and the results of this design

in both the "shroud off" and "shroud on" are presented in Section 6.4 on page 65.

Further data were collected on fan designs that were similar to that of the 5-blade fan in
Figure 4.4. An 8-b1ade and a 3-blade fan, as shown in Figures 4.5 and 4.6, resp., were

tested and these results can be found in Section 6.5 on page 69.

4.2.2 Installed Centerbody Device

The flow Visualization study also suggested that a centerbody device, as shown schemati-
cally in Figure 4.7, might improve the performance and efficiency of the fan system, par-
ticularly when used with the newer blade designs described in Section 4.2.1. The
centerbody was a cylindrical section that has a diameter that matched that of the 5—blade
fan. Figure 4.8 shows a close-up View of the hub region and top portion of the centerbody.
A 2.0 mm gap separated the top edge of the centerbody from the trailing edge of the hub.
The results of the centerbody and 5-blade fan can be found in Section 6.6. This configura-

tion was only evaluated with the 1 1 degree cone.

43

Shroud Airflow Delivery Piping

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4-Blade Fan with
7 Degree Stock Diffuser
Cen ' Blower
L _
Upper Receiver

ofAFRD

FIGURE 4.1 Schematic View of the Production 0.61m Ventilation Fan and Diffuser Cone

 

 

 

 

 

 

FIGURE 4.2 Schematic of Standard Production Assembly

44

0.066 m

 

FIGURE 4.3 Production 4-Blade Propeller

D1,,“ = 0.610 m

 

FIGURE 4.4 Prototype 5-Blade Propeller

45

Dc... = 0.608 m

 

FIGURE 4.5 Prototype 8—Blade Propeller

 

FIGURE 4.6 Prototype 3-Blade Propeller

46

Shroud Airflow Delivery Piping

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5-Blade Fan with
11 Degree D'fl’user
Blower
= Centerbody Device
11
Centerbody Support

 

 

 

FIGURE 4.7 Schematic View of the Production 0.61m Ventilation Fan and Diffuser Cone
with Installed Centerbody

47

5 III ’ \

F 5 Motor

 

 

 

 

 

 

 

 

Cylindrical Centerbody ’ . 2.0 mmJ
5-B1ade Fan

 

 

 

 

 

FIGURE 4.8 Close-up View of Centerbody and S-Blade Hub

48

5.0 Experimental Techniques
5.1 Integral Measurements

The primary quantities of interest in turbomachinery exist in an integral form. These vari-

ables are the volumetric flow rate, Q, expressed in m3/s, the pressure rise across the fan
blades or fan system, which is denoted as Ap, and expressed in Pa, and the system effi-

ciency, 11*, defined as

Q ' (épUtip)

n = p (5.1)

 

The various definitions and interpretations of efficiency are explained in detail in
Section 2.0, but the choice of using 11* was made because it was non-dimensional and it
also has a solid physical interpretation. It is recognized that the industry standard defini-

tion is

Q
11., = F (5-2)

This definition of efficiency, 11,, is found in literature that is standard in the agricultural
ventilation industry; see Ref. [1] and [9]. The data found in these references were con-
verted to the definition of efficiency as found in equation 5.1, since all the information can
readily be obtained in either Ref. [1] and [9]. These values were then compared with
industry accepted standards to ensure the reliability of the measurements. Care was taken
to ensure that the measurement procedure was similar to that of BESS and AMCA testing
standards. The one omission is the placement of shutters upstream of the fan plane. It is
recognized that in practice these are a detriment (albeit necessary for safety issues) to the

performance and efficiency of a ventilation fan, but their inclusion was not feasible given

49

the configuration detailed in Figure 4.1. Since the baseline measurements did not incorpo-
rate shutters, the relative differences (expressed in percentages) are expected to remain

consistent when shutters are included in the final production version.

When the efficiency was evaluated for a fan with the aerodynamic shroud activated,

termed the "shroud on" condition, a more specific definition of equation 5.1 was used:

7
. U“.

Aflfan + APshroud . Qshroud . (1 /ns/1roud)

 

This definition expands the denominator from equation 5.1 to separate the individual

power consumption of the fan, denoted by pfan, from the power consumed from the
added apparatus of the aerodynamic shroud. The term 11 shroud is the efficiency of the aero-

dynamic shroud, which is estimated for this work. Since a non-optimized assembly is used
to pressurize the shroud, evaluating its efficiency directly is not instructive for this study.
Instead, information found in a previous study by Herz (1998) was used to estimate the

value 11 shroud = 0.70 for all the "shroud on" cases.

The value of 'lshroud = 0.70 was chosen based on Herz’s conclusions that the efficiency

range for centrifugal blowers is between 0.60 and 0.80. An efficiency of 0.80 corresponds
to a backward curved centrifugal blower that is optimized for higher flow rates and lower
pressure rises. Similarly, an efficiency of 0.60 is representative of a forward curved design
that is optimized for lower flow rates and very high pressure rises. Since the aerodynamic
shroud would require performance characteristics similar to those provided by a backward
curved assembly, the author speculates that this type of blower, and the corresponding effi-

ciency, could be used for the final assembly. Assuming the use of this type of blower, the

50

value of nshroud = 0.70 is chosen after taking into account losses that will occur in the flow

delivery system downstream of the auxiliary blower.

These quantities were measured using techniques and equipment that are defined in
Section 3.0, using a standard 16-bit A/D board. Once an operating condition was set
(using the throttle on the AF RD as described in Section 3.1.1) the static quantities of flow
rate, pressure rise, shroud pressure and shroud flow rate were sampled for 10 secs at 100
Hz. Other quantities, including input voltage, input current, and shaft RPM, were sampled
at 10 kHz for 10 secs, using a sample and hold card. This sample and hold card insured
that these channels were sampled simultaneously, rather than the small time delay that
exists for the baseline A/D board. This requirement was necessary for calculating the
power as defined in equation 3.15. The sampling time of 10 secs was determined to be

optimal for statistically converged measurements.

Specific operating conditions were targeted for each test, nominally values of Ap = 0 Pa,
12.5 Pa, 25 Pa, 37.5 Pa, 50 Pa, 62.5 Pa, and 75 Pa. These values were selected since they
correspond to the same operating conditions tested in Ford et a1. (1999) and direct compar-
isons could be made with only small interpolations. The plots in Section 6.0 show typi-
cally two measured points at each operating condition. These redundant measurements
were made to determine the repeatability of each data point, and also to indicate if there
were any hysteresis effects. Each test involved first running the fan from the "shut-off"
condition to "free-delivery" condition and then reversing the throttle on the auxiliary
blower of the AFRD so the fan would move from a "free-delivery" condition back to a

"shut-off" condition. There were minimal hysteresis effects in most of the test cases.

51

5.2 Particle Image Velocimetry (PIV) Measurements

A better understanding of the flow field in the fan system environment was obtained
through the use of Particle Image Velocimetry (PIV) measurements. PIV is a measurement
technique that provides instantaneous velocity maps of a particular region of interest with
two or three components of the velocity being recorded. The two components are in the
plane of the light sheet. A two-camera stereo system permits the out-of—plane component
to be determined. These instantaneous maps can be ensemble averaged to provide time
mean velocity information for a particular region of the flow. A description of the PIV
measurement technique, along with specific techniques that were used for this study, will

be provided in this section.

PIV measurements require that the fluid motion can be reliably traced by neutrally buoy-
ant seed particles that are added into the flow. When a thin slice of the flow field is illumi-
nated by a light-sheet, the illuminated seeding scatters the light. This reflected light is
recorded by a camera that is placed at a right angle to the light sheet. The light sheet is
pulsed, very quickly (~5 ns), twice at a precisely known interval (150 - 500 11s for the
present study). The first pulse of the laser captures an image of the initial positions of the
seeding particles in the light sheet. The camera frame is then advanced and the second
frame is exposed to the light scattered by the particles from the second pulse of the laser.

For each desired instantaneous velocity map, a pair of images are recorded.

52

Image 2

Interrogation
Region

Lx Image 1

FIGURE 5.1 PIV Interrogation Region

  

The two images are then processed to find the velocity vectorI map of the flow field. First
each image pair is subdivided into a set of interrogation regions. The displacement of
groups of particles between image 1 and image 2 in each interrogation region is measured

using cross-correlation techniques, which are implemented using FFT algorithms as

described in [8] and [16]. This displacement is tenned: (I . Since the time interval (t)
between these two images is well—known, the velocity can be easily inferred from the fol-

lowing equation:

? = M- (21/1), (5.4)
where M is the magnification factor between the camera’s CCD chip and the measurement

area.

A single-camera digital PIV system2 was used to acquire the velocity data in the wake of
the test fan. A Rosco brand theatrical fog machine was used at the fan inlet to produce a

relatively uniform seeding density at the measurement locations. Two ND-Yag pulsed

 

1. The present technique obtains the projection of the 3-D velocity vector onto the image plane. The image
obtained is a vector: P z in +jv.

2. A Dantec PIV2000 processor and PC based data acquisition system are subcomponents of the Multi-
Phase Flow Facility (MFF) of the College of Engineering at MSU.

53

532nm lasers and a Kodak Megaplus ES1.0 CCD camera was used. These ND-Yag lasers
illuminated the region of interest (nominally 20x20cm) with a 1 mm thick laser sheet. The

two lasers were triggered sequentially with a typical separation time of 0.3 msec.

The digital camera was used to acquire two grayscale digital images. The measurement
region was divided into 32x32 pixel interrogation regions to calculate the individual vec-
tor displacements. The calculation used a Gaussian window function and a 2-D Fourier
transform of the image data to correlate the two images. The individual interrogation
regions were overlapped by 25% (8 pixels) for increased resolution. This resulted in 1722

measured velocities for each set of images acquired.

The configuration described above provided accurate results given that the seeding was
distributed in a nominally uniform manner in each image. However, at some locations
high shear or poor seeding would cause poor signal to noise ratios that would lead to vec-
tor maps with many "bad" or "spurious" vectors. These were post processed using several
validation techniques to ensure the integrity of the data. First, a "peak validation" was used
to ensure that the maximum value of the cross correlation was a minimum of 1.2 times
greater than the next highest value. This calculation removed typically 90% the suspect
vectors. Second, a range validation was used such that the magnitude of the velocity in the
plane of the laser light sheet could not exceed 25m/s, which is 3.2 times the average axial
velocity through the fan. Finally, a moving average validation was implemented as a final
check. This moving average validation is used to accept or reject velocity magnitude

based on a comparison between neighboring values. Continuity of the flow field’s behav-

54

ior is an implicit assumption in the moving-average validation method. Further details on

the algorithms used for these validations can be found in references [8], [14] and [16].

A total of 1000 vector maps were recorded and processed for each of the measurement
regions described in the following section. The time averaged velocity field was then com-
puted from the validated vectors. Note that phase averaging the data to the rotating fan
would have provided more information on the blade dynamics. However, it is the time
averaged quantities that lead to the net flow rates of interest, and therefore the latter repre-

sentations are used in this initial study.

5.2.1 PIV Measurement Locations

As previously stated, there were three distinct regions of interest for the fan/shroud combi-
nations. These locations, shown in Figure 5.2, include: (I) at the exit plane of the discharge
cone, (II) the near wall region of the discharge cone, and (III) the region directly down-
stream of the hub. The selection of these regions was based on the following consider-
ations: Region III was selected after initial observations with wool tufts indicated that the
angle of attack of the blade in the near hub region was not sufficient to impart axial
momentum into the flow. Region II was measured to determine if the boundary layer
along the wall of the 1 1 degree cone remained attached. Previous studies (Patterson (193 8)
and Squire (1953)) had indicated that in general, axial fans equipped with discharge cones
that were greater 9 degrees would exhibit a separated boundary layer (stall condition) near
the wall. This boundary layer separation would result in decreased fan performance.
Region I was selected to measure the exit profile of the fan and more specifically, to deter-

mine if a reversed flow condition existed near the centerline of the fan. Preliminary obser-

55

vations using wool tufts suggested that the velocities were highest at the outer region of
the fan, with much lower velocities near the mid-blade region and near-zero velocities

towards the centerline of the fan.

 

 

 

 

 

 

 

 

 

Motor
Plenum for
Aerodynamic Shroud
1 ‘1 A
4- 1 III 1
‘7 i. _____ .1
II _/ 594.4mm
~-\.JI Direction of Flow
1
fr 1 l 111 1 1 1 :,
+xV i_._LJ_IJ.__1
422.9mm
(i

FIGURE 5.2 Location of PIV Measurement Regions

56

6.0 Results and Discussion - Integral Measurements

6.1 Baseline Measurements - 4-Blade Fan with 7 Deg. Diffuser Cone

6.1.1 Baseline Measurements (Production Assembly) - "Shroud Off" Condition

The repeatability of the measurements was established by taking a series of independent
performance (AP vs. Q) and efficiency (AP vs. r1*) curves before experimenting with vari-
ous shroud configurations. The results of this study are shown in Figure 6.1 and

Figure 6.2. A total of five such data sets were collected and four of these are shown in the
figures. Since this configuration is very close to that of the current production assembly,
these data will serve as the reference condition to which all future data sets will be com-

pared.

These data were compared to published data for the Aerotech AT24Z, Test: 93086, as
found in Ford et a1. (1999). This comparison is to ensure that these results were consistent

with prior performance measures. A table of these values is shown below.

TABLE 3. Comparison of Results Obtained in the AFRD vs. Published Data

 

 

 

for (Ap = 0.10 in H20) (BESS Labs) Data from Published
Data
2.71 2.78 +2.58

Volume Flow Rate (m3/s)

 

Efficiency (11*) 8.59 9.28 +8.03

 

 

 

 

 

 

 

 

 

The Aerotech AT24Z is close to, but not exactly the same as, the fan that was tested in the
AF RD. The AT24Z does use the same propeller as was used on the experimental fan, but
the two fans have shrouds that are quite different. The AFRD data does show a slightly
higher flow rate and efficiency than the published data. This is caused, in part, by the

absence of inlet shutters that are on the production fan. Ford et a1. (1999) states that the

57

addition of inlet shutters will typically reduce the efficiency and flow rate of a fan by 10 to
25%. The differences indicated in Table 3 are below this range. These differences, which
are lower than expected, can be explained by noting the different inlet geometries of the
two shrouds. The inlet side of the experimental fan’s shroud contains the additional appa-
ratus of the delivery pipes for the aerodynamic shroud (even though these are baseline
data, the aerodynamic shroud delivery system was not removed). This piping system cre-
ates additional blockage that would not be found in a final, production configuration. Thus
the close agreement, as shown in Table 3, between the data for the AFRD-tested configu-
ration and those found in the literature indicates that the blockage of the shroud delivery
system is close to the blockage created by the inlet shutters and the safety guards. These
data confirm that the baseline data measured in the AF RD are accurate and reliable, and
consistent with the analysis that was derived in Section 3.6. Finally, it should be noted that
once an optimal configuration is identified, a final production aerodynamic shroud would

incorporate a more aerodynamic inlet to minimize the blockages of the prototype.

The data presented in the following section show the entire performance range (i.e. from
"shut-off" condition to "free-delivery" condition). However, specific percentage gains or
losses will often be noted based on a typical operating condition. This operating condition,
chosen to be nominally 25 Pa, is based on information obtained from Ford et al. (1999)
where they specifically state that a ventilation building fan will operate at a range around
25 Pa. They note that slight variations can occur if the housing and shutters of the fan sys-
tem are excessively dirty (thereby increasing the AP) or if there is a strong headwind fac-

ing the downstream exit plane of the diffuser cone (again increasing the AP). Despite these

58

variations, a typical fan system (fan + shroud + inlet shutters) will operate with a pressure

rise of ~25 Pa for the majority of its operating life cycle.

6.1.2 Production Assembly - "Shroud On" Condition

Once baseline measurements were established to be accurate and repeatable, the effects of
a range of shroud pressures were examined with this experimental configuration. The
resulting pressure rise vs. flow rate (AP vs. Q) and pressure rise vs. efficiency (AP vs. 11*)
are shown in Figure 6.3 and Figure 6.4, respectively. From Figure 6.3, there is a clear
increase in flow rate as the shroud pressure is increased from the "off" condition, to pres-
sures of 249 Pa, 498 Pa and 996 Pa. These gains are obtained at the cost of by the decreas-

ing values of efficiency that are shown in Figure 6.4. Examining the typical operating

condition of AP z 25 Pa, the flow rate increases from 2.77 m3/s to 3.15 m3/s, from the
"shroud off" condition to the "shroud on" condition of 996 Pa. This represents a gain of
nearly 14% in volume flow rate. This would be a notable increase if a comparison of the

efficiencies at the same shroud conditions ("shroud off" condition to APshroud = 996 Pa

were not for the accompanying decrease in efficiency, in this case from approximately
9.65 to 7.07, a decrease in 26.7%. From these two figures, the aerodynamic shroud did not

provide an unambiguous benefit for this production assembly.

6.2 The 4-Blade Fan with an 11 Deg. Diffuser Cone

After initial measurements with the production assembly did not show appreciable
improvements in performance without substantial decreases in efficiency, a larger diffuser
cone, 11 degrees, was incorporated into the system. Numerous references, among them

Wallis (1983), have stated that diffuser angles for axial fans cannot be larger than 7-9

59

degrees before the boundary layer that is formed in the sidewall of the cone will separate.
This separated flow will result in decreased performance and efficiency, offsetting the
static pressure recovery gains that would exist in the inviscid case. Even in the cases noted
by Wallis, a diffuser cone of 9 degrees represents the upper limit and is attainable only in
carefully controlled laboratory conditions. Flow separation would appear in an actual
installed condition for diffuser cone angles higher than nominally 7 degrees. The current
production diffuser cone has an approximately 7 degree exit angle. Aerotech, Inc. has
tested diffusers with larger angles, but the 7 degree diffuser cone yielded the optimal
increases in performance and efficiency of any other devices. It should be noted that they

have evaluated numerous cones with both larger and smaller angles.

Although these larger diffuser cone angles (> 7 degrees) have larger geometric exiting
areas, the separated flow in the exiting wall boundary layer reduces the effective flow
area. A device that would maintain attached flow in the boundary layer could make the use
of these larger diffuser angles feasible. Specifically, the wall jet that is produced by the
aerodynamic shroud could provide sufficiently high momentum to prevent this separated
condition. The current production 7 degree cone has an exit area of 0.46 m2. A new dif-
firser cone, with an angle of 11 degrees was fabricated to test this hypothesis. This nomi-
nally 11 degree cone had a much larger exit area of 0.56 m2. The potential increase in
static pressure recovery from this larger exit area is notable and was one of the primary

motivations for evaluating larger diffuser cones with the aerodynamic shroud.

It should be noted that describing a diffuser by only its exit angle is ambiguous and insuf-

ficient, since it does not fully describe the capabilities of the device. The salient features of

60

these diffusers is shown in Table 4 and schematically explained in Figure 6.5. One can
quickly see that a 7 degree cone with a much larger length, L, could have an outlet area
comparable to that of the 11 degree cone. However, this configuration would also have

much larger sidewall losses, designated as ksh and described in Section 2.0, “Fan Analy-

sis,” beginning on page 8. There is also an additional constraint that is less apparent. The
use of diffirser cones longer than the current production cone (L = 594.4 mm) is problem-
atic because it will fall off from the weight of snow or heavy winds. Therefore, any new

device must remain within the current limits.

TABLE 4. A Comparison of the Salient Features of the Tested Diffuser Cones (see
Figure 6.5 for a description of the variables: a, b, and L)

 

 

 

 

 

 

 

 

 

 

 

 

Cone Size a (mm) b (mm) L (mm) Inlet Area Outlet Area
(m2) (m2)
7 Degree 622.3 762.0 594.4 0.3042 0.4560
11 Degree 629.9 845.8 594.4 0.3116 0.5619

 

 

 

6.2.1 The 4-B1ade F an and 11 Deg. Diffuser - "Shroud Off" Condition

The first tests conducted with the 11 degree cone were performed in the "shroud off" con-
dition. These measurements provided a baseline by which further tests could be compared.
Surprisingly, these data show a negligible difference in performance between the 7 & 11
degree diffusers, as shown in Figure 6.6. When the efficiencies for these two configura-
tions were compared, the results were very similar. This is evident in Figure 6.7. Further

explanation on why this occurred will be discussed in Section 7.0.

6.2.2 The 4-Blade Fan and 11 Deg. Diffuser - "Shroud On" Condition

61

After establishing baseline performance for the 4-blade fan with an 1 1 degree diffuser, this
configuration was evaluated with the aerodynamic shroud. Two shroud pressures, 249 and
498 Pa, were tested. Figure 6.8 shows the performance curves for the 7 degree and 11
degree diffusers at the lower shroud pressure. For reference, the 7 degree - "shroud off"
data are also shown. Each of the two "shroud on" conditions shows enhanced flow rate
from the baseline data, but very little difference exists between the 7 degree - "shroud on"
and the 11 degree - "shroud on" data. Taking an operating point at 25 Pa, a performance
increase of approximately 4.5% occurs at this shroud pressure. Figure 6.9 shows a plot
comparing the efficiencies of the same three configurations. From this figure, there is no
apparent improvement in efficiency between the baseline data, the 7 degree - "shroud on"

and the 11 degree - "shroud on" data.

Next the shroud pressure was increased to 498 Pa and the same configuration was evalu-
ated. Figure 6.10 shows a performance curve that compares the same three test conditions.
Here the 11 degree - "shroud on" condition shows appreciable gains in performance. By
taking an operating point at 25 Pa, there is approximately an 11.5% improvement com-
pared to the baseline data. However, Figure 6.1 1 shows that these performance gains are
accompanied by a decrease in efficiency. Taking an operating point of 25 Pa, the effi-
ciency decreases over 6%. The gains in performance are accompanied by the strong

decrease in efficiency.

6.3 The 4-B1ade Fan and 11 Deg. Diffuser Cone at Variable Shroud Gaps

6.3.1 The 4-Blade Fan and 11 Deg. Diffuser at a Medium Shroud Gap (g = 5.0 mm)

62

After the initial tests with the 4-blade fan in the "shroud on" condition failed to improve
the flow rate without having an appreciable decrease in efficiency, the effect of gap height
was evaluated. Previous tests had used a gap height of 1.0 - 2.0 mm, but additional tests
were taken with a gap height of 5.0 mm. By increasing this height (but maintaining the
same shroud pressure), the volume flow rate contributed by the aerodynamic shroud
would be greater and the resulting wall jet would be thicker. The idea was that this thicker,
higher momentum wall jet might be more effective in preventing separated flow on the
sidewall of the cone. Additionally, the entrainment capabilities of the jet might be
improved. These ideas were stimulated from the observation that the total flow rate mea-

sured was greater than the sum of the "shroud off" condition plus the shroud flow rate.

2
n1, : Qtotal ' (PUtip ) (6 1)
sofa” + APs/rroud ' Qs/rroud ' (1 /nshroud)

 

As shown in equation , the efficiency, 11*, has the total flow rate in the numerator, and the

product of APshmud and Qshroud in the denominator. Therefore, a substantial increase in
Qtotal with respect to a smaller increase in APshmudQShmud will result in increased 11* val-

ues. It was thought that this could potentially be accomplished through a larger gap height.

Figure 6.12 shows the results of performance data for the medium gap height assembly.
Pressures of 125 Pa, 249 Pa and 498 Pa were measured, and for each subsequently higher
shroud pressure resulted in a higher system flow rate. At the operating point of 25 Pa, a 16
percent gain in flow rate is shown between the "shroud off" condition and the highest

"shroud on" condition, APShmud = 498 Pa. However, at this shroud condition, the effi-

ciency, shown in Figure 6.13, was substantially decreased by 20.7%, which detracts from

63

the improvements in performance. Interestingly, at the lower "shroud-on" condition of

APshroud = 125 Pa, the efficiency actually increased from 8.9 to 9.5, which is an improve-

ment of 6.75%, while showing modest improvements in flow rate of 8%. This study
proved that with a different gap height, moderate gains in both flow rate and efficiency
could be simultaneously obtained. However, if the shroud flow rate is set excessively high,
the resulting higher improvements in flow rate are accompanied by drastically lower effi-

ciencies.

6.3.2 The 4-Blade Fan and 11 Deg. Diffuser at a Large Shroud Gap (g = 7.0 mm)

The data collected in Section 6.3.1 (the 4-blade fan with 11 degree diffuser and a medium
gap height) represent the first test condition where simultaneous gains in both flow rate
and efficiency were obtained for any particular combination of diffuser cone, shroud pres-
sure and shroud gap height. These results, specifically the notion of increasing the shroud
flow rate while simultaneously decreasing the shroud pressure, led to the testing of an
even larger gap height for the incoming jet of the aerodynamic shroud. This condition,

described as the "large" gap height condition resulted in g = 7.00 mm.

Figure 6.14 shows the performance data for the large gap height condition at the baseline
"shroud off" condition, and also for shroud pressures of 125 Pa, 249 Pa and 374 Pa. The
highest shroud pressure, while lower than the highest condition evaluated in the medium
gap height condition, represented the limit of the auxiliary blower that was used to pres-
surize the shroud. The performance data clearly show a systematic trend of increased total

flow rate for each subsequently higher shroud pressure condition. Taking the referenced

64

operating condition of AP z 25 Pa, the maximum increase of Q is nearly 16% from the

"shroud off" condition to the APshmud = 374 Pa condition.

The efficiency data for this test condition is shown in Figure 6.15. Although the shroud
pressure of 374 Pa had a substantial increase in flow rate, it was also accompanied by a
significant decrease (19.4%) in efficiency. The only shroud condition that did not experi-

ence a decrease in efficiency was the APShmud = 125 Pa, which had nominally the same

efficiency as the "shroud off" condition. This shroud pressure had small improvement in

flow rate at the operating point of 25 Pa, from 2.76 m3/s to 3.00 m3/s, a gain of 6.4%.

None of the measured shroud pressures at this gap height were able to increase both the
flow rate and the efficiency simultaneously. It was determined that the optimal condition

for the 4-blade fan is the medium gap height (g = 5.0 mm) with a Apshroud = 125 Pa. This

was the one condition for the 4-blade fan that could make simultaneous increases in both

Qandr1*.

6.4 Comparison of the S-Blade and the 4-Blade Fans

Observations made while testing the 4-blade assembly led to an interest in testing another
blade design with the aerodynamic shroud configuration. These observations were ini-
tially made using simple flow visualization techniques like tufting as described by Morris
(1997) and also introducing smoke into the flow. Subsequent PIV measurements con-

firmed and further characterized these initial flow visualization observations.

A radically different fan propeller, shown in Figure 4.4, was obtained from Aerotech, Inc.

This 5-blade propeller had been developed for possible production purposes, but was

65

abandoned when initial tests indicated that it did not provide an enhanced performance (or
in some cases, a reduced performance) when compared with the current production pro-
peller. Nevertheless, this fan had specific features which appeared to address observed
deficiencies in the 4-blade fan. For example, the flow visualization observations of the 4-
blade fan showed that there was a large recirculation region directly downstream of the
hub. This recirculation region was approximately the same width as the upstream motor.
The 5-blade fan has a large bowl shaped hub that could potentially reduce or effectively
eliminate this recirculation. Additionally, the 5-blade fan had thinner, more aerodynamic
blade profiles and the angle of attack increased substantially from the tip to the hub. These
features, it was reasoned, would have improved lift to drag coefficient ratio (CL/CD) and
would provide a more uniform velocity profile than that of the current 4-b1ade fan. These
observations provided the motivation to evaluate the 5-blade fan in both the "shroud off"

and "shroud on" conditions.

6.4.1 The 4-Blade and 5-B1ade Fans - "Shroud-Off" Condition

Initial measurements were collected to verify the previous data of Aerotech that indicated
that the 5-blade propeller would not provide higher flow rates and efficiencies than the
standard 4-blade propeller when measured in identical configurations. Since the 4-b1ade
fan showed nearly identical performance and efficiency characteristics in both the 7
degree and 11 degree diffuser configurations, the l 1 degree cone was chosen as the refer-

ence configuration to examine the 4-blade and S-blade comparisons.

Figure 6.16 shows the performance curves for these two propellers. For pressure rise val-

ues below 31 Pa, these data confirm the expected result that the 4-blade provides a higher

66

volume flow rate at each AP; however, for pressure rise values higher than 31 Pa, the trend
is reversed by roughly the same magnitude as was observed for the lower pressure rise
conditions. These magnitudes, approximately 2% for AP above and below 31 Pa were
considered to be not significant. Thus, the data collected in the AFRD assembly confirmed
the previous results; namely that the 4-blade propeller is better-suited to this application

than the 5-blade propeller.

A comparison of the efficiency values is shown in Figure 6.17. These data repeat the trend
exhibited in the performance data, where again an inflection point exists around AP z 31
Pa. Similar to the performance curve, the 4-blade fan has a higher efficiency at pressure
rise values below 31 Pa and a lower efficiency at pressure rises above 31 Pa. The magni-
tude of change is slightly larger for the efficiency, averaging approximately 5% over the
entire curve (but varying depending on each specific operating point). However, similar to
the performance data, the 4-blade propeller would be chosen over the 5-blade propeller,
since the operating point for these fans is normally around AP = 25 Pa. As shown in both
Figure 6.16 and Figure 6.17, the 4-blade propeller out-performs the S-blade at this pro-

jected operating point and its surrounding values.

6.4.2 The 4-Blade and 5-Blade Fan - "Shroud-On" Condition

After verifying the performance of the 5—blade fan in the "shroud—off" condition, this same
fan was then tested in the "shroud-on" condition. Considering the "shroud-off" conditions,
the 5-blade propeller was a less-desirable choice for a fan system, but it was thought that
this fan might perform differently when placed into the "shroud-on" environment. This

idea was based on qualitative observations made using the previously mentioned (tutting,

67

 

etc.) flow Visualization techniques performed during the initial testing of the "shroud-off"
condition. These observations showed the surprising result that the near wall flow
appeared to be separated near the exit of the diffuser cone. Given the large cone angle of
11 degrees, this result is expected; however, one would also then expected the perfor-
mance and efficiency of the 5-b1ade fan to be considerably worse than that of the 4-blade
fan. As shown by the detailed results of Section 6.4.1, this was not the case. Hence, further
experiments were conducted. A gap height of g z 5.0 mm was selected based on the previ-

ous 4-blade studies showing this to be an Optimal height for the delivery of the shroud jet.

The results of the 5-blade fan with shroud pressures of 249 Pa and 373 Pa, in addition to
the baseline (production) case of the 4-blade fan with a 7 degree diffuser cone, are shown
in Figure 6.18. Figure 6.18 shows a substantial increase in volume flow rate for both the

"shroud-on" conditions of APshroud: 249 Pa and APshmud= 373 Pa. For these data, the ref-

erence operating point (AP z 25 Pa_ shows performance (flow rate) gains of 20.5% and

32.8%, respectively.

Figure 6.19 shows the efficiency data for the 5-b1ade fan with the same shroud pressures,
249 Pa and 373 Pa. Examining the reference operating point of AP z 25 Pa, these two
shroud pressures indicate efficiency gains of 10.8% and 7.8%, respectively. These data
show substantial improvements in both performance and efficiency simultaneously. This
result had not been seen on any of the 4-blade tests except for the medium gap height, but
the percentage improvements measured here were significantly higher. These tests were
subsequently re-taken several times to ensure the validity of the results, and they were

shown to be repeatable to within the uncertainty of the measurements.

68

6.5 A New Blade Design - Comparisons of 3, 5 and 8-Blade Fans

The success of the new propeller design, shown in the S-blade tests of Section 6.4.2, led to
an investigation of similarly designed fans with slightly different characteristics. These
two new propellers, shown in Figures 4.5 and 4.6, are closely related to the 5-blade pro-
peller (larger hub-to-tip ratio, similar angles of attack at various radii, etc.), but they do
differ in the number of blades and solidity. A study was conducted to see if maintaining
the general salient features of the fan, but changing the number of blades could improve
upon either the already significant gains in efficiency and flow rate of the 5-blade fan.

Both the "shroud off" and "shroud on" conditions were evaluated in this study.

6.5.1 Comparisons of 3, 5 and 8-Blade Fans - "Shroud Off" Condition

Before evaluating the new propellers with the aerodynamic shroud, a series of baseline
data were acquired to determine how each one compares with the existing 4 and 5-blade
designs. As previously stated, the new 3 and 8-blade designs were quite similar to that of
the 5-blade fan, but the chord lengths were slightly different. The 3-blade fan had a
slightly larger chord and the 8-blade fan had a slightly smaller chord. The performance
data are shown in Figure 6.20. Previous results had shown that the 4-blade and 5-blade
had very similar volume flow rates at the reference operating point of AP z 25 Pa, with the
4-blade fan having a slightly higher flow rate. Figure 6.20 shows that the 3-blade fan gen-
erates a much lower flow than the baseline 4-blade fan (-8.3%). However, the 8-blade fan
generates a higher flow rate, showing an increase in volume flow rate of 6.5%. Consider-
ing the efficiency, shown in Figure 6.21, it is evident that both the 3-blade and the 8-blade
fans have much lower efficiencies at the reference operating point, by -15.0% and -31.2%,

respectively.

69

6.5.2 Comparisons of 3, 5 and 8-Blade Fans - "Shroud On" Condition

The 3-blade and 8-blade fans were then tested in the "shroud on" condition to determine if
the aerodynamic shroud could improve upon the initial performance and efficiency

results. Based on previous data, the shroud pressure, APshroud, was set to 373 Pa. This

shroud pressure had previously provided the optimal trade-off between flow rate increase
and efficiency increase. Since the overall fan designs were quite similar, it was reasoned
that this would be the best case for the 3 and 8-blade fans. The performance curve results
are shown in Figure 6.22. This figure shows that each of the new blade designs (3, 8 and
5-blade) have substantial improvements in flow rate over those of the current production

4-blade fan. The 8-blade fan has the highest performance, producing a volume flow rate of
3.77 m3/s at the reference operating point of AP z 25 Pa. The 5-blade fan generates the

next highest flow rate at 3.69 m3/s, and the 3-blade fan is the third highest with 3.52 m3/s.
These data indicate that the 8-blade fan with the aerodynamic shroud would be the choice

for maximum ventilation in an agricultural environment.

Figure 6.23 shows the corresponding efficiencies for the data in Figure 6.22. Once again
the 4-b1ade production fan is compared with the more recently designed fans with 3, 5 and
8-blade propellers in the "shroud on" condition. These data show that the 8-blade fan,
which had the highest flow rate, has the lowest efficiency, with an 11* value of 9.1. The 4-
blade production fan has an 11* value of 9.6. This is somewhat surprising given the sub-
stantially different flow rate values between these two fans (the 8-blade had a 36%
increase in flow rate). Both the 3 and 5-blade fans had 11* values higher than the produc-

tion fan, there values were nominally 10.3 and 10.4, respectively.

70

 

These data show that the 5-blade fan provides the best compromise between increased
flow rate and increased efficiency. This fan was able to generate a volumetric flow rate
that was very close to the 8-blade fan, without suffering the significant decreases in effi-
ciency experienced by the 8-blade fan. The 5-blade fan was therefore used as the reference

fan for further attempts at refining the overall fan system.

6.6 The S-Blade Fan with an Installed Centerbody Device

The results of Section 6.5 show that each of the later propeller designs, the 3-blade, 5-
blade and 8-blade can simultaneously improve the flow rate and efficiency of a the fan
system when they are used in conjunction with the Aerodynamic shroud. However, the 5-
blade provided the optimal increase of both flow rate and efficiency. Further flow Visual-
izations made while testing the S-blade fan suggested that other devices could possibly be
used in conjunction with the aerodynamic shroud to reduce losses, as described by
equation 2.20. These flow visualizations showed that a large region of reverse flow (i.e.
back into the control volume) existed near the center of the downstream exit of the control
volume of Figure 2.1. This reversed flow was substantially increased when the fan was
operating in the "shroud on" condition. Reversed flow near the exit of the diffuser cone is
a detriment to the system performance and efficiency in two distinct ways: (1) It repre-
sents lost kinetic energy that detracts from the net system flow rate and, (2) this reversed
flow creates a higher value for or, since it creates a highly non-unifonn profile at the exit
plane. These flow visualizations were later confirmed and quantitatively assessed by the
PIV measurements of Section 7.0. Subsequently, a new device, called a "centerbody" was
proposed to address these issues and determine if even further improvements could be

obtained.

71

6.6.1 The Cylindrical Centerbody

The first centerbody that was installed was a simple cylindrical shape, which was approx-
imately the same width as the hub on the 5-blade fan. This shape was chosen because it
corresponded to the approximate width of the reversed flow region and also because it
would block any recirculating flow downstream of the hub of the 5-blade propeller. This
straight centerbody also did not appear a priori to provide any significant blockage to the

primary exiting flow path.

6.6.1.1 The Cylindrical Centerbody — "Shroud Off" Condition

The results of the flow visualization indicated that the straight centerbody device would
most likely not benefit the fan in the shroud off condition. This was primarily because
there appeared to be very little reversed flow at the exit plane of the diffuser when the 5-
blade fan was tested in a "shroud off" configuration. However, there still existed the possi-
bility that the fan could benefit in other ways (i.e. downstream of the hub). Several sets of
data were collected to test this idea and also to provide a methodology which was consis-
tent with each previous step ("shroud off" vs. "shroud on"). Figure 6.24 shows the perfor—
mance curve for the 5-blade fan in the "shroud off" condition with the cylindrical
centerbody installed. The previously tested conditions of the baseline production assembly
(4-blade, 7 degree cone, and "shroud off") and the 5-blade, 11 degree cone in the "shroud
off" condition are shown for reference. From Figure 6.24, the performance is slightly
increased throughout the range of potential operating points by adding the centerbody
device. This increase is approximately 2.8% for the reference operating point of AP z 25

Pa.

72

 

Figure 6.25 shows a comparison of the efficiencies of the same three configurations that
were shown in Figure 6.24. These results show that the efficiency, which was essentially
the same for both the baseline production assembly and the 5-blade fan (both in "shroud
off") is slightly increased by the addition of the cylindrical centerbody. Again, the 5-blade
fan, "shroud off" condition is shown as a reference, along with the baseline 4-b1ade assem-
bly. Previous results had shown that the 5-blade fan had a decreased efficiency at the
higher flow rates, and Figure 6.25 shows that the efficiency is slightly increased at these
higher flow rates. This improvement, while minimal, shows that the 5-blade fan, when
evaluated with the centerbody device, can deliver a higher flow rate with a nominally
equivalent efficiency to that of the baseline production case. Using the reference operating
point of AP z 25 Pa, the efficiency is increased from z 9.7 for the baseline 4-blade assem-
bly to z 9.8 for the 5-blade assembly with cylindrical centerbody. This is also an improve-
ment from the efficiency of the 5-blade without cylindrical centerbody, which had an
efficiency of z 9.3. These data indicate that both the flow rate and the efficiency are

slightly increased by the addition of the centerbody for the "shroud off" case.

6.6.1.2 The Cylindrical Centerbody - "Shroud On" Condition

The cylindrical centerbody was subsequently evaluated with the aerodynamic shroud in
the "Shroud On" condition. The results of Section 6.4 indicated that the optimal shroud

pressure for the 5-blade configuration was APshmud = 373 Pa. This information was used

as the starting point for the "shroud on" condition, but additional shroud pressures were
also evaluated. Figure 6.26 shows the initial performance tests that were conducted at

APShmud = 373 Pa for the 5-blade fan with centerbody. The previously shown results for

73

the 5-blade fan at the same shroud pressure without centerbody, as well as the 4-b1ade
baseline fan are also shown. Figure 6.26 indicates that the addition of the centerbody is
able to increase the flow firrther than the same configuration without the centerbody.

Referring to the reference operating condition of AP z 25 Pa, the 4-blade fan has a Q z 2.8
m3/s, the S-blade with centerbody has a Q z 3.7 n13/s and the 5-blade with centerbody has

a Q 2: 3.8 m3/s. The 5-blade without centerbody configuration increases the flow rate by

32.0% while the 5-blade with centerbody configuration increases the flow rate by 35.7%.

The data in Figure 6.27 show the efficiency curves for these same three configurations.
The general trend in this figure is a consistent improvement in efficiency throughout the
range of operating conditions. Referencing Figure 6.27 and looking at the target operating
point, the 4-blade baseline fan has an efficiency of 9.6, while the 5-blade fan without cen-
terbody has an efficiency of 10.4 and the 5-blade fan with centerbody has an efficiency of
11.1, This is an improvement of 8.3% and 15.6%, respectively. The addition of the center-
body to the 5-blade fan operating with a "shroud on" condition of 373 Pa is able to

improve upon the already substantial gains in both flow rate and efficiency.

The effect of shroud pressure when operating with the centerbody was subsequently stud-
ied to determine if the optimal shroud pressure would be different than that of the same
fan without a centerbody. Figure 6.28 shows the performance curves for this configuration
when tested with both the "shroud off" condition. as well as "shroud on" conditions for
249 Pa, 373 Pa and 498 Pa. This figure shows the significant performance gains for each
subsequently higher "shroud on" condition. At an operating point of AP 2: 25 Pa, the

"shroud on" conditions of 249, 373 and 498 Pa show improvements of 23.5%, 34.3% and

74

38.5%, respectively. Figure 6.29 shows the associated efficiency for each of the three
tests. These results show that the efficiency (at AP z 25 Pa) is increased for each of these

three shroud pressures, APshmud = 249 Pa, 373 Pa, and 498 Pa to 11* = 10.9, 11.2, and 10.2

respectively. Comparing these 11* values to those of the 4-blade baseline production case,
which has an 11* = 9.8, these shroud pressures represent efficiency improvements of

11.2%, 14.3%, and 4.1%.

These data clearly show that the centerbody, when used in conjunction with the aerody-
namic shroud, does improve the performance and efficiency of the 5-blade fan. The maxi-
mum increase in flow rate occurs when the Apshroud is set to 498 Pa, and this maximum
increase in flow rate (38.5%) is accompanied by a 4.1% increase in efficiency. While this
is a notable increase in flow rate, it is only a modest improvement in efficiency. However,
if the shroud pressure is lowered slightly, to a APShmud = 373 Pa, then the flow rate is
increased by 34.3% with an increase in efficiency of 14.3%. This combination would
appear to be optimal; however, there may be situations where a maximum flow rate is
required and a more modest increase in efficiency could be tolerated. If this were the case,

then the shroud pressure of APshroud = 498 Pa might be preferred.

75

Shroud off, 4-Blade, 7 Degree Performance Comparison

 

 

  
  
  
  
  

 

 

 

 

 

 

70
7 Legend
60 [—
: + Testl
50 :— + TestZ
: Tet3
A 40 1? f s
a: e + Tests
v _
Q _
20 ——
E
10 L
(1 :_
1_
_1() T M l A I 1 l 1 l 1
1.00 1.50 2.00 2.50 3.00 3.50
3
Q(m /s)

FIGURE 6.] Baseline (Shroud 011) Performance Measurements for the Production 4-
Blade Fan with 7 Degree Diffuser. The Multiple Tests Show the Repeatability of the
Measurements

 

 

  
  
  
  
   

 

 

 

 

 

 

70
1.
t 1. a
(10 __ egen
.. + Test]
50 :— + Test2
A 40 L + Test3
as r
a C + Tests
a“ 30 :—
4 _
20 -
1__
10 ~—
0 L
-10 T 1 1 1 I 1 1 1 l 1 1 1 l 1 1 1 l 1 1 I
4 (1 8 10 12 14

FIGURE 6.2 Baseline (Shroud 0ft) Efficiency Measurements for the Production 4-Blade
Fan with 7 Degree Diffuser. The Multiple Tests Show the Repeatability of the
Measurements

76

Performance Comparison for "Shroud On" Conditions for 4-Blade Fan with 7 Degree Cone

 

 

 
   
 

 

 

 

  

 

 

 

80 E
70 Legend
1 -o— Shd off
60 - -I- Shd 249 Pa
l + Shd 498 Pa
5 5" 1‘ + Shd 996 Pa
E: 40 [_ I
<1 1
30 l—
.l
20 1-
L
10 [7
0 i» ._ _I_ I l I ' .___ Iw-’ l
1.00 1.50 2.00 2.50 3.50

Q (m3/s)

FIGURE 6.3 Performance Data for the 4-Blade Fan with 7 Degree Diffuser at Various
Shroud Pressures

Efficiency Comparison for "Shroud On" Conditions for 4-Blade Fan with 7 Degree Cone

 

 

 

 

 

 

 

 

80 , _. _, -_--_____
\ ,
70 -
- x 1 Legend
“0 l‘ \ ‘+ Shd off
so i - + Shd 249 Pa
3 1 + Shd 498 Pa
V 40 :5 |+ Shd 996 Pa
flu L \ .
<1 30 1 -
l .
20 1
l \‘
10 1—
0 E
-10 i ._I* A l i____l___._L_.i , I_.. 1
4 0 8 10 12 14

T111!

FIGURE 6.4 Efficiency Data for Data for the 4-Blade Fan with 7 Degree Diffuser at
Various Shroud Pressures

77

FIGURE 6.5 Geometric Variables for the Diffuser Cone (as described in Table 4)

for 7 vs. 11 Degree Cone

 

 

     
  
    

 

 

 

 

 

 

80 - — —
1. ._.
70 4 Legend
: -§ _ Shd Off. 4-Blade7Degree
(0
’ - _.L Shd Off.4-Blade,110egree
50 5 r“
7.? -
a. 40 -.
v
a“ .
Q 30 1:
20 g
10 4'
0 -
-10 _ _1__ A g i _1__. l 1 l P
1.00 1.50 2.00 2.50 3.00 3.50

Q (mi/s)

FIGURE 6.6 Performance Curves for the 4-Blade Fan for 7 vs. 11 Degree Diffuser
(Shroud Off)

78

7 vs. 11 Degree Cone

 

80

 

   
   
 

70 Legend

0 _ ._ Shd Off, 4-Blade 7 Degree
6

+ Shd Off. 4-Blade, 11 Degree

 

 

 

50

 

40

AP (Pa)

30

20

 

 

 

FIGURE 6.7 Efficiency Curves for the 4-Blade Fan for 7 vs. 11 Degree Diffuser (Shroud
0“)

 

80

70

60

50

40

 

AP (Pa)

30 L - Legend

+ Shd Off, 4-Bladc, 7 Degree i \
20 — + Shd 249 Pa, 4-Blade, 11 Degree 1
1+ Shd 249 Pa 4-Blade, 7 Degree

1“; ,
'0 F I I \
0 L- -#L I 1 l _ _ 1 1 i 1

1.00 1.50 2.00 2.50 3.00 3.50

Q (ms/s)

 

 

 

FIGURE 6.8 Performance Data for the 4-Blade Fan with 11 Degree Diffuser - Low
Shroud Pressure (APshroud = 249 Pa)

79

 

80

 
  
  
  
 

 

70 Legend

60 + Shd Off, 4-Blade. 7 Degree

+ Shd 249 Pa. 4-Blade. 11 Degree
50
_*_ Shd 249 Pa 4-Blade, 7 Degree

 

 

  

40

 

AP (Pa)

30

20

 

lll‘111I11—IT1III1lll1lll1lll1llT1flT

 

_1() l l 44 l l l I l l l I l I I 4L 1 I l

 

is
o
oo
5
i3
3

FIGURE 6.9 Efficiency Data for the 4-Blade Fan with 11 Degree Diffuser - Low Shroud
Pressure (APshroud = 249 Pa)

Shroud On (High), 4-Blade, 7 vs. 11 Degree Performance Curves

 

 

   
   

 

 

 

  
  

 

 

80 {j _ ‘1
70 1_ '1 Legend
1 1 + Shd Off, 4-Blade, 7 Degree
60 1- + Shd 498 Pa 4-Blade, 7 Degree
’1 -o— Shd 498 Pa. 4-Blade, 11 Degree
A 50 " l -
a t
EL 40 f
7 i
Q
30 1'—
2. l
10 L“
0 L 1 I l I l I l_, 1 I

1.00 1.50 2.00 2.50 3.00 3.50 4.00

Q (m3/s)

FIGURE 6.10 Performance Data for the 4-Blade Fan with 11 Degree Diffuser at a High
Shroud Pressure (APshmud = 498 Pa)

80

 

‘ 11

 

 

Shroud On (High), 4-Blade, 7 vs. 11 Degree Performance Curves

 

80
70 r:-
(10 I“

50 ’f

 

AP (Pa)

   
   

Legend
' + Shd Off, 4-Blade, 7 Degree

+ Shd 498 Pa 4-Blade, 7 Degree
+ Shd 498 Pa, 4-Blade, 11 Degree

 

 

 

6

FIGURE 6.11 Efficiency Data for the 4-Blade Fan with 11 Degree Diffuser - High Shroud

80

Pressure (APshroud = 498 Pa)

 

 

 

 

 

  
  

 

 

 

 

 

70 L
()0 :—
50 L
Q i
40 ”
E: E
g 30 qt Legend
20 : _,_ Shd ORA-Blade,” Degree
E + Shd 125 Pa 4-Blade, 11 Degree
10 1+
: --D— Shd 249 Pa, 4-Blade. 11 Degree
0 S“ + Shd 49s Pa,4-Blade. 11 Degree
-l() H I I 4 11 I I 1 I 1 L 1
l.()() 1.50 2.00 2.50 3.00 3.50 4.00
Q (m3/s)
FIGURE 6.12 Performance Data for the 4-Blade Fan with 11 Degree Diffuser - Medium

Gap Height (g z 5.0 mm)

81

 

 

Medium Gap (g = 5.0 mm), 4-Blade, 11 Degree Efficiency Comparison

80

 

 

 

 

 

 

 

: Legend
70 -
: —+— Shd Off, 4-Blade,ll Degree
60 j \ + ShdlZS Pa 4-Blade,ll Degree
50 ; —O——— Shd 249 Pa, 4-Blade, 11 Degree
A E _*_ Shd 498 Pa, 4-Blade, 11 Degree
a: 40 _ \
g 30 L
20 E—
E \.
IO — \
0 L ’
-10 FA #1 I J I I J I I I I I I l I I L I4I I I I
2 4 () a 10 12 14
9:
TI

FIGURE 6.13 Efficiency Data for the 4-Blade Fan with 11 Degree Diffuser - Medium Gap

Height (g z 5.0 mm)

Large Gap (g = 7.0 mm), 4-Blade. 11 Degree Performance Comparison

 

 

 

 

 

 

  

 

 

 

 

80
70 L
(30 L
50 I.
a 40 _
V : Legend
3 30 :r
_ .-__._. Shd Off. 4-Blade." Degree
20 Z-
: _*_ Shd 125 Pa 4-Blade. 11 Degree
0 Z-
I : —D__ Shd 249 Pa. 4-Blade. ll Degree
0 S“ _. t— Shd 374 Pa. 4-Blade. 11 Degree
-|() P I I I I 1 I 1 I I I 1
1.00 1.50 2.00 2.50 3.01) 3.50 4.00
3
Q (m /s)
FIGURE 6.14 Performance Data for the 4-Blade Fan with 11 Degree Diffuser — Large

Gap Height (g z 7.0 mm)

82

 

Large Gap (g = 7.0 mm), 4-Blade, 11 Degree Efficiency Comparison

 

 

 

 

 

 

 

 

80 P
: \ i* Shd Off. 4-Blade,ll Degree
()0 —
5 ) E + Shd 125 Pa 4.31m, 11 Degree
( __
Q : "‘Cf— Shd 249 Pa. 4-Blade, 11 Degree
40 L \
EL : ---*_ Shd 374 Pa. 4-Blade, 11 Degree
m 3( '— .
) _
q :
20 L
H.) E—
o I'—
I:
_| 0 I— I I I I I I I I I I I I I I 4 I 41 I I
4 6 8 IO I2 M
T] *

FIGURE 6.15 Efficiency Data for the 4-Blade Fan with 11 Degree Diffuser - Large Gap
Height (g z 7.0 mm)

Shroud Off, 11 Degree, 4-Blade vs. S-Blade Performance

 

 

 

70 I- g _ m“.
I I Legend

()0 I I -I— Shd Off, 4-Blade, 11 Degree

50 I —o— Shd Off, S-Blade, 11 Degree

 

 

40 I—

AP (Pa)

 

 

 

0 I_ r__ i_______I___ _ . _.____._.-- _ , , ‘ I _
1.50 2.00 2.50 3.00 3.50

Q (m3/s)

FIGURE 6.16 Performance Data for the 4-Blade and 5-Blade Fans

83

 

 

 

 

 
  

 

 

 

 

 

70 f —’-— 7— w;

00 i I Legend
i I + Shd Off, 4-Blade, 11 Degree

50 L 1+ Shd Off, S-Blade, 11 Degree

2? , .

m 40 E

v

9.1

Q 30 m

20

 

 

 

FIGURE 6.17 Efficiency Data for the 4-Blade and S-Blade Fans

Performance Data for S-Blade "Shroud On" Condition
I Legend

+ Shd Off, 4-Blade, 7 Degree Cone

—-t- Shd 249 Pa, S-Blade, 11 Degree

+ Shd 373 Pa, S-Blade, 11 Degree

 

  
 
 
 

 

 

 

AP (Pa)

        

 

 

0' 1 I 1 I 1 I I I 1
1.00 1.50 2.00 2.50 3.011 3.50 4.00 4.50

Q (m3/s)

   

FIGURE 6.18 Performance Curves for 5-Blade Fan "Shroud On" Condition vs. Baseline
Fan (4-Blade)

84

 

 

     
   
 

+ Shd Off, 4-Blade, 7 Degree Cone
-*— Shd 249 Pa, 5-Blade, 11 Degree
—n—— Shd 373 Pa, S-Blade, 11 Degree

 

 

 

40L

AP (Pa)

30 *—

 

20

 

 

 

2.0 4.0 (1.0 8.0 l().() 12.0 14.0 [6.0

 

FIGURE 6.19 Efficiency Curves for S-Blade Fan "Shroud On" Condition vs. Baseline
Fan (4-Blade)

Comparison of Performance, Shd Off: 4-Bl vs. 5-Bl vs. 3-Bl vs. 8-Bl
70

 

L

()0

IfIrIITTIaj
|

 

AP (Pa)

Legend

+ Shd Off, 4-Blade, 7 Degree
+ Shd Off, 5-Blade, l 1 Degree

+ Shd Off Pa, 3-Blade, 1 1 Degree
+ Shd Off Pa, 8-Blade, 11 Degree

 

J
1 I I I 1

1.50 2.00 2.50 3.00 3.50

Q (m3/s)

 

 

 

[j—T I I T—TT—r‘rf‘gifiTI

u—
f».
V
3

FIGURE 6.20 Performance Curves for 4-Blade fan ("Shroud Off") vs. S-Blade ("Shroud
Off") vs. 3-Blade ("Shroud Off") vs. 8-Blade ("Shroud Off")

85

Comparison of Efficiency, Shd Off: 4-Bl vs. S-Bl vs. 3-Bl vs. 8—Bl

 

 

 
   
    

 

 

_ I Legend
60 I- x
I I + Shd Off, S—Blade, 11 Degree
50 F I+ Shd Off, 4-Blade, 7 Degree
1‘ + Shd Off, 8-Blade, 11 Degree
’6? 40 I_ I—t— Shd Off, 3-Blade, 11 Degree
5" - 7
9.;
<1 30 —

 

 

FIGURE 6.21 Efficiency Curves for 4-Blade fan ("Shroud Off") vs. S-Blade ("Shroud
Off") vs. 3-Blade ("Shroud Off") vs. 8-Blade ("Shroud Off")

Comparison of Performance: "Shroud On" (4-Bl vs. 5-Bl vs. 3-Bl vs. 8-Bl)

 

 

 

 

 

 

 

  

 

 

80 .
7. I
60
A 50 :2.
N
E:
40
On
Q C—
30 t, r
I Legend
20 —I— Shd 373 Pa, S-Blade, 11 Degree
+ Shd Off, 4-Blade, 7 Degree
10 + Shd 373 Pa, 3-Biade, II Degree
, + Shd 373 Pa, 8-Blade, I 1 Degree
0 IL 1 1 r ”7 ,,,1t::1_"f 1 ,WAJ I 1
I .00 1.50 2.00 2.50 3.00 3.50 4.00 4.50
3
Q (m /s)

FIGURE 6.22 Performance Curves for 4-Blade fan ("Shroud off") vs. S-Blade ("Shroud
On") vs. 3-Blade ("Shroud On") vs. 8-Blade ("Shroud On")

86

AP (Pa)

FIGURE 6.23 Efficiency Curves for 4-Blade fan ("Shroud off") vs. S-Blade ("Shroud

AP (Pa)

FIGURE 6.24 Comparison of Performance for 5-Blade with and without Installed
Centerbody at the "Shroud Off" Condition. The Baseline Production Case (4-Blade,

80

70

()0

50

40

30

20

IO

0

70

 

 

 

 

 

 

 

 

(10 ‘

50

40

30

20

k .

I I
TI
II
I? _

 

Legend
I _..._ Shd Off. 4-Blade. 7 Degree
: _fi._. Shd 373. 8-Blade, ll Degree
r
I: _-__ Shd 373 Pa. S-Blade, 11 Degree
; \ + Shd 373. 3-Blade, 11 Degree
f \
I...
I— \
C
L \
: \ \ ,
F 1 I 1 I 1 I 1 I 1 I 1
4.0 0.0 8.0 10.0 12.0 14.0 16.0
*
'1

On") vs. 3-Blade ("Shroud On") vs. 8-Blade ("Shroud On")

4-Bl vs. S-Bl w/o Centerbody vs. S-Bl w/Centerbody

 

  

 

   

Legend
+ Shd Off, 4-Bl, 7 Deg

+ Shd Off, 5-Bl, 11 Deg w/O CB I
+ Shd Off, S-Bl. 11 Deg w/CB I

 

 

A _ I . _
2,00 2.50

Q (m3/s)

1.50 3.50

"Shroud Off") is shown as a reference

87

 

 

4-Bl vs. S-Bl w/o Centerbody vs. 5-Bl w/Centerbody

 

 

 

      
  

 

 

 

 

 

 

 

7o _
()0 f : Legend
; I + Shd Off, 4131, 7 Deg
50 f’ + Shd Off, S-Bl, 11 Deg w/O CB
A ;; + Shd Off, 5431. 11 Deg w/CB
a -.
9-1 40} -,________.__
V .
9‘ ,,
<1 30}
20 17
I'
10 j—
I
OIIILIHIHII .
2 4 o 8 111 12 14
11*

FIGURE 6.25 Comparison of Efficiency for S-Blade with and without Installed
Centerbody at the "Shroud Off" Condition. The Baseline Production Case (4-Blade,
"Shroud Off") is shown as a reference

4-Bl vs. S-Bl w/o Centerbody vs. 5—Bl w/Centerbody

' ___-_. I

 

 

 

so I

70

Legend

1,..-

  
  

 

 

  

‘ —a— Shd 0024-81.? Deg
_ _ + Shd 373 Pa. 5-131, 11 Deg Wm (‘B
60 g I + Shd 373 Pa. 5-131, 11 Deg w/CB
A 50 I:
a ._
e: ~
a... 40 :’
<1
30
[,
20 ~-
10
0 1.— I __ I I I I

 

 

     

 

_ _. .A.-..

 

1.00 1.50 2.00 2.50 3.1111 3.51) 4.00 4.50

Q (m 3/s)
FIGURE 6.26 Comparison of Performance for 5-Blade with and without Installed

Centerbody at the "Shroud On" Condition. The Baseline Production Case (4-Blade,
"Shroud Off") is shown as a reference

88

 

 

4-Bl vs. S-Bl w/o Centerbody vs. 5-Bl w/Centerbody

 

 

 

 

 

 

 

 

80 I"

70 L— I Legend

( 0 f“ e ,, I —a— Shd Off, 4-31, 7 Deg

’ - I + Shd 373 Pa. S-Bl. 11 Deg w/o (B

Z _ + Shd 373 Pa, 5-131, 11 Deg w/CB

A 50 W L—--—-
N I
E: -

40 -
m L
<1 :

30 a

10 I:—

40 0.0 8 0 10.0 12.0 14.0 16.0

11*

FIGURE 6.27 Comparison of Efficiency for 5-Blade with and without Installed
Centerbody at the "Shroud On" Condition. The Baseline Production Case (4-Blade,
"Shroud Off") is shown as a reference

5-Blade Fan at Various Shroud Pressures

 

j

 
 
 
  
 
  

I Legend
I —l— Shd Off, 4-Bl, 7 Deg Cone

()0 ' + Shd 249 Pa, 5131 w/CB

 

 

 

 

 

 

+ Shd 373 Pa, S-Bl w/CB

.6? 50 __ + Shd 498 Pa, 5-131, w/CB
a; ._
m 40 “'—
< .

30 '

20

IO

0 . , 1

1.50 2.00 2.50 3.00 3.50 4.00 4.50
3
Q (m /s)

FIGURE 6.28 Performance Data for the S-Blade Fan with Installed Centerbody at
Various Shroud Pressures

89

L
e;

5-Blade Fan at Various Shroud Pressures
80 r - * A,

 

 

  
  
  
  
   
  

I
I + Shd Off, S-Bl w/CB

‘ + Shd 249 Pa, S-Bl w/CB
I + Shd 373 Pa. S-BI w/CB
I + Shd 498 Pa, 5-Bl, w/CB

P, f. —

 

Legend

 

AP(Pa)
*‘T‘Ififi? *Trr—

 

I

0 _ ._._ I I ' .____ _._I__

4.0 (1.0 8.0 10.0 12.0 14.0 l6.0

Q (m 3/s)

 

FIGURE 6.29 Efficiency Data for the S-Blade Fan with Installed Centerbody at Various
Shroud Pressures

9O

 

7.0 Results and Discussion - Velocity Measurements

7.1 PIV Measurements on the Production 4-Blade Fan Configuration

The initial set of PIV measurements was collected on the 4-blade production fan assembly
in three separate regions, as previous defined by Figure 5.2. The choice of these three
regions was based on observations made using flow visualization techniques such as tuft-
ing and smoke visualization. These flow visualization studies were carried out concur-
rently with the integral measurements and they provide insight on the flow field in this fan
system. Referring to Figure 5 .2, Region I was selected so the exit velocity profile of the
diffuser cone could be evaluated. Data collected in this region provide information about

the exit velocity profile (i.e. larger or smaller OLC values in equation 2.20), including the

existence of reversed flow. Additionally, these data can yield information about flow sepa-
ration near the cone wall at the exit plane. Region II was selected to determine if the flow
is attached to the sidewall of the diffuser cone at a position upstream of the exit plane.
Finally, Region III was selected to measure the effect of the (required) motor assembly.
Specifically, in Region III, it was speculated that the interior blade regions of the 4-blade
propeller were ineffective at imparting a positive velocity (in the direction of the cone exit
plane) because these regions are located in the wake region of the motor. This would indi-
cate that the hub was undersized for this application. The results of each of these three
regions are presented in the following sections. These plots are primarily shown as color

contours of the z—component of velocity with streamlines added to show the path a particle

would follow if the flow field were constant with time. These streamlinesl are found as the

 

1. Streamlines, as calculated in the PIV data, are defined by the differential equation vrdz = vldr

91

 

contour curves of the stream function, which is calculated using a predictor-corrector inte-

gration algorithm.

7.1.1 Near Hub Region (4-B1ade Fan with 11 Degree Diffuser Cone)

Measurements collected in Region 111 from Figure 5.2 were obtained by cutting a small
observation window in the sidewall of the 11 degree diffuser cone for the camera and by
shining the laser sheet from the upper receiver of the AFRD, pointing in the upstream
direction. The laser was aligned such that it cut the center of the fan’s rotation and the cor-
responding field of view from the camera was approximately 200 mm wide by 200 mm
tall. Smoke particles were introduced from an upstream reservoir by using a Rosco Model
No. 1600 theatrical fogging machine. The Rosco machine provides particles that range in
size from 0.25 pm - 1.0 pm in diameter. The data presented in this section are all presented

as time-averaged (i.e. an ensemble of 1000 random blade position images) contour plots.

7.1.1.1 Near Hub Region - "Shroud Off" Condition

Figure 7.2 shows the time-averaged velocity contour plot of Region III. The center of the
axis of rotation for this fan is located at (0,-594.4) with z = -594.4 mm representing the
downstream edge of the motor shaft. The approximate size and locations of the 4—blade
hub and the upstream motor are shown at the top of the figure. These data show that there
is a strong recirculating region downstream of the motor, with a reversed flow that has a
peak z-component of nearly - 2 m/s. Further downstream, there appears to be a stagnation
point at approximately r = 0 mm and z = -480.0 mm. The region adjacent to the recirculat-
ing region is characterized by a very low positive velocity, ranging from nearly 0 m/s to a

peak of z 1.5 m/s. At larger radii, the velocity continually increases, with a peak velocity

92

 

 

of nearly 4.0 m/s for this measurement volume. Figure 7.2 shows that this region of lower
speed fluid (0 < v s 1.5 m/s) that lies just outside of the recirculation area appears to be
converging towards the center of the cone in the downstream direction, but the higher
speed fluid (v 2. 2.0 m/s) appears to be diverging towards the outer edge of the cone. This
is most likely a result of entrainment by the higher speed flow as it follows the diverging
angle of the diffuser cone. It should be noted that the region of flow in the upper left-hand
corner that appears to be moving in the horizontal direction (per the streamlines that are
shown in this region) is actually the fan blade that passed through the image plane and cor-
related on itself. Because of the direction of the blade, this only appeared in the left-hand

side.

7.1.1.2 Near Hub Region - Low-Pressure Shroud Condition (APshroud = 498 Pa)
The region directly downstream of the hub was subsequently examined in the low-pres-

sure shroud condition (Apshroud = 498 Pa). This is the identical condition to the one that

was measured for numerous "shroud on" conditions in Section 6.0. Although the aerody-
namic shroud actively controls the outer blade region, it could potentially affect this hub
region by altering the radial velocity profile directly downstream of the blade. The contour
plot of this shroud condition is shown in Figure 7.3, again with color contours of the 2-
component and the corresponding streamlines. This figure shows that the downstream hub
region is essentially the same as in the "shroud off' condition in both qualitative flow fea-

tures and magnitudes of velocity.

7.1.1.3 Near Hub Region - High-Pressure Shroud Condition (APshroud = 996 Pa)

93

 

A second shroud condition, termed the "high-pressure shroud" condition and correspond-

ing to a APshroud = 996 Pa, was also measured for the near hub region. This represents the

highest condition that was tested in the integral measurements of Section 6.0 and also the
condition which provide the most significant improvement in flow rate for both the 4-
blade and S-blade propellers. Figure 7.4 shows the contour plot of the z-component of
velocity and the corresponding streamlines. These results are similar to those of Figure 7.3

for the low-pressure shroud condition, with nearly identical qualitative flow features and

velocity magnitudes. One notable difference is that the low velocity region directly down
stream of the recirculation area appears to be biased to the left. This could be a result of
the high entrainment rate of the wall jet of the aerodynamic shroud. That is, it is suggested
that this region is being "pulled" along. The streamlines are most likely tending to the lefi
because the fan’s direction of rotation is counter-clockwise when the viewer is looking in
the upstream direction. From the perspective of the camera used for these measurements,
the projection of this rotation would point towards the lefi. It is quite possible that this
downstream flow pattern has a slight counter-clockwise swirl component. This 3-D swirl

component is being measured as the projection onto the image plane of the CCD camera.

7.1.2 Diffuser Wall Interior Region (4-Blade Fan with 11 Degree Diffuser Cone)

Measurements were made in the interior of the cone for the 4-blade fan with the 11 degree
diffuser cone to determine if there was attached flow at a region that was upstream of the
exit plane of the diffuser cone. Initial flow visualizations conducted during the integral
measurements indicated that for the 4-b1ade and 11 degree diffuser cone combination, the
flow remained attached along the entire length of the cone. However the results of

Figures 6.6 and 6.7 clearly show that there is neither an improvement in performance nor

94

in efficiency by adding the 11 degree cone, despite its significantly larger exit area. These
PIV measurements were collected with the intent of measuring the flow field along the
interior wall of the 11 degree diffuser cone for the "shroud off“ condition, along with two

different "shroud on" conditions.

Optical access was achieved by placing the camera inside the cone and pointing the laser
sheet upwards, as shown in Figure 7.5. This creates an unusual perspective since the laser
sheet actually “cuts” through the cone at the perimeter of the images. The results are
shown in Figures 7.6, 7.7, and 7.8. These figures are shown with a local coordinate system
rather than the global coordinate system used in the PIV measurements previously pre-
sented and also the rest of the PIV measurements in this section; the x' and y'—component
values that are shown in the plots are independent of the r, 0, z coordinate system that is
used in the remainder of the data. The y-component is aligned with the surface of the cone
wall. As a result, the negative y-direction in the these figures is aligned with the stream-
wise direction of the flow. Similarly, a negative z-component of velocity would represent a
flow moving towards the exit plane of the diffuser cone and a positive velocity would rep-
resent reversed flow, which would most likely be a result of flow separation. Figure 7.6
shows the "shroud oft“ condition for the 4-blade fan with the 11 degree cone. At this con-
dition, the peak value for the z-component of velocity, found in the center of the image
plane, is approximately -6.00 m/s. This peak velocity decreases in both the positive and
negative x-directions as the laser sheet is positioned closer to the surface of the diffuser
cone. At the outer most regions, the velocity approaches zero. From Figure 7.6, it is appar-
ent that there is no reversed flow (positive z-component velocity) evident across this sec-

tion of the cone wall. The region of peak velocity increases from the top of the image to

95

 

 

 

 

the bottom, which corresponds to the local radius of curvature increasing. Finally, it is
notable that the streamlines follow a line that is pointing towards the bottom left of the
image. This is evidence of a slight swirl component in the interior wall of the diffuser,

which was also observed in the flow visualization measurements.

Figure 7.7 shows the same physical location as in Figure 7.6, only the fan system is oper-

ated in the "shroud on" condition, with a APshmud = 498 Pa. This figure shows that the

peak velocity, located at the center of the image, has been increased to approximately -7.5
m/s. This is a result of the increased momentum that is imparted by the wall jet from the
aerodynamic shroud. Similar to the results of Figure 7.6, this peak velocity decreases in
both the positive and negative x-directions, finally resulting in a near zero velocity close to

the wall. There also appears to be a slight swirl similar to that observed in Figure 7.6.

The final shroud condition that was evaluated was for APshmud = 996 Pa. These results are

shown in Figure 7.8 and, as anticipated, the peak velocity has increased further, to a peak
velocity of approximately -8.5 m/s. The overall results of this shroud condition are quali-

tatively very similar to those of both the "shroud off" and APShroud = 498 Pa conditions.

These results show that there is no evidence of flow separation in the 11 degree diffuser
cone when used in conjunction with the 4-blade fan. Given this observation, it is reason-
able that there would be an increase in either the performance or the velocity of this con-
figuration compared with the baseline production assembly (4-blade fan with 7 degree
diffuser cone). However, the results of Figures 6.6 and 6.7 show that this is not the case.
Additionally, it can be seen from these results that the aerodynamic shroud is capable of

increasing the momentum flux adjacent to the wall of a diffiiser cone. As expected, this

96

' L

 

 

momentum flux increased with increasing shroud pressure. Finally there is a swirl compo-
nent observed in the region adjacent to the wall. Since the initial jet from the aerodynamic
shroud has only an axial component, this observation suggests that the wall jet is slightly

altered by the swirl component of the fan.

7.1.3 Exit Plane of the Diffuser Cone (7 Degree Cone)

A set of measurements were made near the exit plane of the 7 degree diffuser cone to
resolve the exit velocity profile and additional features of the flow field near the exit of the
control volume, as described in Figure 2.1. The setup used for these measurements is
shown in Figure 7.9. This figure shows that the laser sheet was tilted to a 6° angle to main-
tain a perpendicular relationship with the CCD camera, which had to be tilted up 6° to
gain optical access to the centerline of the cone. The region of interest was from the cen-
terline of the cone, (0,0) in Figure 5.2, to the outer wall of the cone, which is a distance of
381 mm for the 7 degree cone (the 11 degree cone, as described later, has a diameter of
422.9 mm). The experimental configuration, as shown in Figure 7.9, allowed for a field of
view that was approximately 185 mm wide; therefore three different radial views were uti-
lized, with a large overlapping region in the middle. This allowed for the entire radial span
to be interrogated and the overlapping areas also provided a redundant check on the sepa-
rate data sets. These three radial measurement regions are referred to as the "center
region" "overlap region", and "near wall region", respectively from the center of the cone

3

moving to the outermost radius.

Since the shroud and housing for this fan is axisymmetric, only one azimuthal position

was measured. The operating conditions for all data in this section can be described as a

97

pressure rise: AP = 25 Pa and three different shroud pressure conditions: i) APshroud = 0 Pa
("shroud off“), ii) APshmud = 498 Pa and iii) 996 Pa. These values represent the range of

shroud conditions that were evaluated in the integral measurements of Section 6.0. The
time-averaged data are presented as plots that are shown as color contours of the z-compo-
nent of velocity with added streamlines of the time-averaged flow field. The trailing edge
of the diffuser cone is drawn as a solid black line to give the approximate location of the

exit plane of the control volume as shown in Figure 2.1.

7.1.3.1 Center Region of the Diffuser Cone Exit Plane (7 Degree Cone)

The center region is shown schematically in Figure 7.10. The field-of—view for this region
was approximately 185 mm by 185 mm, and extended from the centerline of the fan exit
plane at r = O outward to a radial position, r = 185 mm. The image plane was positioned
such that the exit plane of the diffuser cone was at the top of the image, this exit plane cor-

responds to z = 0 mm.

Figure 7.11 shows the center region results for the 4-b1ade fan with the 7 degree diffuser
with the "shroud off' condition. The dark black line across the top of the figure (that lies
along the z = 0 mm line) designates the location of the exit plane of the diffuser cone. Sev-
eral interesting observations are evident from this figure. First, there is a region with a
negative z-component in the upper right-hand portion of the figure. This region of
reversed flow extends from the centerline at r = 0 mm and continues out to approximately
r = 60.0 mm. This is portion is problematic since it reduces the overall flow rate of the sys-
tem. The second interesting observation is the region of nearly zero velocity just outside

the region of reversed flow. This region extends from r = 60.0 mm outward to r = 175.0

98

 

p,

 

mm. The flow in this region is positive, yet it has a very low velocity (less than 1 m/s).
Finally, at the upper left hand comer of the image, a region of higher velocity is shown.
The flow in this region has a velocity of approximately 1.5 m/s. Therefore, three distinct
regions are shown: an inner region with negative z-component velocities (r = 0 - 60.0
mm), a larger middle region of small (less than 1 m/s) positive z-component velocities (r =

60.0 -175.0 mm) and an outer region of higher (greater than 1 m/s) positive velocity flow.

This center region was then measured in the "shroud on" condition to determine the effects
of the aerodynamic shroud on this region of the flow. The center region is quite far from
the outer edge of the diffuser cone wall, but it was reasoned that the aerodynamic shroud
might alter the velocity across the entire span of the diffuser cone exit plane. The results of

the aerodynamic shroud in the low-pressure shroud condition, Apshroud = 498 Pa are

shown in Figure 7.12. The general features of this region appear to be similar to those of
the "shroud off" condition, however the region of reversed flow in the center has
increased. At this condition the reversed flow region spans from approximately r = 0 mm
out to approximately r = 105 mm. Adjacent to this reversed flow region is a region of
small magnitude (less than 1 m/s) positive z-component velocity and a slightly larger outer
region of higher magnitude (greater than 1 m/s) positive velocity flow. Overall, the gen-
eral features of the flow field are similar, but the presence of the "shroud on" condition has
widened the region of reversed flow and slightly increased the size of the higher magni-

tude region.

A second "shroud on" condition was also examined, and the results are shown in

Figure 7.13. This second condition had a APShmud = 996 Pa. These results show that there

99

 

again exists a region of reversed flow that is larger than the shroud off condition. The
magnitudes of this region appear to be slightly larger than both the "shroud off' and low-

pressure shroud (APshroud = 996 Pa) cases.

7.1.3.2 Near Wall Region of the Diffuser Cone Exit Plane (7 Degree Cone)

The near wall region is defined by the schematic of Figure 7.14. This region had a field of
View comparable to that of the center region in Section 7.1.3.1 that measured approxi—
mately 185 mm by 185 mm. This region extended just beyond the side wall of the diffuser

cone.

Figure 7.15 shows the results for the 4-blade fan with the 7 degree cone when operated in
the "shroud off" condition. The location of the exit plane of the diffuser cone, along with
the side of the cone wall are shown by the dark black line. Streamlines are placed to show
the approximate flow paths. This figure shows that there is a region of high velocity (>
4m/s) near the outer edge of the flow (approximately between r = 280 mm and r = 380
mm), that decreases steadily towards the center of the cone. The peak magnitude in this
region is approximately 4.8 m/s. There is no evidence of reversed flow nor any negative 2-
component flow. The large region of nearly zero velocity that exists near the cone wall is
the result of the diffuser cone being visible in the measurement field-of—view. The glare
from the laser contacting this surface created a small region adjacent to the solid surface
where very few correlations from the PIV image pairs exist. These "washed out" areas are

indicated by the ~0 m/s velocity contours.

Next the "shroud on" condition was evaluated with a A'Pshroud = 498 Pa and these results

are shown in Figure 7.16. This figure shows the aerodynamic shroud is able to increase the

100

 

 

velocity in this region; the aforementioned high velocity region of the "shroud off" condi-
tion has increased and the peak z—component velocity in this region now exceeds 5.0 m/s.

A second "shroud on" condition of APshmud = 496 Pa was measured and the results are in

Figure 7.17. This figure shows that the high velocity region is further increased at this
higher shroud pressure. For this case the high velocity region is increased to well over 6.0
m/s. Both of these "shroud on" conditions show that the aerodynamic shroud is capable of

increasing the velocity in the near wall region of the diffuser cone.

7.1.3.3 Overlap Region of the Diffuser Cone Exit Plane (7 Degree Cone)

Limitations in the field-of—view of the CCD camera resulted in a measurement region that
was approximately 185 m2. Therefore three measurement regions were required in order
to measure the entire exit plane of the diffuser cone from the center (0,0) to the outer edge
of the diffuser cone (0, 381.0 mm). The third region is referred to as the "overlap" region
and it covers the portion not measured by either the "center region" nor the "near wall
region"; and it also overlaps into each of these regions. The overlap region was selected to
provide a redundant check on a portion of the measurement planes of the center and near

wall regions. The overlap region is shown schematically in Figure 7.18.

The "shroud off" results of the 4-blade fan with the 7 degree diffuser cone are shown in

Figure 7.19. As expected, this measurement location shows the transition from the higher
speed flow of the near wall region to the low velocity of the center region. The location of
the diffuser cone exit plane is again shown by the solid dark line that exists at z -= 0 mm.

The subsequent results of the "low-pressure shroud" condition of APshroud = 496 Pa and

the "high-pressure shroud" condition of APshroud = 998 Pa are shown in Figures 7.20

101

 

 

and 7.21, respectively. Again the general trends are the same, but the magnitude of the
peak velocity component increases for each condition. This is the result of the high speed

jet that is produced from the aerodynamic shroud.

7.1.3.4 Composite View of the Diffuser Cone Exit Plane (7 Degree Cone)

The composite view of each of the three measurement regions: i) center region, ii) near
wall region, and iii) overlap region is shown for the "shroud off“ casein Figure 7.22. This
composite view provides a comprehensive look at the exit plane flow field of the diffuser
cone. There is slight shift in the upper and lower limits of the z-component positioning 'of
each of the three views, this is a result of having to re-position the camera for each of these
measurement sets, which resulted in slight shifts in orientation and also slight changes in
the field-of-view for each of these images. The location of each image plane was checked
prior to each measurement set, so the resulting locations are accurate and reliable. The
slight vertical shifts in the measurement planes of Figure 7.22 show the agreement
between these independent measurement sets. This is particularly evident at the high
velocity region that exists in the "near wall region" and the right hand side portion of the

overlap region.

Figures 7.23 and 7.24 show the composite views for the two "shroud on" conditions of
APShmud = 496 Pa and APShmud = 998 Pa, respectively. As was shown in the results of
Figure 7.22, it is evident that there is good agreement between the overlapping portions.
Also, the trend of increased flow in the larger radii (approximately from r = 250 mm to r =
360 mm) for the "shroud on" conditions can be seen. There also exists a slightly higher

reversed flow for the "high-pressure shroud" condition than for the "low-pressure shroud"

102

 

condition. A slightly higher reversed flow is also seen for the "low—pressure shroud" con-

dition compared to the "shroud off" condition.

7.1.4 Exit Plane of the Diffuser Cone for the 4-Blade Fan with 11 Degree Cone

The effects of changing the diffuser cone were examined by replacing the 7 degree dif-
fuser with the 11 degree diffuser and repeating the measurements of Section 7.1.3.
Although the radius of the exit plane of the 11 degree diffuser (r = 423.0 mm) is consider-
ably larger than that of the 7 degree diffuser (r = 381.0 mm), the entire span could still be
measured by three measurement planes. Therefore, the same terminology will be used to
describe the three measurement regions (center region, near wall region and overlap
region). These approximately represent the same relative locations as those used in for the

7 degree diffuser.

7.1.4.1 Center Region of the Diffuser Cone Exit Plane (11 Degree Cone)

The center region of the 11 degree diffuser covered a region that was very similar to that
of the 7 degree cone, and a visual description of this area can again be found in the sche-
matic drawing of Figure 7.10. The field-of—view was very close, but not identical to, that
of the center region measured for the 4-b1ade fan with the 7 degree diffuser. Figure 7.25
shows the results of the center region of the 4-blade fan with the 11 degree cone for the
"shroud off" condition. This figure shows that there exists a strong region of reversed flow
in the center of the cone. Whereas the velocity magnitudes in the reversed flow portions
for the 7 degree cone were small (< 0.5m/s), this configuration has a much higher magni—
tude. The peak velocity in this region of reversed flow is nearly - 1.6 W3, and this reversed

flow region extends farther downstream than the reversed flow region of the 7 degree cone

103

 

configuration. The left side of Figure 7.25 shows a region of positive velocity, with a peak
velocity slightly greater than 2.0 m/s. Between this inner region of reversed flow and the
outer region of positive velocity flow, there exists a large swirling region. This region
spins in a clockwise direction and is driven by the two boundaries of opposite direction
flow. The regions of reversed flow and the swirling region represent areas of the flow field
that do not make positive contributions to the overall performance and efficiency of this

fan system.

Figure 7.26 shows this same region in the "shroud on" condition, specifically for the case

of APshmud = 498 Pa. These results show that with the aerodynamic shroud activated, the

region of reversed flow has increased in size, extending out to a radial position of approx-
imately r = 150 mm. The magnitude of this reversed flow has remained the same as that of
the "shroud off“ case. The large region of swirling flow in the center of the image is no
longer present, and the region of positive direction flow has clearly shifted to a higher
radius. This is most likely a result of the increased entrainment from the aerodynamic

shroud wall jet in the cone wall region.

The results for the a higher "shroud on" condition. APShroud = 996 Pa, are shown in
Figure 7.27. This figure shows that for this higher "shroud on" condition, the radial span
of reversed flow has increased even farther beyond the reversed flow region that for the
low-pressure shroud case (APshmud = 498 Pa). This region of reversed flow extends from
the center of the cone (r = 0 mm) out to around r = 140 mm. Again, the large region of
swirling flow in the center of the image does not exist and the positive direction flow has

shifted outward to a higher radius. It appears that the effect of increased entrainment from

104

 

the wall jet of the aerodynamic shroud is even more pronounced than the observed effect

from the lower "shroud on" case (APShmud = 498 Pa).

7.1.4.2 Near Wall Region of the Diffuser Cone Exit Plane (11 Degree Cone)

The near wall region, as shown schematically in Figure 7.14, was measured for the same

three shroud conditions: "shroud off", "low-pressure shroud" (APShmud = 498 Pa) and
"high-pressure shroud" (APshroud = 996 Pa). Figure 7.28 shows the results for the "shroud

off" condition. The edges of the diffuser cone sidewall and exit plane are marked in this
figure by the solid black line. From this figure, there is a region of high speed flow (V >
6.0 m/s) that exists away from the wall of the diffuser cone, approximately between the
radial positions of r = 280.0 mm and r = 370.0 mm. The outer edge of the cone is at r =
423.0 mm. In this figure, the region between r = 370.0 mm and r =423.0 mm has a much

lower velocity.

Figure 7.29 contains the data for the same region in "low-pressure shroud" condition

(APshroud = 498 Pa). Again, similar to the results of Figure 7.28, there exists a region of

high speed fluid (V > 6.0 m/s) near the outer cone wall. However, for these results, the
radial position of this region has shifted to a larger radial position; it now occupying the
region between r = 315.0 mm and r = 395.0 mm. Since the percentage of total exit plane
area represented by this region is much higher, this would represent an increase in total

volume flow rate. The results for the "high-pressure shroud" condition (APshroud = 996 Pa)

are shown in Figure 7.30. Similar to the results of "low-pressure shroud" condition, the
region of high speed fluid (V > 6.0 m/s) has shifted towards the outer cone wall, and is

occupying a region that is nominally identical to that of the "low-pressure shroud" condi-

105

 

"1‘

tion. The primary difference at this condition is that the peak velocity within this region of
high speed fluid has increased slightly, to a peak z-eomponent of velocity of slightly
higher than 7.0 m/s. This would most likely indicate a higher volume flow rate than both
the "low-pressure shroud" and "shroud off" conditions, but not nearly as dramatic as the

change from the "shroud off“ to the "low-pressure shroud" case.

7.1.4.3 Overlap Region of the Diffuser Cone Exit Plane (11 Degree Cone)

The overlap region for the 4-b1ade fan with the 11 degree cone is again shown schemati-
cally in Figure 7.18. Given the large exit area of the l 1 degree cone, there was slightly less
redundant information between this region and the two adjacent regions (near wall region
and center region). The results for this region with the "shroud off" condition are shown in
Figure 7.31. This figure shows that there is no reversed flow (negative z-component of
velocity) across the entire radial span, and also that the velocity increases with increased

radius.

Figure 7.32 shows the same location for the "low-pressure shroud" condition (APshroud =

498 Pa). The surprising result here is that the low-pressure shroud condition shows a
region of reversed flow (negative z-component of velocity) at the right edge of the figure.
Additionally, the region of high speed fluid (V > 5.0 m/s) has shifted to a higher radial
position. This appears to be the result of increased entrainment from the wall jet of the
aerodynamic shroud. Referring to Figures 6.10 and 6.1 1, this condition was able to

increase the volume flow rate, Q, but it was accompanied by a decrease in efficiency.

Figure 7.33 shows the results for the "high-pressure shroud" condition (APshroud = 996

Pa). These results are very similar to those of Figure 7.32, once again show that there is

106

reversed flow that did not exist for the "shroud off' case and also the location of the high
speed fluid (V>5.0 m/s) has shifted radially outward. The peak magnitude of this region of
high speed fluid has also increased, which is most likely caused by the increased speed of

the wall jet at this "high-pressure shroud" condition.

7.1.4.4 Composite View of the Diffuser Cone Exit Plane (1 1 Degree Cone)

By combining the three measurement regions: i) center region, ii) near wall region, and iii)
overlap region, a composite view of the exit plane velocity was created. The three shroud
conditions, "shroud off", "low-pressure shroud", and "high-pressure shroud" are shown in
Figures 7.34, 7.35, and 7.36, respectively. These composite views clearly show the two
key observations that were seen in the individual measurement regions, namely the out-
ward shifi in the high speed fluid (V > 5.0 m/s) for the two "shroud on" conditions and the
subsequent increase in reversed flow at the center of the cone for the two "shroud on" con-

ditions.

7.2 PIV Measurements on the 5-Blade Fan Configuration

The results of Section 6.4 on page 65 clearly showed that the 5—B1ade fan, when tested
with the aerodynamic shroud in the "shroud on" condition, could provide substantial
increases in both volume flow rate, Q, and efficiency, 11*. While these gains were appar-
ent, the mechanisms that caused this dramatic increase was not. Because of these improve-
ments, there was considerable interest in using PIV at the exit plane of this diffuser cone.
It was thought that information obtained from these measurements could provide an
insight into the specific reasons why the aerodynamic shroud worked so well with this

configuration, but not nearly as well with the previous configurations. Data collected in

107

 

Section 6.0 indicated that the optimal "shroud on" condition for this configuration was

APShroud = 373 Pa. This was selected as optimal because it provide substantial increases in
both the flow rate and the efficiency, 33.6% and 12.3% respectively, when examined at the
operating condition of AP = 25 Pa. Because of these findings, only one "shroud on" condi-

tion of APshmud = 373 Pa was utilized for the following PIV data. Additionally, two Oper-

 

ating points were examined for these data, to determine the effect of the system pressure

"a

rise on the fan exit plane in both the "shroud off' and "shroud on" conditions. These two

operating points were AP = 25 Pa and AP = 50 Pa.

IT
1

7.2.1 Exit Plane of the Diffuser Cone (11 Degree Cone)

Similar to the results of both Section 7.1.3 and Section 7.1.4, the exit plane of the 11
degree diffuser cone was measured in three separate regions: i) center region, ii) near wall
region, and iii) overlap region. These regions are nominally identical to those used in

Section 7.1.4 for the 4-b1ade fan with the 11 degree diffuser cone.

7.2.1.1 Center Region of the Diffuser Cone Exit Plane (11 Degree Cone)

The center region is defined schematically in Figure 7.10 for the 5-blade fan and the 11
degree cone. Figure 7.37 shows this region for the "shroud off" condition at a pressure rise
of AP = 25 Pa. The exit plane of the diffuser cone is again indicated by the solid black line
at the z = 0 mm position. This figure shows that the velocity near the exit plane at this
operating point has a positive magnitude across the entire span; there is no area with a
reversed flow (negative velocity) condition. Although there is no area showing reversed
flow, the flow closest to the center of the cone has a very small magnitude. Specifically,

the velocity magnitudes in the region from r = 0 mm out to r = 60 mm are less than 1.0 m/

108

5. However, when the pressure rise of the system is increased as shown in Figure 7.38,

there is a region of reversed flow that forms near the centerline (0,0) of the exit plane.

Figure 7.39 displays the data for the center region when the "shroud on" condition

(APShmud = 373 Pa) is applied. The operating point for this data set is AP = 25 Pa. Whereas

Figure 7.37 showed that there was no reversed flow in this region, the "shroud on" condi-
tion shows that a region of reversed flow has appeared at the exit plane of the cone. These
results are consistent with the findings of Section 7.1.4.1, which showed that the "shroud
on" condition increased the region of reversed flow near the center of the 11 degree cone.
It also interesting to note that the velocity gradient across this radial section is much
smaller than for the "shroud off" case. The maximum velocity is approximately 2.66 m/s
and the minimum velocity is - 0.42 m/s. The "shroud off" condition showed a much
steeper gradient across this same section, with a peak velocity of 5.6 m/s and a minimum
velocity of approximately 0.40 m/s. From these data, it is apparent that the wall jet from

the aerodynamic shroud has made the velocity profile more uniform.

The "shroud on" condition was also evaluated at the higher back, AP = 50 Pa, for this
region. The results of this test are shown in Figure 7.40. The results here are somewhat
surprising because previous data had indicated that the higher operating point had a large
region of reversed flow near the center of the cone. Yet when the "shroud on" condition
was tested, this region disappeared, and only the exit velocity is positive across this entire
radial span. It is also interesting to note that the exit velocity is once again much more uni-
form for the "shroud on" condition. Whereas the data in Figure 7.39 had a maximum

velocity of nearly 2 m/s and a minimum of approximately -0.50 m/s at the exit plane, the

109

 

data for the "shroud on" condition show a maximum velocity of approximately 0.70 m/s
and a minimum velocity of approximately 0.10 m/s. It is apparent that the aerodynamic
shroud is capable of creating a more uniform exit velocity profile for each of these operat-

ing points.

7.2.1.2 Near Wall Region of the Diffuser Cone Exit Plane (11 Degree Cone)

Data for the near wall region, as shown schematically in Figure 7.14, were measured for

the same four conditions that were tested in Section 7.2.1.1. The results of the "shroud off“
condition at an operating point of AP = 25 Pa are shown in Figure 7.41. This figure shows
that a region of high velocity fluid (V > 5.0 m/s) exists away from the wall, approximately
from r = 250.0 mm out to r = 315.0 mm. A maximum velocity of about 6.8 m/s exists in

this high velocity area, which steadily decreases to a minimum of approximately - 0.21 m/
s. This minimum velocity occurs near the edge of the cone wall, where there is evidence of

a very small reversed flow region.

Figure 7.42 shows the same near wall region in the "shroud off' condition for a higher
operating point, AP = 50 Pa. Again there exists a region of high speed fluid (V > 5.0 m/s)
that exists adjacent to the wall, but for this operating condition the region is much larger
and is situated closer to the wall, approximately between r = 250.0 mm and r = 400.0 mm.
The peak velocity across this section is nearly 6.4 m/s and the minimum is near zero, but

overall, the velocity profile is more uniform than that of the lower operating point.

The results of the "shroud on" condition for the lower operating point, AP = 25 Pa, are
shown in Figure 7.43. There exists a region of high speed fluid (V > 5.0 m/s) that is simi-

lar to that found in Figure 7.41, only with the aerodynamic shroud in the "shroud on" con-

110

 

 

‘ a1

 

dition, this region has shifted outward, and now covers wider portion of this measurement
region, spanning from r = 250.0 mm out to r = 350.0 mm. The results of Figure 7.41 indi-
cated that there was a small region of reversed flow at the outer radial positions, but these
appear to have vanished in the "shroud on" condition. Overall, the "shroud on" condition

has caused the exit profile of this fan to be more uniform.

The "shroud on" condition was also measured for the operating point of AP = 50 Pa.
These results are shown in Figure 7.44. This figure shows a similar result to those found
for the lower operating point; namely that the "shroud on" condition shows a more uni-
form velocity profile than that for the "shroud off" condition. There exists a large region of
high speed fluid (V > 4.0 m/s) that extends across almost the entire radial span of this

region, from approximately r = 250.0 mm to almost r = 420.0 mm.

7.2.1.3 Overlap Region of the Diffuser Cone Exit Plane (11 Degree Cone)

The overlap region for the 5-blade fan with the 11 degree diffuser cone is again shown
schematically in Figure 7.18. The results for this region in the "shroud off" condition and
at the lower operating point of AP = 25 Pa are shown in Figure 7.45. This figure shows
that unlike the previous cases for the 4-blade fan, the concentration of flow with the high-
est velocity occurs in this overlap region, as opposed to the near wall region. The central
portion of this region, from approximately r = 170.0 mm to r = 250.0 mm has a z—compo-
nent of velocity that is greater than 6.5 m/s. The velocity then begins to decrease for radial

locations, r > 250.0 mm and r < 170.0 mm.

The "shroud off' data for the higher operating condition, AP = 50 Pa, is shown in

Figure 7.46. These data show that the velocity increases as the radial position increases.

111

 

 

 

 

Also, similar to the results for the lower operating point, the region of maximum velocity
occurs in this region. Specifically, its location is approximately between r = 240.0 mm and
r = 280.0 mm. The outer wall of the 11 degree diffuser is at approximately r = 423.0 mm;

therefore this region of maximum velocity is significantly far away from the wall.

The results for the "shroud on" condition (APShmud = 373 Pa) when operated at the lower

operating point of AP = 25 Pa are presented in Figure 7.47. These data show, somewhat
surprisingly, that the addition of the aerodynamic shroud has actually reduced the velocity
across the entire span of this region. The exit velocities along the z = 0 mm line range
from approximately V = 0.5 m/s at the smallest radial position shown in this region to a
maximum velocity of V = 5.2 m/s at the highest radial position shown. Figure 7.48 shows
that this trend is consistent for the "shroud on" condition for the higher operating point.
The exit velocities are also reduced at each radial position for this case when compared

with those data for the "shroud off" condition.

7.2.1.4 Composite View of the Diffuser Cone Exit Plane (11 Degree Cone)

The composite view for each of the previous cases was created by combining the three
measurement regions: i) center region, ii) near wall region, and iii) overlap region.
Although specific details of each region can be more easily understood by looking at the
figures presented in the previous sections, the composite results allow the reader to look at
the entire radial span for a general comparison of the flow fields that exist for each operat-

ing point and shroud condition.

Figures 7.49 and 7.50 show the composite views for both the higher (AP = 50 Pa) and

lower (AP = 25 Pa) back pressure operating points at the "shroud off' condition.

112

 

 

 

Figure 7.49 shows that the concentration of the highest velocity fluid is significantly far-
ther away from the wall than the highest velocity fluid for the 4-blade fan with the 11
degree diffuser cone. Also there is very little reversed flow at any radial position, except
for a small region at the outermost radial location. The innermost radial locations from
around r = 0 mm to r = 60.0 mm have exit velocities that are small (V < 1.0 m/s), but are in
the positive direction. For the higher back pressure operating point shown in Figure 7.50,
this region of high velocity fluid occurs at a slightly larger radial position. This condition
shows that reversed flow does exist at the inner most radial positions, from r = 0 mm to

around r = 60.0 mm.

The composite views for the "shroud on" condition of both the lower and higher operating
conditions are shown in Figures 7.51 and 7.52. These figures show that for each operating
point, the "shroud on" condition shifted the location of the maximum velocity outward
toward the wall of the diffuser, while simultaneously reducing the magnitude of the peak
velocity region. This creates a more uniform exit profile for each of the two operating

points.

7.3 Exit Plane Velocity Profiles and Mass Flow Rate Distributions

The previous sections that presented the results of the exit plane PIV for the 7 and 11
degree diffuser cones with both 4 and 5-blade fans have focused largely on the line along 2
= 0 mm. This line indicates the location of the downstream edge of the cone wall and rep-
resents the exit of the control volume that was originally presented in Section 2.0. Using
the PIV data from Sections 7.1.3, 7.1.4, and 7.2.1, individual data points were extracted
along the exit plane of the diffuser cone. These data are the z-components of velocity that

exist along the z = 0 mm line. A total of 300 points were extracted from the line along 2 =

113

0 mm and these points were interpolated between the measured locations. These data are
presented spatially in non-dimensional form by normalizing the radius, r, by the total
radius, R. It should be noted that these velocity data can also be represented non-dimen-

sionally by normalizing VZ by <Ve>, where <VC> = Q/Ae. These data are not presented in
this manner here since <Ve> is different for every combination of shroud condition and

diffuser cone; therefore, normalizing the velocity data makes it difficult to make relative

comparisons between the various shroud conditions.

Since the outer radial sections of the cone have the largest percentage of the total exit area,

they make large contributions to the exit flow rate. This distribution is also presented in
this section by plotting the mass flow rate, m , as a function of r/R, as defined by:

l

n} = p27tR2( V8) III
0

Vz)dr

'<—I7—>' E (7.1)

7.3.1 Exit Plane Velocity Profiles: 4-Blade Fan with 7 Degree Diffuser Cone

The extracted velocity profiles for the 4-blade fan with the 7 degree cone are shown in
Figure 7.53. This figure shows all three of the shroud conditions that were tested: "shroud

off", the "low-pressure shroud" condition (APshmud = 498 Pa), and the "high-pressure
shroud" condition (APshmud = 996 Pa). Focusing on the inner region of the cone, from r/R

= 0 out to r/R = 0.50, it is apparent that the addition of the aerodynamic shroud had very
little effect on the velocity profile. A small region of reversed flow exists from r/R = 0 out
to approximately r/R = 0.24. The velocity increases monotonically in the positive direc-

tion from r/R = 0.24 out to r/R = 0.50.

114

 

For the region at the center of the exit plane, slight increases in velocity are observed for
the two "shroud on" conditions when compared with the "shroud off" condition. This
region is approximately from r/R = 0.50 out to r/R = 0.72. It should be noted that there
appears to be a fluctuating aspect to the velocity in this region. This is actually “noise”
caused by several interrogation regions that had low numbers of validated vectors. This
was most likely caused by a small distortion on these portions of the lens of the CCD cam-
era, which is possibly caused by condensed vapor droplets or small pieces of dirt. Evi-
dence of these can be seen by referring back to Figs 7.19 - 7.21, which were described in
detail in Section 7.1.3.3 on page 101. As shown in these figures, these small regions that
had low validation numbers actually represent a small fraction of the total plot area.
Despite this, they do lie on the line of z = 0 mm. By examining the regions that surround
this area, it is apparent that the velocity in this region is increasing with increasing radius,

but it is expected that this increase would be more monotonic.

In the outermost region of the 7 degree cone, from r/R = 0.72 out to r/R = 1.0, there exists
a systematic increase in the z-component of velocity as the shroud pressure is increased.
The velocity distribution in this region increases to a maximum value and then "flattens"
until the velocity values decrease in the near wall region of the cone. That is, a relatively
thick boundary layer is observed for these conditions. This increase in velocity occurs at
the largest radial values of the diffuser, where the largest percentage of the exit area exists.
It is evident that this velocity increase accounts for the enhanced volume flow rate for both

the "low-pressure shroud" condition (APshmud = 498 Pa) and the "high-pressure shroud"

condition (APshmud = 996 Pa) when compared with the "shroud off' condition.

115

7.3.2 Exit Plane Velocity Profiles: 4-Blade Fan with 11 Degree Diffuser Cone

The velocity profiles for the 4-blade fan with the 11 degree diffuser cone are shown in
Figure 7.55. A similar set of three shroud conditions with one system pressure rise was
measured for this configuration, and the results of all three are presented in this figure.
This figure clearly shows that a region of reversed flow exists for the "shroud off" condi-
tion between the radial positions of r/R = 0 and approximately r/R = 0.26. The size of this
reversed flow region increases substantially for both of the "shroud on" conditions. That

is, VZ < 0 are observed for the radial positions from r/R = 0 to r/R = 0.37. The velocity pro-

file shows that the velocity for each of the shroud conditions increases monotonically in
the positive radial direction, reaching a maximum before it decreases in the boundary
layer region. In this region of monotonic growth, the two "shroud on" conditions show
velocity profiles that are lower than that of the "shroud off position”. The maximum exit
velocity increases with increasing shroud condition, with the "shroud off" position having

a maximum velocity, V = 6.9 m/s and the "high-pressure shroud" condition (APshmud =

996 Pa) yielding a maximum velocity, V = 7.6 m/s.

7.3.3 Exit Plane Velocity Profiles: 5-Blade Fan with 11 Degree Diffuser Cone

The 5-blade fan was of particular interest since it was capable of substantially increasing
the flow rate, while simultaneously increasing the efficiency. The previous results of the 4-
blade fan with both the 7 and 11 degree diffuser cone showed that the aerodynamic shroud
was able to increase the velocity profile in the outer regions of the exit plane of the dif-
fuser. This increase in velocity accounts for the slight increases in volume flow rate; how-
ever, these increases in velocity were accompanied by a more non-uniform velocity profile

that caused a decrease in the efficiency of the fan. These decreases were detailed in

116

‘1‘

 

 

 

Section 6.1.2 for the 4-blade fan with the 11 degree cone, where there was 14.0% increase
in volume flow rate, Q, but a 26.5% decrease in system efficiency, 11*. Also, Section 6.3.1
showed a similar trend for the 4-blade fan with the l 1 degree cone, where there was a

16.0% increase in volume flow rate, Q, accompanied by a 20.7% decrease in system effi-

ciency, 11*.

Figure 7.57 shows the extracted exit velocity profile for the 5-blade fan with the 11 degree
diffuser cone. Both of these tests were run at a system pressure rise, AP = 25 Pa, which
was the target operating point identified in the comparisons from Section 6.0. The data of
Figure 7.57 show that in the interior region of the cone there exists no reversed flow for
the "shroud off' condition, but instead a small positive velocity across the entire radial

span. When the "shroud on" condition is evaluated (APShroud = 373 Pa), a region of reverse

flow appears in the central portion of the cone, from r/R = 0 out to approximately r/R =
0.18. The "shroud off“ condition has a much higher exit velocity throughout the region
spanning r/R = 0 to r/R = 0.70 and this condition reaches a maximum velocity, V = 7.0 m/
s at a radial location, r/R = 0.51. The "shroud on" condition reaches a maximum exit
velocity, V = 5.4 m/s at a radial position, r/R = 0.65. The "shroud on" condition has a con-
sistently higher exit velocity in the outermost regions of the cone, from r/R = 0.70 mm out
to the edge of the cone at r/R = 1.0. This shows that the 5-blade fan, when operated with
the "shroud on" condition is able to provide a higher flow in the outer radial positions.
These outer radial locations represent the largest percentage of the exit area and therefore
the volume flow rate is higher. Referring back to Section 6.4.2, there was a 33.6% increase
in volume flow rate, Q, when the "shroud off“ condition is compared with the "shroud on"

condition of APShmud = 373 Pa. Additionally, the "shroud on" condition has a much more

117

uniform exit profile than the "shroud off" condition, which accounts for the subsequent
increase in efficiency. This was also shown in Section 6.4.2 to be an increase in 11* of

12.3%.

The effect of increasing the back pressure was also measured for the 5-blade fan with the
l 1 degree diffuser. These data were collected based on the observations from Figure 6.18
and Figure 6.19 which showed that the "shroud on" condition for 5-blade fan was also able
to increase the Q and 11* for higher back pressure operating points. Previous results with
the 4-blade fan had shown that even when improvements were made in the flow rate at
lower operating points, the higher operating points showed very little, if at all, increases in
flow rate. Similarly, the 4-blade fan typically showed reductions in efficiency at lower
operating points, and a much greater reduction in the efficiency at the higher back pressure
operating points. These PIV observations help to identify why the 5-blade fan was able to

function more efficiently at these higher operating points.

The extracted velocity profiles for the higher operating point, AP = 50 Pa, are shown in
Figure 7.59. These data show that in the innermost region of the diffuser cone, from r/R =
0 out to r/R = 0.25, the "shroud on" condition exhibits a higher exit velocity. This exit
velocity is much more uniform, as shown by the nearly flat profile. Also, in contrast with
the situation for the low operating point, the "shroud on" condition has a region of reverse
flow from r/R = 0 out to r/R = 0.21. The "shroud off' condition shows a higher exit veloc-
ity between the radial positions of r/R = 0.25 out to r/R = 0.95, with a maximum exit
velocity of 6.75 m/s occurring at r/R = 0.59. The "shroud on" condition has a maximum

exit velocity that is much lower, approximately 5.0 m/s, and it occurs at the same approxi-

118

mate radial location of r/R = 0.59. In the outermost region of the cone, from r/R = 0.59 out
to r/R = 1.0, the "shroud on" condition has a larger exit velocity. This higher exit velocity
in the outermost radial positions coupled with the absence of reverse flow accounts for the
higher Q that was measured for this condition in Figure 6.18. Additionally, the more uni-
form profile of the "shroud on" condition accounts for the increased 11* that is shown in

Figure 6.18.

119

 

 

 

 

 

 

 

 

 

 

Near Hub Region

FIGURE 7.1 Schematic Defining the "Near Hub Region" PIV Measurement Plane

 

   

 

 

 

- 4-Blade Hub Region " Shroud Off "Condition
: ' ::.OTOR '
-600 :— HUB
-575 :—
-550 :—
E-525 _—
“L500 —
-475 :—
-450 :—
-425 :-
150 100 50 0 -5o 1 -100
r (mm)
FIGURE 7.2 The 4—Blade Fan with 11 Deg. Cone - "Shroud Off" Condition - Near Hub
Region

120

4-Blade Hub Region AP’MM = 498 Pa

 

 

  
   

 

 

  
   

 

 

    

 

 

 

 

 

 

-600 -
-550 -
E L
.5.
N-500 -
—450 I—
100 7 I so 0 -50 A -100
r (mm)
FIGURE 7.3 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition (APshroud =
498 Pa) Near Hub Region
4-Blade Hub Region APshroud = 996 Pa
-600 I'-
, _ V1 (m/s)
- 3.62
-575 - 3.05
; mg 2.49
-550 — 1:33
A _ 0.80
E-525 l- 3::
g ' 089
N ; 1.46
-500 ,-
-475 I-
450 5
-425 I:- .:_ . .
100 50 0 ~50 -100
r (mm)
FIGURE 7.4 The 4—Blade Fan with 11 Deg. Cone - "Shroud On" Condition (Apshroud =
996 Pa) Near Hub Region

121

“'1

‘ ‘ w—W—‘v'fi‘w
3 . . w

 

 

 

 

 

 

    

CCD Camera
\Laser Sheet
Sheet Optics

 

 

 

Nd:YAG Laser
FIGURE 7.5 Setup for Interior Wall PIV Measurements

Interior Wall Region - 7 Degree Cone " Shroud Off"

    

 

 

 

0 50 150 200

100
x' (mm)
FIGURE 7.6 The 4-Blade Fan with 11 Degree Cone - "Shroud Off" Condition, Cone Wall
Interior Region

122

Interior Wall Region - 7 Degree Cone (AP,W= 498 Pa)
200 .- ‘ . fife/Si

 

 

 

 
 

O 50 150 200

100
x' (mm)
FIGURE 7.7 The 4-Blade Fan with 11 Degree Cone - "Shroud On" Condition (APshroud
= 498 Pa), Cone Wall Interior Region

 

Inten'or Wall Region - 7 Degree Cone (APmM: 996 Pa)

 

 

200 I- ' v(mls)
_ ., -0.63
. -1.51
*_ -2.39
175 g . -321
i -4.14
150 - 22:33
-6.78
.7.
A125 — ' .123
g g
yr 100 L
> .
75 -
50 -
25 -

 

0 H 50 150 200

100
x' (mm)
FIGURE 7.8 The 4-Blade Fan with 11 Degree Cone - "Shroud On" Condition (APshroud
= 996 Pa), Cone Wall Interior Region

123

 

 

 

 

 

 

 

 

 

 

 

 

  
 

 

\ 6° . Laser Sheet
Camera Sheet Optics
Nd:YAG Laser

FIGURE 7.9 Setup for Exit Plane PIV Measurements

 

+r
+x

 

 

 

(E Exit Plane of
the Diffuser Cone
Center Region

FIGURE 7.10 Schematic Defining the "Center Region" for the Exit Plane PIV

124

 

7 Degree Cone, Center Section " Shroud Off" Condition

      

 

2 (mm)

100 i 75 50 25
r (mm)

FIGURE 7.11 The 4-Blade Fan with 7 Deg. Cone - "Shroud Off" Condition Center
Region

4-Blade Fan wl7 Deg. Cone, Center Section, APshroud = 498 Pa

 

2 (mm)

 

r (mm)

125

FIGURE 7.12 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition (APshroud =
498 Pa) Center Region

4-Blade Fan WI 7 Deg. Cone, Center Section, APW = 996 Pa I

 

150 ‘100

r (mm) I
FIGURE 7.13 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition (APshroud =
996 Pa) Center Region

 

 

 

 

(E Exit Plane of

the Diffuser Cone
Near Wall Region

FIGURE 7.14 Schematic Defining the "Near Wall" for the Exit Plane PIV

126

2 (mm)

 

 

 

250
r (mm)

FIGURE 7.15 The 4-Blade Fan with 7 Deg. Cone - "Shroud Off" Condition Near Wall

  

400 3

Region

4-Blade Fan wl7 Deg. Cone, Near-Wall Section APWW,= 498 Pa

    

 

I1
I
1
I
30

127

FIGURE 7.16 The 4—Blade Fan with 7 Deg. Cone - "Shroud On" Condition (APshroud =
498 Pa) Near Wall Region

4-Blade Fan wl7 Deg. Cone, Near-Wall Section APIhmd= 996 Pa

    

20

4o

60

2 (mm)

 

400 350 300 I 250
r (mm)

FIGURE 7.17 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition (APshroud =
996 Pa) Near Wall Region

 

 

 

+X

 

CE Exit Plane of
the Diffuser Cone

Overlap Region

FIGURE 7.18 Schematic Defining the "Overlap" for the Exit Plane PIV

128

4-Blade Fan WI 7 Deg. Cone, Overlap Section, " Shroud Off " Condition

om

 

  
       
  
 
 
  
 
   

    

25

l'liTT

50

 

 

 

z(mm)
N
01
lilll'llllllili'li

 

125
250 200 150
r (mm)
FIGURE 7.19 The 4-Blade Fan with 7 Deg. Cone - "Shroud Off" Condition Overlap
Region

4-Biade wl7 Degree Cone, Overlap Section APflmd= 498 Pa

25

1 I I 1 I

01
C
1 r 1 I

 

.-
O
O
11'1111'
/

.11
N
U!

1T|1

 

200 150
r (mm)
FIGURE 7.20 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition (Apshroud =
498 Pa) Overlap Region

129

4-Biade wl7 Degree Cone, Overlap Section APsmm= 996 Pa

2 (mm)

 

 

2510 I - 200 150 100
r (mm)

FIGURE 7.21 The 4-Blade Fan with 7 Deg. Cone - "Shroud On" Condition (APshroud =
996 Pa) Overlap Region

4 Blade Section 81. 7 Deg Cone, Composite View APM= 996 Pa

2 (mm)

 

 

 

 

 

 

IIIIII r1rrr1rr.1 .1rr.1 11.1.1
400 350 300 250 200 150 100 50 0
r(nm)
FIGURE 7.22 Composite View of the 4—Blade Fan with 7 Deg. Cone - "Shroud Off"
Condition

130

 

 

 

          

400 350 300 250 200 1 50 100 50 0

r(rrm)
FIGURE 7 .23 Composite View of the 4-Blade Fan with 7 Deg. Cone - "Shroud On"
Condition (APshroud = 498 Pa)

 

4 Blade Section & 7 Deg Cone, Composite View APM= 996 Pa

    

 

7111111 I 1" IIIIII [LII]
400 350 300 250 200 150 100 50 0

FIGURE 7.24 Composite View of the 4-Blade Fan with 7 Deg. Cone - "Shroud On"
Condition (APshroud = 996 Pa)

131

4-Blade Fan WI 11 Deg Cone, Center Section " Shroud Off " Condition

 

 

 

 

    

 

 

 

 

 

,1 ,1 ‘
0
-. 1’
-/
25 71 ‘
I
E 50 -
E. ;
N .
75 I
E
100 T
1
I
125 'I
I .1 g -
150 100 50 0
r (mm)
FIGURE 7.25 The 4-Blade Fan with 11 Deg. Cone - "Shroud Off" Condition Center
Region
4-Blade Fan w/ 11 Deg Cone, Center Section, APsmud = 498 Pa
0
25
E 50
E
N
75 ~
100
125 '
150 I H H 100 50 o
r (mm)
FIGURE 7.26 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition (APshroud =
498 Pa) Center Region

132

4-Blade Fan wl11 Deg Cone, Center Section, APWWd = 996 Pa

  
  

J

 

25

100 50
r (mm)

FIGURE 7.27 The 4-Blade Fan with 11 Deg. Cone - "Shroud 0n" Condition (APshroud =
996 Pa) Center Region

all Section " Shroud Off " Condition

I

 

4-Blade Fan w I 11 Deg Cone, NearW

    

 

 

400 350
r (mm)
FIGURE 7.28 The 4-Blade Fan with 11 Deg. Cone - "Shroud Off" Condition Near Wall
Region

133

 

 

 

 

r (mm)

FIGURE 7.29 The 4-Blade Fan with 11 Deg. Cone - "Shroud 0n" Condition (APshroud =

498 Pa) Near Wall Region

4-Blade Fan w/ 11 Deg Cone, Near Wall Section, A 5mm = 996 Pa

   

    

 

 

400 250

r (mm)

FIGURE 7.30 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition (APshroud =

996 Pa) Near Wall Region

134

2 (mm)

 

 

 

300 250 200 150
r (mm)
FIGURE 7.31 The 4-Blade Fan with 11 Deg. Cone - "Shroud Off" Condition Overlap
Region

4-Biade Fan WI 11 Deg Cone, Ovenap Section, AP"mud = 498 Pa

 
 
   
    

25

  
  

IrlllNki

 

E 50 —

E, :

N I

75 r

100 —
125

300 250 200 150
r (mm)
FIGURE 7.32 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition (APshroud =
498 Pa) Overlap Region

135

4-Blade Fan w I 11 Deg Cone, Overlap Section, APmmd = 996 Pa

  

25

50

2 (mm)

75'

 

125

250 200 150
r (mm)

FIGURE 7.33 The 4-Blade Fan with 11 Deg. Cone - "Shroud On" Condition (APshroud =

996 Pa) Overlap Region

4-Blade Section & 11 Deg Cone, Composite View " Shroud Off "

 

 

 

 

0
50
100
150
200
I l L ‘ ‘ I ' I l I I I
400 300 200 100 0
r (mm)
FIGURE 7.34 Composite View of the 4-Blade Fan with 11 Deg. Cone - "Shroud Off"
Condition

136

2 (mm)

2 (mm)

4-Blade Fan WI 11 Deg Cone, Composite View, APsm= 498 Pa

  
  

50
100
150

200

1 1 x l l

l
100 0

t i L J; 1

1
200
r (mm)
FIGURE 7.35 Composite View of the 4-Blade Fan with 11 Deg. Cone - "Shroud On"
Condition (APshroud = 498 Pa)

1 i , , l
400 300

4-Blade Fan w/ 11 Deg Cone, Composite View, APW= 996 Pa

200

1 1

l
100 0

l , 1 l i
400 300 200

r (mm)
FIGURE 7.36 Composite View of the 4-Blade Fan with 11 Deg. Cone - "Shroud 0n"
Condition (APshroud = 996 Pa)

137

5-Blade Section, Center Section " Shroud Off " & AP=25 Pa

  
    

    

 

 

'20 — Vz(m/s)
’ w; 5.61
5.03
0 4.45
3.87
_ 3.29
— 2.71
20 T 2.13
~ 1.55
— 0.97
E 40 L 0.39
E C
N I
60 L-
80 —
100
150 V 100 o

50
r (mm)

FIGURE 7.37 The S-Blade Fan with 11 Degree Cone - "Shroud Off" Condition, AP = 25
Pa - Center Region

5-Blade Section, Center Section " Shroud Off" & AP=50 Pa

 

 

 

 

2 (mm)

 

 

l
150 100 50
r (mm)

FIGURE 7.38 The 5-Blade Fan with 11 Degree Cone - "Shroud 0ft" Condition, AP = 50
Pa — Center Region

138

=373 Pa & AP=25 Pa

shroud

20

40

Illl'llll'lil

2 (mm)

 

60

80

llllliillll

 

 

150‘ > 100 50
r (mm)
FIGURE 7.39 The S-Blade Fan with 11 Degree Cone - "Shroud 0n" Condition (APshroud
= 373 Pa), AP = 25 Pa - Center Region

5-Blade Section, Center Section, AP =373 Pa & AP=50 Pa

shroud

 

7 V2 (mls)

T} 1.54

1 1.35
i

‘rrfif 0.98

 

 

 

 

 

150 "" 100

50
r (mm)

FIGURE 7.40 The S-Blade Fan with 11 Degree Cone - "Shroud On" Condition (APshroud
= 373 Pa), AP = 50 Pa - Center Region

139

5 Blade Section, Near Wail Section " Shroud Off", AP=25 Pa

 

  
 

25

 

 

 

50

 

2 (mm)

75

400 ' 376.0 250
r (mm)
FIGURE 7.41 The S-Blade Fan with 11 Degree Cone - "Shroud Off" Condition, AP = 25
Pa - Near Wall Region

 

5-Blade Section, Near Wall Section "Shroud Off", AP=5O Pa

25

 

 

 

50

 

2 (mm)

75

l l

« 1 1 L1 L _|_
400 3 250

FIGURE 7.42 The S-Blade Fan with 11 Degree Cone - "Shroud Off" Condition, AP = 50
Pa - Near Wall Region

140

5-Blade Section, NearWall Section, APshroud = 373 Pa, AP=25 Pa

. vz (mls)
—j .03
’ 4.47
W12?

 

   

       

 

 

 

 

0 3.91

3.35

2.79

2.23

25 1.68

1.12

A 0.56

E 0.00
E 50

N

75
100

   

i I l l l I I I ‘ I
400 350 300 250
r (mm)
FIGURE 7.43 The 5-Blade Fan with 11 Degree Cone - "Shroud On" Condition (APshroud
= 373 Pa), AP = 25 Pa - Near Wall Region

5-Blade Section, Near Wall Section, APshroud = 373 Pa, AP=50 Pa

    

7V1 (m/s)
f“. 5.22

25

 

50

2 (mm)

75

100

 

FIGURE 7.44 The 5-Blade Fan with 11 Degree Cone - "Shroud 0n" Condition (APshroud
= 373 Pa), AP = 50 Pa - Near Wall Region

141

5-Blade Section, Overlap Section, " Shroud Off" & AP=25 Pa

  

v
‘1

l (mls)
7.26

     

 

 

 

 

 

z (mm)

 

 

 

 

 

 

 

' l J I i l l I J

 

 
 

200
r (mm)
FIGURE 7.45 The 5-Blade Fan with 11 Degree Cone - "Shroud Off" Condition, AP = 25
Pa - Overlap Region

5-Blade Section, Ovedap Section, " Shroud Off" & AP=50 Pa

 

 

- . V,(m/S)

20 % 2. 6.69

_ 6.00

_ 5.32

0 4.63

I 3.94

_ 3.26

:— 2.57

20 P 1.89

_ 1.20

A i 0.52
E 40 _
5. :
N I
60 .—
C
80 _-
100 H

 

 

 

150

200
r (mm)
FIGURE 7.46 The 5-Blade Fan with 11 Degree Cone - "Shroud Off" Condition, AP = 50
Pa - Overlap Region

142

5-Blade Section, Overiap Section, APWM=373 Pa 81 AP=25 Pa

 

20

 

40

 

2 (mm)

60

80

llIlIlll'TIIllil‘l—Tllllllllll

 

 
 

 

250 200

r (mm)

FIGURE 7.47 The 5-Blade Fan with 11 Degree Cone - "Shroud On" Condition (APshroud
= 373 Pa), AP = 25 Pa - Overlap Region

5—Blade Section, Overlap Section, APmoud=373 Pa & AP=50 Pa

 

7V1(m/s)

-20 i 5.34

_ 4.79

i m 4.24

0 3.69

: 3.14

_ 2.59

~ 2.05

20 _— 1.50

- 0.95

A I 0.40
E 40 _
E :
N E

60

80 :—
100 {-

 

250 150

200
r (mm)
FIGURE 7.48 The S-Blade Fan with 11 Degree Cone - "Shroud 0n" Condition (Apshroud
= 373 Pa), AP = 50 Pa - Overlap Region

143

5-Blade Section, Composite View, " Shroud Off ", AP=25 Pa

'5

  

 

 

 

 

 

,_ITI|IIrlllllilllillfrllllrl

I 1 1 1_l 1 L L 1 I v ‘ ‘ l 1 1 I I l l

A 1
400 300 200 100 0
r (rrm)

FIGURE 7.49 Composite View of the 4—Blade Fan with 11 Deg. Cone - "Shroud Off"

Condition and AP = 25 Pa

 

   

 

 

 

i I l l L l i ‘ l J l l I I i
300 200 100 0

r (m)

I 1 4

I
400

FIGURE 7.50 Composite View of the 4-Blade Fan with 11 Deg. Cone - "Shroud Off"

Condition and AP = 50 Pa

144

 

   

 

 

1 l

. 1 1 . l
400 300 200

r (run)

FIGURE 7.51 Composite View of the 4-Blade Fan with 11 Deg. Cone - "Shroud On"
Condition (APshroud =373 Pa) and AP = 25 Pa

 

l i ‘ 1 1 l l
100 0

 

Ilrlii

5-Blade Section, Composite View, APsm= 373 Pa, AP=50 Pa

 

 

 

 

 

 

 

. 1 1 1 . 1 . L 1 .
300 200 1 00 0
r (m)

 

Composite View of the 4—Blade Fan with 11 Deg. Cone - "Shroud On"
Condition (APshroud = 373 Pa) and AP = 50 Pa

145

Exit Velocity Profiles - 4-Bl & 7 Deg Diffuser

_ ___t m T

 

 

o1 A

   

LEGEND 1

 

A w .9. 1911511591111“ 5.
1’ + A:shroud: 498 :a 1 1
g ' -A- A shroud-I €96 8 .1
>" :
1
7 _ ‘- . . 2272-. L. ._._ 1.. A-__..__.i1
1 0.9 0.8 0.7 0.6 0.4 0.3 0.2 0.1 o

W 0.5
rIR
FIGURE 7.53 Exit Velocity Profiles: 4-Blade Fan with 7 Degree Cone, AP = 25 Pa

Mass Flow Rate Distribution - 4-Bl 81 7 Deg Diffuser
-1. ------ . f ...----- -. . --- - , , . , .

 

 

 

_I

N

 

 

 

   

 

mass flow rate distribution (kg/s)

3
. LEGEND
4‘ 1 +Tsh‘r60d Off“‘“"‘
1 AP = 498 Pa 1
‘ + shroud_
5' —A— APShrOUd- 996 Pa 1 I
6, . 1
3 1
1 ~.
7 1 "0.9 T613 '07 0.6 0.5 0.4 0.3 0.2 "6? OJ
r/R

FIGURE 7.54 Mass Flow Rate Distribution: 4-Blade Fan with 7 Degree Cone, AP = 25 Pa

146

 

 

   

 

 

A 2‘.

%’ LEGEND

" 3; Wd'Oiffifi—‘h—fi

>" 1 1 6 AP =498 Pa 1 1
4 + P31115115: 996 Pa 1
1 —A— A shroud ‘
51 1
1 1
61 .
7 A1
8‘1 “ * 51'3" 0.5 “”54 0112 b “'52

HR

FIGURE 7.55 Exit Velocity Profiles: 4-Blade Fan with 11 Degree Cone, AP = 25 Pa

1 Mass Flow Rate Distributlon - 4-BI & 11 Deg Diffuser

 

 

 

 

 

A 0
g 1
a) .
5 1
8 2. LEGEND
“3 1 .9. "Shroud Off" 1
.Q 1 = !
+
5 4 + A shroud- a 1 1
.9 1 ’“ FF 1
g 51
E 6 1
LL .
U) 71
a) 1
(U
E 81 1
‘ 1
9 -. . .-.-_ . . A __ 1
1 0.8 0.6 0.4 0.2 o -0.2
r/R
FIGURE 7.56 Mass Flow Rate Distribution: 4_Blade Fan with 11 Degree Cone, AP = 25
Pa

147

 

Exit Velocity Profiles-S- BI& 11 Deg Diffuser (AP= 25 Pa)

-1 7 __fi ,,,,, fi________

 

6 5, ~1 LEGEND
.9. "91115115611" "
1 —i-—# APshroud= 373 Pa “1 1

8 ,Ali - . - A,“ _ AW 1 _____--
1 0.8 0.6 0.4 0.2 0 -0.2

r/ R
FIGURE 7.57 Exit Velocity Profiles: 5-Blade Fan with 11 Degree Cone, AP = 25 Pa

Mass Flow Rate Distribution - 5-BI & 11 Deg Diffuser

 

 

 

mass flow rate distribution (kg/s)

LEGEND
.9; "Shroud Off" W

I APshroud = 373 Pa 11

2.5 . . 7 - .77 — . _ ,- — ..

1 0.9 0.8 0.7 0.6 0.5 0 4 0.3 0.2 0.1 0

r/R
FIGURE 7.58 Mass Flow Rate Distribution: 5-Blade Fan with 11 Degree Cone, AP = 25
Pa

148

 

Exit Velocity Profiles - 5-Bl 8. 11 Deg Diffuser (AP = 50 Pa)

.1” _ h f h

 

 

 

A

1’ 2'.
E

v
"3
>

   

LEGEND

 

41 _ -_ , .
.9. "Shroud Off"
1 + APshroud= 373 Pa -
5, EN
6.
7 . M_ . Av ,, , - I , Afi____ ,:
1 0.8 0.6 0.4 0.2 o -0.2

rl R
FIGURE 7.59 Exit Velocity Profiles: 5-Blade Fan with 11 Degree Cone, AP = 50 Pa

Mass Flow Rate Distribution - 5-Bl 8: 11 Deg (AP= 50 Pa)

flyfi

 

 

-31 . AA_ 1 ”A

1
411 LEGEND

 

1 .9; "Shroud ow"

1 1 + 7 "Apshroud= 373 Pa 1 a

 

 

 

 

 

mass flow rate distribution (kg/s)

 

 

6 1’ of9 ofa’i F07 0.6 7 6.5 0.4 0.3 ’ 0.2 ($1 7 6
r/R
FIGURE 7.60 Mass Flow Rate Distributions: 5-Blade Fan with 11 Degree Cone, AP = 50
Pa

 

149

8.0 Conclusions, a Proposed New Fan System, and Future Work
8.1 Conclusions

The aerodynamic shroud has been applied to a new type of fan system, an agricultural
building ventilation fan. This fan system is completely different from an automotive cool-
ing fan system, which was the motivation problem for the application of the aerodynamic
shroud, see Morris (1997). Although the aerodynamic shroud was originally intended for
fan systems with large tip clearances and high back pressure operating conditions, it has
now been successfully applied to a small tip clearance fan system that operates at very low
pressure rise conditions. The results, as summarized below, show that the effectiveness of
the aerodynamic shroud is highly dependent on the blade design that is used for the fan.
Additionally, it has been shown that a blade design that is considered inferior when used
without the aerodynamic shroud can be dramatically superior when used with the aerody-
namic shroud. However, the proper choice of shroud operating conditions is required to

realize the maximum gains (in both volumetric flow rate and efficiency).

Section 2.0 provides a control volume energy equation balance for the specific type of fan
system that has been experimentally evaluated in the present investigation. This section
also describes the loss mechanisms in detail and incorporates them into the energy balance
for the fan system and aerodynamic shroud. The limitations of the popular definition of
static efficiency (11) were demonstrated and a newer definition, 11*, was presented. This
definition accounts for the kinetic energy that is created by a low pressure rise fan which,
along with volumetric flow rate, Q, are argued to be useful outputs of the fan system. The
conventionally used definition of Q/ g) is also acknowledged here, but it is suggested that

n* is the preferred definition because it incorporates the kinetic energy and also because it

150

is dimensionless. Finally, a formulation of 11* is derived which clearly shows the neces-
sary terms that must be minimized in order to obtain a maximum efficiency for the fan

system in question.

A comprehensive study of the integral quantities: i) system pressure rise, AP, vs. volumet-
ric flow rate, Q, as well as the ii) system pressure rise, AP, vs. system efficiency, n*, is pre-
sented in Section 6.0. These variables were presented in a manner described above
because the pressure rise operating condition (nominally AP z 25 Pa) is nearly constant for
the majority of the life cycle of an agricultural building ventilation fan. Therefore, the ben-
efits, or lack of benefits, for a particular combination of parameters can be quickly identi-
fied from the figures in Section 6.0. The uncertainty of the data measured in the AF RD
facility was derived analytically and then evaluated by comparing them with published
data. The reliability of the measurements was demonstrated by showing the small varia-

tion of several test results over several days.

There was a systematic increase in volumetric flow rate, Q, for the each of the 4-blade fan
tests when the aerodynamic shroud in the "shroud on" condition. The magnitude of this
increase increased as the shroud pressure, APshroud9 was increased. The tests that used the
11 degree diffuser cone showed greater increases for a given shroud pressure than those
that used the 7 degree diffuser cone. However, these gains were accompanied by signifi-
cant decreases in system efficiency, 11*. The best results that involved the 4-blade propel-

ler were those with the 11 degree cone and a shroud pressure, APshmud= 498 Pa. This

configuration yielded an 11.5% gain in volumetric flow rate with a decrease in efficiency

of 6.0%. The Particle Image Velocimetry results from these 4-blade tests confirm that the

151

aerodynamic shroud does increase the velocity and mass flow rate in the outer regions of

both of the diffuser cones. These increases result in higher exit velocities, Vc, and more
non-uniform exit profile, represented by OLE. These PIV measurements also showed that
there are strong secondary losses, kscc, in the region directly downstream of the hub of the

4-blade fan.

A 5-b1ade propeller was obtained and utilized for a new set of measurements. This blade
design did not provide significant increases in Q or 11*, with respect to those of the pro-
duction 4-blade fan, when operating in the "shroud oft" condition. However, when this
propeller was operated in conjunction with the aerodynamic shroud, this combination
showed substantial increases in both volumetric flow rate and system efficiency. Numer-
ous combinations of shroud parameters were evaluated and a final optimal configuration
was identified. This optimal configuration yielded an increase in volumetric flow rate, Q,
of 33.6% and an increase in system efficiency of, 11*, of 12.3% when compared to the val-
ues for the 4-blade production fan system. The effects of increasing and decreasing the

number of blades were examined, but the 5-blade was identified as the optimal propeller.

Particle Image Velocimetry measurements at the exit plane of 11 degree diffuser cone

show that the 5-blade fan generates a velocity profile that reaches a maximum at a median
radius (r/R z 0.5) of the exit plane. The aerodynamic shroud is able to shift this maximum
velocity to a larger radius (r/R z 0.65) and also reduce the exit velocity across much of the
exit plane. The resulting velocity profile is reduced over much of the exit plane, resulting

in lower values of VC, and it also more uniform, which would indicate a lower value for

OLE. Both of these result in higher system efficiency values as was shown in equation 2.20.

152

Finally, a new device, termed the "centerbody" was added and its effects were measured.
To the best of the author’s knowledge, this type of device has not been used in an agricul-
tural building ventilation fan, nor in any fan that uscd the aerodynamic shroud. This cen-
terbody device, when used with the 5-blade fan and the 11 degree diffuser cone in the
"shroud on" condition, provided even further improvements in Q and n* when compared
to those values for the 4-blade production fan system. The volumetric flow rate, Q,

increased 35.7% and the system efficiency, 11*, increased 15.6%.

These data show that the effectiveness of the aerodynamic shroud is highly dependent on
the design of the propeller that is used. The optimal blade design in this study provided
lower flow rate and efficiency than the production design when operated in the "shroud
off' condition, yet significant increases in both flow rate and efficiency when operated in
an optimal "shroud on" condition. These data also suggest that higher diffuser angles than
the widely regarded limit of 7-9 degrees are feasible when the aerodynamic shroud is

operated in an optimal configuration.

8.2 Proposed New Fan System

The results, which were presented in Section 6.0 and Section 7.0 and then summarized
above, have led to the identification of a new fan system. The features of this fan system
are shown in Figure 8.1. It should be noted that these specific modifications are suggested
for a 0.61 in agricultural ventilation building ventilation fan. However, these improve-
ments could potentially apply to any fan which has a low pressure rise operating condition
and a high volume flow rate. It is suggested that the aerodynamic shroud be implemented
for this fan system. This feature can be integrated as shown in Figure 8.1 with a shroud

gap of 5.0 mm and a shroud pressure of 373 Pa. The 5-blade propeller should be installed

153

 

in this fan system and a cylindrical centerbody should be placed directly downstream of
the propeller. This centerbody should have a diameter which matches the hub of the pro-
peller. Since the centerbody is stationary and the propeller rotates, the spacing between
these two devices should be minimal. The stationary centerbody can be supported by the
downstream protective guard screens. A diffuser that is no smaller than 11 degrees should
also be installed on this fan system. It is acknowledged that further improvements in flow
rate and efficiency might be possible with larger-angled cones, but this has not been con-

firmed in this study.

8.3 Future Work

The present investigation represents a systematic study of several different blade designs

and their effectiveness when used in conjunction with active control elements, such as the
aerodynamic shroud, and passive elements such as a centerbody device. Although an opti-
mum blade design was identified among those evaluated, it is anticipated that many design
parameters exist that could be modified to improve upon the results presented here. Future
work in this area could involve numerical computations and/or an optimization scheme to

be used to identify a blade design that would best fit the assembly presented in Figure 8.1.

The usefulness of the centerbody device in this application, particularly when used with
the aerodynamic shroud, has been demonstrated in the present study. The only centerbody
shape that was evaluated was the cylindrical cone that is shown in Section 6.6. Additional
shapes could possibly provide further improvements beyond those found here. Also, the
use of small stator elements, which create pressure rise in a manner similar to diffusers,

could be used on the surface of the centerbody.

154

 

The aerodynamic shroud has been shown to dramatically improve the performance of a
fan that is completely different than that of the original application. Many other applica-
tions exists which have specific challenges different than those of this study and the previ-
ous study. The use of an aerodynamic shroud in turbomachinery applications of all type
could have a profound effect on the design approaches and methodologies currently used

to design fans.

Shroud Pressure = 373 Pa
\ 5-Blade Propeller Cylindrical Centerbody

(with Wider Hub) /

 

Shroud Gap = 5.0 mm 1

 

 

 

 

1

11 Degree Diffuser

FIGURE 8.1 The Proposed New Fan Design

155

REFERENCES

1.

AMCA Directory of A gricultural Products with Certified Ratings (2000), Air Move—
ment and Control Association, Inc. Publication 262-00.

. AMCA Laboratory Methods of Testing F ans for Rating, (1985), Air Movement and

Control Association, Inc. Publication 210-85.

. AMCA Laboratory Certified Ratings Program Air Performance, (1994), Air Move-

ment and Control Association, Inc. Publication 211-94.

. Bickert, W.G. (1988), “Full wall ventilation to improve warm weather environments in

naturally ventilated livestock buildings,” Livestock Environment 111, Proc. of the Third
Int’l. Livestock Envir. Symp. ASAE, 2950 Niles Rd., St. Joseph, MI. pp. 1-7.

Bleier, F. P. (1998), Fan Handbook: Selection, Application and Design, McGraw-Hill,
New York.

Brody, S., Ragsdale, A.C., Thompson, H.J. and Worstell, D., (1954), “The effect of
wind on milk production, feed and water consumption and body weight in dairy cat-
tle,” Mo. Agr. Exp. Sta. Res. Bull. 545. Univ. of Missouri, Columbia, MO.

Chastain, J .P. (1991), “Heat and mass transfer through the bovine hair coat during
evaporative cooling,” Unpublished PhD dissertation. University of Kentucky, Lexing-
ton, KY.

8. F lowMap Installation and User ’3 Guide. (1997), Dantec Measurement Technology.

10.

11.

12.

13.

14.

15.

16.

17.

. Ford, S.E., Christianson, L. L., Riskowski, G. L., Funk, TL. (1999), Agricultural Ven-

tilation Fans, Performances and Efficiencies, UILU ENG 99-7001.

F uquay, J .W. (1981), “Heat stress as it affects animal production,” .1. Animal Sci.
52(10:l64-174).

Hanan, J .J . (1998), Greenhouses, Advanced T eclznology Protected Horticulture, CRC
Press, Boca Raton, FL.

The Handbook of Fluid Dynamics, edited by Johnson, Richard W., (1998), CRC Press.

Herz, S. A. (1998), “Centrifugal Blower Considerations for the Purpose of Aerody-
namic Shroud Pressurization,” ME 490 independent study final report, summer semes-
ter, Michigan State University College of Engineering.

Host-Madsen, A. and McCluskey, DR. (1994), “On the Accuracy and Reliablity of
PIV measurements,” Proceedings of the Seventh International Symposium on Applica-
tion of Laser Techniques to Flow Measurements, Lisbon.

Johnson, H.D. (1965), “Environmental temperature and lactation,” Int ’1 J. Biometeor.
92103-116.

Keane, RD. and Adrian, R]. (1992), “Theory of cross-correlation analysis of PIV
images,” Applied Scientific Research, vol. 49, pp 191-215.

Lakshminarayana, B. (1996), Fluid Dynamics and Heat Transfer of T urbomachinery,
Wiley and Sons, Inc., New York.

156

18.

19.

Mangold, D.W., Hazen, TE. and Hays, V.W. (1967), “Effect of air temperature on per-
formance of growing-finishing swine,” Trans. ASAE lO(3):370-375, 377.

Mass Airflow Test Stand (MAFS) Operator '5 Manual, (1995), Engineered Systems,
Inc., Project Nr. P93045A.

20. Mass Airflow Test Stand Commercial Literature, 80. 2169, Model 11359 Testek, Inc.,

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

Livonia, MI.

Morris, SC. (1997), MS. Thesis, Department of Mechanical Engineering, Michigan
State University.

Morris, S.C., Good, J .J ., Foss, J.F. (1998), “Velocity Measurements in the Wake of an
Automotive Cooling Fan,” Experimental Thermal and Fluid Science, 17 pp 100-106.

Morris, S.C., Foss, J.F. and Pakkala, J .E. (1997), “The Automotive Cooling Fan
Research and Development Facility,” Society of Automotive Engineering Paper, No.
971784.

Morris, SC. and Foss, J .F. (1997), “Performance Measurements and Detailed Flow
Field Observations for a Light Truck Cooling Fan,” Society of Automotive Engineer-
ing Paper, No. 971794.

Morris, S.C., Neal, D.R., Foss, J .F., Cloud, G.L. (2001, February), “A Unique Air Flow
Measurement Device,” Measurement and Science Technology, pp N9-Nl 3.

Morris, S.C., Foss, J .F. (2001, June), “An Aerodynamic Shroud for Automotive Cool-
ing Fans,” Journal of Fluids Engineering, Vol. 123.

Neal, DR. and Foss, J .F. (1999) “The Development of the Aerodynamic Shroud for
Enhanced Ventilation in Livestock Buildings,” USDA SBIR Phase I Final Report.

Neal, DR. and Bardouwalla, A. ( 1998), “A Study of the Agricultural Livestock Venti-
lation Fan Market,” Management of Technology and Innovation Final Report - MSC
890, Michigan State University College of Business.

Nienaber, J .A., Shanklin, M.D., Hahn, G.L. and Manak, RC. (1985), “Performance of
neonatal and newly-weaned pigs as affected by temperature and diet,” Trans. ASAE
28(5):1626-1634.

Nienaber, J .A., Hahn, G.L. and Yen, J .T. (1987), “Thermal environmental effect on
growing-finishing swine: Part I - Growth, feed intake and heat production,” Trans.
ASAE 30(6):1772-1775.

Overhults, D.G. and Gates, RS. (1993), “Energy Use in Tunnel Ventilated Broiler
Houses,” Livestock Environment IV, Fourth Intl Symposium, Coventry, England, July
6-9, pp. 339-346.

Smith, RH. and Wang, OT. (1944), “Contracting Cones Giving Uniform Throat
Speeds” Jour. Aero. Sci. pp 356-360.

Tillotson, R]. and Bickert, W.G. (1994), “Dairy housing ventilation modification due

to level of milk production,” Dairy Systems/or the 21st Century, Proc. of the Third
Int’l Dairy Housing Conf. ASAE, 2950 Niles Rd, St. Joseph, MI. pp. 317-326.

157

34.

35.
36.

37.
38.
39.

40.
41.

42.

43.

44.

Tyrrell, H.F., Brown, R.S., Reynolds, P.J., Haaland, G.C., Bauman, D.E., Peel, OJ. and
Steinhour, W.D. (1988), “Effect of somatotropin on metabolism of lactating cows:
energy and nitrogen utilization as determined by respiration calorimetry,” J. Nutr.
118:1024-1030.

Osborne, W. C. (1966), Fans, Pergamon Press, Oxford.

Overhults, D.G. and Gates, RS. (1997), “Energy requirements for operating tunnel
ventilated broiler houses,” Proceedings of the Fifth International Livestock Environ-
ment Symposium, Bloomington, MN, May 29-31, pp.859-865.

Patterson, G.N. (1938), Modern Diffuser Design, Aircr. Engng. 10, 267.
Potter, MC. and Foss, J .F . (1982), Fluid Mechanics, Great Lakes Press, Okemos.

Squire, HE. (1953), “Experiments on Conical Diffuser” R. and M. 2751 Aero. Res.
Council.

Wallis, R. A., Axial Flow Fans and Ducts, Wiley and Sons, Inc., New York, 1983.

Wheeler, E. F. and Both, A. J. (2002), “Evaluating Greenhouse Mechanical Ventilation
System Performance,” Agricultural and Biological Engineering Fact Sheet I-42, The
Pennsylvania State University, University Park PA. 4 pp.

Wheeler, E. F. and Both, A. J. (2002), “Evaluating Greenhouse Ventilation Fan Capac-
ity,” Agricultural and Biological Engineering Fact Sheet I-43, The Pennsylvania State
University, University Park PA. 4 pp.

Wilcox, DC. (1993), Turbulence Modeling for C FD, DCW Industries, La Canada,
California.

Yousef, MK. and Johnson, H.D. (1966), “Calorigenesis of dairy cattle as influenced
by throxine and environmental temperature,” J. Animal Sci. 25:150-155.

158