Fan Handbook, Selection, Application and Design, Bleier-text, Notas de estudo de Engenharia Mecânica

Fan Handbook, Selection, Application and Design, Bleier-text, Notas de estudo de Engenharia Mecânica

400 páginas
50Números de download
1000+Número de visitas
60 pontos
Pontos de download necessários para baixar
este documento
Baixar o documento
Pré-visualização3 páginas / 400
Esta é apenas uma pré-visualização
3 mostrados em 400 páginas
Esta é apenas uma pré-visualização
3 mostrados em 400 páginas
Esta é apenas uma pré-visualização
3 mostrados em 400 páginas
Esta é apenas uma pré-visualização
3 mostrados em 400 páginas
Fan Handbook, Selection, Application And Design, Bleier


ank P. Bleier


Other McGraw-Hill Handbooks of Interest

Avallone and Baumeister * marks' standard handbook for mechanical ENGINEERS

Brady etal • materials handbook

Brink • handbook of fluid sealing

Chironis and Sclater • mechanisms and mechanical devices sourcebook

Czernik • gasket handbook

Harris and Crede • shock and vibration handbook

Hicks • handbook of mechanical engineering calculations

Lingaiah • machine design data handbook

Parmley • standard handbook of fastening and joining

Rothbart » mechanical design handbook

Shigley and Mischke • standard handbook of machine design

Skousen • valve handbook

Suchy • die design handbook

Walsh • electromechanical design handbook

Walsh * McGRAW-HILL machining and metalworking handbook

FAN HANDBOOK Selection, Application,

and Design

Frank P. Bleier, re. Consulting Engineer for Fan Design

Boston, Massachusetts Burr Ridge, Illinois

Dubuque, Iowa Madison, Wisconsin New York, New York San Francisco, California St. Louis, Missouri

Library of Congress Cataloging-in-Pub!ication Data

Bleier. Frank P.

Fan handbook : selection, application, and design / Frank P.


p. cm.

Includes index.

ISBN 0-07-005933-0 (alk. paper) 1. Fans (Machinery)—Handbooks, manuals, etc. I. Title.

TJ960.B58 1997

621.6'1—dc21 97-23697 CIP

McGraw-Hill A Division o/'TheMcGraw-Hill Companies

Copyright Q 1998 by The McGraw-Hill Companies, Inc. All rights reserved. Printed in the United States of America. Except as permitted under the

United States Copyright Act of 1976. no part of this publication may be reproduced or distributed in any form or by any means, or stored in a data

base or retrieval system, without the prior written permission of the


5 6 7 8 9 BKM BKM 0987654321

ISBN 0-07-005933-0

The sponsoring editorfor this book was Harold B. Crawford, the editing supervisor was Paul fi. Sobel, and the production supervisor was Tina

Cameron. It was set in Times Roman by North Market Street Graphics.

McGraw-Hill books are available at special quantity discounts to use as premiums and sales promotions, or for use in corporate training programs. For more information, please write to the Director of Special Sales, McGraw-Hill. 1 1 West 19th Street, New York. NY 1001 1. Or contact your local bookstore.

Information contained in this work has been obtained by The McGraw-Hill Companies, Inc. ("McGraw-Hill") from sources believed to be reliable. However, neither McGraw-Hill nor its authors guarantee the accuracy or completeness of any informa- tion published herein and neither McGraw-Hill nor its authors shall be responsible for any errors, omissions, or damages arising out of use of this information. This work is published with the understanding that McGraw-Hill and its authors are supplying information but are not attempting to render engineering or other professional services. If such services are required, the assistance

of an appropriate professional should be sought.


Foreword xi

Preface xiii

List of Symbols xv

Conversion Factors xvii

Chapter 1. Basics of Stationary and Moving Air 1.1

Atmospheric Pressure / /. /

Static Pressure / 1.3

Airflow Through a Round Duct of Constant Diameter. Velocity Pressure / 1,3 Airflow Through a Converging Cone / 1.9 Airflow Through a Diverging Cone / /.// Aerodynamic Paradox / 1.12 Tennis Ball with Top Spin / 1. 13 Airflow Through a Sharp Orifice. Vena Contracta / 1.14 Venturi Inlet / 1.14

Airflow Along a Surface / 1.16

Chapter 2. Airfoils and Single-Thickness Sheet Metal Profiles 2.1

Description and Function of an Airfoil / 2.1 Influence of Shape on Airfoil Performance / 2.3 Lift Coefficient, Drag Coefficient / 2.3 Characteristic Curves of Airfoils / 2.4 Single-Thickness Sheet Metal Profiles / 2.8 Function of Airfoil Blades in Axial and Centrifugal Fans / 2.10

Chapter 3. Types of Fans, Terminology, and Mechanical Construction 3.1

Six Fan Categories / 3.1 Axial-Flow Fans / 3.1 Centrifugal Fans / 3.8 Axial-Centrifugal Fans / .?./.? Roof Ventilators / 3.16 Cross-Flow Blowers / 3.19 Vortex or Regenerative Blowers / 3.20 Conclusion / 3.20

Chapter 4. Axial-Flow Fans 4.1

Nomenclature / 4.1 Mathematical Fan Design versus Kxperimental Cut-and-Try Methods / 4.1


Axial Flow and Helical Flow / 4.2

Blade Twist, Velocity Distribution / 4.3

Two-Stage Axial-Flow Fans / 4.28

Influence of Tip Clearance on the Performance of Vaneaxial Fans / 4.38

Vaneaxial Fans with Slotted Blades / 4.39

Applications with Fluctuating Systems / 4.44

Noise Level / 4.44

Outlet Diffuser and Outlet Tail Piece / 4.45

Selection of Axial-Flow Fans / 4.46

Overlapping Performance Ranges / 4.53

Sample Design Calculation for a 27-in Vaneaxial Fan / 4.56

Axial-Flow Fans Driven by Compressed Air / 4.69

Chapter 5. Fan Laws 5.1

Conversion of Fan Performance / 5. 1

Variation in Fan Speed / 5. / Variation in Fan Size / 5.4

Variation in Both Fan Size and Fan Speed / 5.6

Variation in Size and Speed with Reciprocal Ratios / 5.7

Variation in Density / 5.9

Machining Down the Fan Wheel Outside Diameter / 5. 12

Chapter 6. System Resistance 6.1

Airflow Systems / 6.1

Airflow Through a Pool of Stationary Liquid / 6.1 Airflow Through Filter Bags / 6.3 Airflow Through a Grain Bin / 6.4

Airflow Through a Ventilating System / 6.4

Comparison of System Characteristic Curves and Changing Speed Curves / 6.6

Shifting the Operating Point Out of the Stalling Range / 6.6 Pressure Losses in Ventilating Systems / 6.7

How the Static Pressure Varies Along a Ventilating System / 6.8

Chapter 7. Centrifugal Fans 7.1

Flow Pattern / 7.1 Operating Principle / 7.1

Drive Arrangements / 7.2 Types of Blades / 7.2

Pressure Blowers, Turbo Blowers / 7.42 Turbo Compressors / 7.46

Two Centrifugal Fans in Parallel / 7.55 Volume Control / 7.55 Summary / 7.58

Chapter 8. Fan Selection, Specific Speed, Examples 8.1

Selection of Axial-Flow Fans / 8.1

Selection of Centrifugal Fans / 8.1

Specific Speed N„ Specific Diameter D, / 8.2


Examples of the Selection and Application of Fans / 8.3

Conclusion / 8.12

Chapter 9. Axial-Centrifugal Fans 9.1

Flow Patterns for Various Configurations / 9.1

Performance of Axial-Centrifugal Fans / 9.4

Chapter 10. Roof Ventilators 10.1

Four Ways to Subdivide Them / 10.1 Ten Configurations of Roof Ventilators / 10.2

Chapter 11. Ventilation Requirements and Duct Systems 11.1

OSHA Regulations / //./ Comfort Conditions / 11.1

Five Methods to Calculate the Air Volume Required for a Space / 1 1.2 Four Methods to Design Duct Systems and to Determine the Required Static Pressure / 11.8

Chapter 12. Agricultural Ventilation Requirements 12.1

Types of Ventilation for Agricultural Requirements / 12.


Chapter 13. Cross-Flow Blowers 13.1

Review / 13.


Flow Pattern and Appearance / 13.1 Performance of Cross-Flow Blowers / 13.3 Advantages and Disadvantages / 13.3

Chapter 14. Flow Coefficient and Pressure Coefficient 14.1

Review / 14.1 Two Methods for Comparing Performance / 14.1 Comparison of Coefficients for Various Fan Types / 14.6 Plotting y versus <p Instead of Static Pressure versus Air Volume / 14.7

Chapter 15. Vortex Blowers 15.1

Review / 15.1 Flow Pattern and Appearance / 15.


Principal Dimensions of the Vortex Blower / 15.1 Performance of the Vortex Blower / 15.2 Noise Level of the Vortex Blower / 15.3 Applications of the Vortex Blower / 15.4


Chapter 16. Various Methods to Drive a Fan 16.1

Prime Movers / 16.1 Types of Motor Drive / 16.1 Types of FJectric Motors Used to Drive a Fan / 16.1

Summary / 16.2

Chapter 17. Fanless Air Movers 17.1

Principle of Operation / 17.1

Performance / 17.1

Chapter 18. Performance Testing of Fans 18.1

Description of the AMCA Test Code ' 18.1 Various Laboratory Test Se t ups . / 8.


Accessories Used with Outlet and Inlet Ducts / 18.3

Instruments / 18.6

Test Procedure / 18.17

Test Forms / 18.19 / 18.20

Chapter 19. Vacuum Cleaners 19.1

Review / 19.1

Configurations / 19.1

Testing / 19.2

Performance / 19.8

Chapter 20. Fan Performance as Shown in Catalogs 20.1

Three Ways to Present Fan Performance / 20.1 Computation of Rating Tables for Belt Drive, Derived from Test Performance

Curves / 20.2

Computation of Rating Tables for Belt Drive from Another Rating Table / 20.3

Analyzing Rating Tables Published by Manufacturers / 20.4

Chapter 21. Air Curtains 21.1

Flow Pattern and Function / 21.1 Outside Wind / 21.1 Induced Airflow / 21.1

Requirements / 21.2

Three Examples / 21.7

Performance Testing / 21.8

Summary / 21.8

Chapter 22. Ceiling Fans 22.1

Description / 22.1

Function During the Heating Season / 22.1


Function During the Summer / 22.1 Performance / 22.2

Applications / 22.3

Summary / 22.4

Chapter 23. AMCA Standards 23.1

Standards Handbook / 23.


Centrifugal Fans / 23.1

Drive Arrangements for Centrifugal Fans / 23.1

Operating Limits for Centrifugal Fans / 23.4

Spark-Resistant Construction / 23.6

Summary / 23.8

Chapter 24. Mechanical Strength 24.1

Centrifugal Force / 24.1

Tensile Stress and Yield Stress / 24.1

Axial Reaction Force / 24.2

Shaft Torque / 24.3

Test Pit / 24.3

Chapter 25. Trouble Shooting and Problem Solving 25.1

Guidelines / 25.1

Converging Cone / 25.2 Wrong Rotation / 25.2 Wrong Inlet Spin / 25.2 Wrong Units (Metric) / 25.2

Chapter 26. Installation, Safety, and Maintenance 26.1

Safety Precautions / 26.1

Airflow at Fan Inlet and Outlet / 26.2

High-Temperature Fans / 26.2 V-Belt Drive / 26.2

Lubrication of Bearings and Couplings / 26.4

Vibration / 26.4

Protection of Fan While Not in Use / 26.5

Index follows Chap. 26 1.1


After receiving a degree in applied physics, Mr. Bleier worked for three fan manu-

facturers, first as a draftsman in Lyons, France, next as a test engineer in East Moline,

Illinois, and then as a director of research in Chicago, Illinois. Since then, he has

worked as a consulting engineer for 137 fan manufacturers, most of them in the

United States, some in Canada, and one in Germany. Over the years, Mr. Bleier has designed and tested close to 800 fans, among them

most of the units pictured in this book. His designs ranged in size from a 4-in-

diameter vaneaxial fan, used to ventilate a copy machine, to a 1000-hp, four-stage

turbo blower, producing more than 300 in of static pressure, for pneumatic convey-

ing of grain from cargo ships. Some of his other interesting assignments have included the design of low-noise exhaust fans for invasion ships used by the Navy during World War II and the development of a pressure blower used to pressurize the flotation bags of an Army tank to make it amphibious.

Mr. Bleier has written seven articles for technical magazines and for two engi- neering handbooks in simple, easy-to-understand language. He also has held twelve seminars at universities and for industrial groups. He is listed in Who's Who in Engi- neering and holds several patents on mixed-flow fans.

By sharing his broad experience with others, Mr. Bleier will help engineering stu- dents and people engaged in the design, manufacture, selection, application, and

operation of fans. If you are active in one of these fields, you will benefit from read- ing this book.

Jerome R. Reich, Ph.D.

Chicago, Illinois



Let me say a few words about Robert Andrews Millikan. He was the American physicist who performed the so-called oil-drop experiment to determine the electric charge of an electron. Millikan also was a good teacher and was proud of it. He once made the statement: "I can explain anything to anybody "That's quite a statement. It impressed me. In writing this book, I have kept this statement in mind and have tried

to produce an understandable text and to present some effortless reading material. The story of fans is about airflow considerations, such as velocities, pressures, and

turbulence losses. This book will give explanations of these concepts and present sample calculations to enable engineers and nonengineers to design fans and sys- tem&/to select and apply fans for systems, and to meet requirements for air volume, static pressure, brake horsepower, and efficiency.

If the reader is familiar with high school mathematics, he or she will be able to understand and apply the principles, graphs, and formulae presented here. Calculus and differential equations are not used in this book. Instead, a "feel" for aerody- namics will be developed gradually, a judgment of what an air stream will or will not

do.The early chapters present the basics that will be needed to understand the prin- ciples discussed in later, more advanced chapters.

In grateful memory to

Mr. Archibald H. Davis

my former boss and teacher.

Frank P. Bleier Chicago, Illinois



Cvmbol Meaning Unit

AR Air ratio 1 A Area ft Aa Annular area ft


AR Aspect ratio I a Angle of attack 5 Air angle past blade

ahp Air horsepower hp

BP Barometric pressure inHg

P Relative air angle

p + ct Blade angle o

/ Blade width in

bhp Brake horsepower hp

ft /mincfm or Q Rate of flow D Wheel diameter in d Hub diameter in DB Dry-bulb temperature op

WB Wet-bulb temperature op t Temperature


T Absolute temperature K dB Decibel sound level dB DR Diffuser ratio 1 Ds Specific diameter in

Specific speed min"' £ J Frequency, musical note s

V Volts V A Amps A W Watts W kW Kilowatts kW ME Mechanical or total efficiency % SE Static efficiency % r Radius in

P Air density lb/ft*

Re Reynolds number rpm Revolutions per minute min" 1

SP Static pressure, positive inWC, psi SP Static pressure, negative inWCinHg TP Total pressure inWC V Vacuum inHg

ft'V Volume V Velocity fpm VP Velocity pressure

Flow coefficient


Pressure coefficient




1 ft = 1 2 in 1 in = 0.0833 ft 1 yd = 3 ft = 36 in

1 mile = 1760 yd = 5280 ft


1 ft 3 = 144 in' 1 in 2 = 0.00694 ft : ! yd

2 = 9 ft 2 = 1 296 in :


1 ft' = 1728 in'


1 fpm (foot per minute) = 0.01 1364 mph {miles per hour) 1 mph = 88 fpm Acceleration due to gravity f> = 32. 17 ft/s



1 inHg = 13.595 inWC = 0.4912 psi (lb/in 2 ) (used for high vacuums)

1 psi = 2.036 inHg = 27.68 inWC (used for high pressures) 1 atm (atmosphere) = 29.92 inHg = 406.8 inWC = 14.7 psi 1 inWC = 0.0736 inHg = 0.0361 psi


1 hp = 0.746 kW = 746 W = 42.42 Btu/min 1 kW= 1000W= 1.341 hp = 56.89 Btu/min





Thus (example) 80°F* <80 - 32) x| ^ 26.7°C

Absolute temperature T=T + 459 7 Example: 80"F= 539.7 K (absolute temperature)





Our planet earth has an average diameter of about 7914 mi or a radius of 3957 mi. It is surrounded by a comparatively thin layer of air. The air pressure is highest close to earth, due to compression by the weight of the air above. At higher altitudes, as the height of the air column above becomes less, the air pressure decreases, as shown in Fig. 1 . 1 . At sea level, the atmospheric or barometric pressure is 29.92 inHg. At an alti- tude of 15 mi or 79,200 ft, which is only 0.4 percent of the earth's radius, the atmo-

spheric pressure is only 1.00 inHg (3 percent of the sea level pressure). However, some rarified air extends about 500 mi up, which still is only 13 percent of the earth's radius.

The air consists mainly of nitrogen (about 78 percent by volume) and oxygen (about 21 percent by volume) plus less than 1 percent of other gases. Air is a physi-

Altitude (Miles)

FIGURE 1.1 Atmospheric pressure %'ersus altitude.



cal mixture (not a chemical compound) of these gases. Normally, air also contains some water vapor. This reduces the air density, as will be discussed in Chap. 18, on testing.

According to the National Advisory Committee for Aeronautics (NACA), later succeeded by the National Aeronautics and Space Administration (NASA), tem- perature, atmospheric pressure, and air density at various altitudes are as shown in Table 1.1.

TABLE 1.1 Temperature, Pressure, and Density versus Altitude

Atmospheric Air Altitude, Temperature, pressure, density,

ft op

inHg lbm/ft1

0 59.0 29.92 0 07AS

1,000 55.4 28.86 0 0743

2,000 51.8 27 82 0 07? 1

3,000 48.4 26 81 o 07fm

4,000 44.8 25 84

5,000 42.1 24 89

6,000 37.6 23 98

7,000 34.0 23 09 n 06?n

8,000 30.6 22 22 0 0601

9,000 27.0 21 38 v.l/JOJ

10,000 23.4 20.58 0 05(SS

11,000 19.8 19.79 0.0547

12,000 16.2 19.03 0.0530

13,000 12.6 18.29 0.0513

14,000 9.2 17.57 0.0497

15,000 5.5 16.88 0.0481

20,000 -12.3 13.75 0.0407

25,000 -30.1 11.10 0.0343

30,000 -48.1 8.88 0.0286

35,000 -65.8 7.04 0.0237

40,000 -69.7 5.54 0.0188

45,000 -69.7 4.35 0.0148

50,000 -69.7 3.42 0.0116

55,000 -69.7 2.69 0.0092 60,000 -69.7 2.12 0.0072 65,000 -69.7 1.67 0.0057

Source: Robert Jorgensen, Fan Engineering. Buffalo Forge Co.. Buffalo,

The standards used in fan engineering are slightly different: Here the density used for standard air is 0.075 lbm/ft3 . This is the density of dry air at an atmospheric pressure of 29.92 inHg or at 29.92 x 25.4 = 760 mmHg. We are not discussing fans yet, but in order to get a comparative idea of fan pres-

sure versus atmospheric pressure, let us anticipate for a moment and pretend that we know already what static pressure is and how much static pressure can be produced by certain types of fans.

Since 1 inHg equals 13.6 inWC (inches of water column), 29.92 inHg equals 13.6 x 29.92 = 406.8 inWC. In other words, the standard barometric pressure of 29.92 inHg also can be expressed as 406.8 inWC. This means that a fan producing a static


pressure of 3 inWC (a good average value) will increase the absolute air pressure by less than 1 percent.

On the other hand, 1 inHg equals 0.491 lb/in 2 (psi). Therefore, 29.92 inHg equals 0.491 x 29.92 = 14.7 lb/in : . In other words, the standard barometric pressure of 29.92 inHg also can be expressed as 14.7 lb/in 2 . For high-pressure centrifugal units (either units running at very high speeds or multistage units), static pressure is usually mea- sured in pounds per square inch (psi). Such units often produce as much as 7 lb/in 2 . They will increase the absolute pressure by a significant 48 percent.


Figure 1.2 shows a cylinder with a piston that can be moved up or down. It also shows a U-tube manometer indicating zero pressure. This means that the pressure below the piston is the same as the barometric pressure in the surrounding air. As the piston is moved down, the air volume below the piston is compressed, and the

manometer will register a positive static pressure relative to the atmospheric

pressure, which is considered zero pres-

sure. This compressed air then has potential energy, i.e., the potential to

expand to its original volume. If, on the other hand, the piston is raised, the air

volume below the piston is expanded, and the manometer will register a nega- tive static pressure relative to atmo-

spheric pressure. This expanded air also has potential energy, i.e., the potential to

contract to its original volume. This

explains the concept of positive and neg- ative static pressure in stationary air.

Positive and negative static pressure exists in moving air as well as in station- ary air. A fan blowing into a system (including such resistances as ducts,

elbows, filters, dampers, and heating or cooling coils) produces positive static pres- sure, which is used to overcome the various resistances.A fan exhausting from a duct system produces negative static pressure, which again is used to overcome the resis- tance of the system.*

FIGURE 1.2 Cylinder with piston and manometer. As the piston moves, the static pres- sure below the piston will become either positive or negative.


Air flowing through a straight, round duct of constant diameter has a velocity distri-

bution, as shown in Fig. 1.3, with the maximum air velocity near the center and with zero velocity at the duct wall. For small duct diameters of 6 to 10 in, and for air veloc-

* Some of the materia! in this section was taken from Bleier, F. P., Fans, in Handbook of Energy Systems Engineering, copyright © 1985 by John Wiley and Sons, New York.


ities of 1000 to 3000 ft/min (fpm), the average velocity V is approximately equal to 91 percent of the maximum velocity at the center. To find the average velocity in larger ducts and for larger air velocities, a so-called Pitot tube traverse across the duct is taken (Fig. 1.4). From the average velocity V (in feet per minute) and the cross-sectional area ofthe duct (in square feet),we can calculate the volume of air Q (in cubic feet per minute, or cfm) as

Q=A xV (1.1)

Furthermore, from the average velocity V (in feet per minute) and the air density d (in pounds mass per cubic foot),we can calculate the velocity pressure VP (in inches of water column, inWC) as


- \T

FIGURE L3 Velocity distribution for the airflow through a round duct,

FIGURE 1A A 27-in-diameter test duct with two supports for Pitot tube tra- verses and with a throttling device at the end of the duct.


or for standard air density of d = 0.075 Ibm/ft !


\4005 '

Velocity pressure is the pressure we can feel when we hold our hand in the air stream. It represents kinetic energy.

For a straight, round duct with constant diameter and smooth walls, the friction loss /is

/= 0.0195-^ VP (1.4)

where / = pressure loss, in inches of water column L = length of the duct, in feet D = duct inside diameter, in feet VP- average velocity pressure, in inches of water column

Example: Let's consider a 100-ft-long duct, 2 ft = 24 in i.d., so that the duct area A = 3.14 ft 2 . If the airflow through this duct is Q = 8855 cfm, the average air velocity will be V = 8855/3.14 = 2819 fpm = 47.0 fps, and the corresponding velocity pressure will be VP= (2819/4005) 2 = 0.50 inWCThe friction loss then will be/= 0.0195 (100/2)0.50 = 0.49 inWC

Figure 1 .5 shows a chart for determining friction loss in straight, round ducts. For

our example, proceed as follows: On the horizontal abscissa on top, find the point for 8855 cfm (just slightly below 9000 cfm). From this point, move straight down until you reach the inclined line for a 24-in-diameter duct (pipe). This point will give you two results: (1) Another inclined line near this point indicates that the duct velocity will be slightly more than 2800 fpm. (2) Moving from that point straight across to the vertical ordinate at the right indicates a friction of 0.49 inWC for 100 ft of duct, the same as we obtained above from Eq. (1.4).

Reynolds' Number

If this airflow through the duct were laminar (smooth, streamline, free of eddies), the friction loss would be smaller than that just computed. Unfortunately, laminar air-

flow is seldom found in fan engineering. In most ventilating systems, the airflow is

turbulent. Let's see whether this air flow through the 24-in i.d. duct is really turbu- lent. We can check this by calculating the Reynolds' number for this example. The English physicist Osborne Reynolds studied experimentally the flow of liquids and gases and arrived at a dimensionless parameter that is characteristic for certain flow

conditions. The formula for this Reynolds' number is



where p = gas density, in slugs per cubic foot V = average air velocity, in feet per second R = one-half the duct inside diameter, in feet fi = coefficient of viscosity, in pounds second per square foot

For standard air, p = 2.378 x 10" 3 and |i = 3.73 x 10"


, resulting in

Re = 6375VR (1.5a)

For the preceding example, V = 47.0 and R = 1, resulting in Re = 300,000. Reynolds found that whenever Re is smaller than 1000 to 2000, the airflow will be laminar, but above 2000, it starts to become turbulent. Our example is far above this critical range. The transition from laminar to turbulent airflow is gradual. There are various degrees of turbulence, such as slightly turbulent, very turbulent, and

extremely turbulent. In most fan applications, the air flow is slightly or very turbu-

lent, and this has to be accepted. But extreme turbulence, as is found in a vaneaxial fan without an inlet duct and without a venturi inlet (see the section on venturi inlet), should be avoided.*

Total Pressure, Air Horsepower, Brake Horsepower, Efficiencies

Total pressure TP is defined as the sum of static pressure SP and velocity pressure VP:

TP = SP+VP (1.6)

In this equation, VP is always positive. SP and TP may be positive or negative. Here are three examples, illustrating this:

5P = +2.2inWC = -0.5 inWC SP = -1.4inWC

VT = 0.8inWC VP = 0.8 inWC yP = 0.8inWC

rP = +3.0inWC rP = +0.3inWC TP = -0.6 inWC

Let us now consider another example: a fan having an outlet area OA - 4.00 ft 2 ,

blowing 16,000 cfm into a system, and producing 3 in SP in order to overcome the resistance of the system. The fan will have an average outlet velocity of

cfm 16.000

and a velocity pressure of

MOOOV . „vp '

(4005 - 100 inWC

The total pressure will be TP = 3.00 + 1.00 = 4.00 inWC, and the power output of the fan (called air horsepower, ahp) will be

* Some of the material in this scclion was taken from Bleier, F. P.. Fans, in Handbook of Energy Systems Engineering, copyright © 1985 by John Wiley and Sons, New York.


ahp= CfmX TP

=10.07 hp (1.7)dn 6356

H ;

Ff the motor output (= fan input) is 15 brake horsepower {bhp). the fan efficiency at

this point of operation will be the mechanical efficiency (also called total efficiency)

ME = TE = ="777r = °-67 = 67 percent ( 1 .8)bhp 13.U

Another efficiency that is sometimes used is called the static efficiency. It can be

calculated as follows: First, we calculate the so-called static air horsepower (which, however, is not the real power output of the fan):

cfm x SP 16,000 x 3 ahp^~63^ = ^35o- =7 -55hp (L7a)

And then we calculate the static efficiency:

ahp, 7.55 SE =—— = T777 = 0.50 = 50 percent ( 1 .8a)bhp 15.0

As you can see, the static efficiency is easier to calculate than the total efficiency because we do not have to calculate the velocity pressure first. For this reason, the static efficiency is sometimes used, even though it does not represent the real fan

efficiency. The total or mechanical efficiency is the real fan efficiency. Coming back to our example, to demonstrate the various types of efficiencies,

let's assume that the motor input at this point of operation is 12.7 kW.The motor effi- ciency (or electrical efficiency) then will be

EE = ' P = 0.88 = 88 percent ( 1 .9)

K. W

This equation is used more often for calculating the brake horsepower when input (in kilowatts) and electrical efficiency are known:

kW x EE 12.7x0.88 bhp =-^T = -0746— = 150 (1 9a)

Finally, the efficiency of the set (fan plus motor), called the set efficiency, is

Set eff. = ME x EE = 0.67 x 0.88 = 0.59 = 59 percent (1.10)

In selecting a fan for a certain application, the fan efficiency is of great impor- tance, because with a higher efficiency, we can obtain the same air horsepower with less power input. This not only will reduce the operating cost but also will save energy at the same time. High-efficiency fans, on the other hand, normally are more expensive, as shown in Fig. 1.6. It should be attempted, therefore, to find a balance between first cost and operating cost, taking into consideration that the first cost of the fan unit itself often is only a small portion of the system's total cost.*

* Some of the material in this section was taken from Bleier. F. P.. Fans, in Handbook of Energy Systems Engineering, copyright <D 1985 by John Wiley and Sons. New York.


Fan efficiency

FIGURE 1.6 Cost versus fan efficiency. Selecting a fan of higher efficiency normally results in higher first cost, but in

lower operating cost.

Scalars and Vectors

The physical quantities used to describe airflow phenomena can be divided into two groups: scalars and vectors. Scalars are quantities such as time, temperature, volume,

and mass. They have only magnitude and can be added simply. Vectors are quantities

such as force, velocity, and acceleration or deceleration. They have magnitude and direction and can be added only by way of vector diagrams, such as the velocity dia- grams that will be discussed in later chapters.

When cars on a crowded highway reach a point where the highway narrows, one of two things will happen: (1) the cars upstream will have to slow down, or (2) the cars

past the point of constriction will have to speed up. Possibly both (I) and (2) will

happen. In any case, the cars downstream will travel faster than the cars upstream,

but obviously, the number of cars will remain the same. A similar condition exists when air flows through a converging cone, as shown in

Fig. 1.7. The air volume Q is

where Q = air volume, in cfm (ft Vmin) A =duct area, in square feet (ft : ) V = average air velocity, in fpm (ft/min)

As the airflow passes through the converging cone, the air volume (cfm) obviously will remain the same ahead and past the converging cone. This can be expressed as


Q=AxV (1.1)

or A { x V, = /4,x V: (lib)



FIGURE 1.7 Air flowing through a converging cone: It accelerates.

Bernoulli's Principle of Continuity

This simple and obvious equation is called the principle ofcontinuity because the air

volume (cfm) continues to be the same before and after the point of constriction. As the air passes through the converging cone, it will accelerate from V

] to a larger air

velocity V2 because the area A 2 is smaller than the area A i.

Example: Let's again assume that the upstream duct i.d. is 24 in, so A x = 3.14 ft 2 .

Next let's assume that the downstream duct i.d. is 18 in, so A-, = 1.767 ft 2 . If Q = 8855 cfm, we get V, = 8855/3.14 = 2819 fpm and VP

{ = 0.50 inWC and V, = 8855/1.767 =

501 1 fpm and VP2 = 1.57 inWC. This substantial increase in velocity pressure from 0.50 to 1.57 inWC, of course,

will result in an increased kinetic energy, which will be obtained at the expense of a

decreased static pressure. Basically, this is Bernoulli's theorem, which in its simplest form says: When the air velocity increases, the static pressure will decrease. This is easy to understand, but Bernoulli's theorem also says: When the air velocity decreases (as in a diverging cone), the static pressure will increase. This increase is

called static regain. This is more difficult to understand. It will be discussed in more detail in the section on the diverging cone. A converging cone past a fan is often used to increase the air penetration for such

applications as snow blowing or comfort cooling. A converging cone past the scroll housing of a centrifugal fan usually works with-

out any problem. Care has to be taken, however, on a converging cone past an axial- flow fan because there often is an air spin past an axial-flow fan, even if it is a vaneaxial fan with guide vanes that are supposed to remove the air spin. If a little air spin remains past the fan, it is multiplied manyfold as the air travels to a smaller duct diameter because it tends to retain its circumferential component. As a result, at the smaller diameter, the revolutions per minute of the air spin becomes considerably larger, just like a watch chain spun around a finger turns faster and faster as the chain becomes shorter.

Here is an example illustrating the phenomenon that the air spin increases as the converging cone becomes smaller. Back in 1949, I designed and tested a 14-in vaneaxial fan with a 12-in hub diameter, resulting in an unusually large hub-tip ratio of 86 percent. This was done because the requirements were for a small airflow (cfm) and a high static pressure. A centrifugal fan would have been a better selection, but


FIGURE 1.8 A 14-in vancaxial fan with a 12-in hub diameter blowing into a converging transition cone and a 6-in test duct.

the customer insisted on a vaneaxial fan. Since the annular area was so small, I used

a 6-in test duct past the unit (in order to get a good air velocity through the duct)

plus a transition cone from 14 in down to 6 in, as shown in Fig. 1.8. This test setup, of course, was not in accordance with the Air Movement and Control Association (AMCA) test code, which requires that the test duct area be within 5 percent of the fan outlet area. A better test would have been on a nozzle chamber instead of on a test duct, but in 1949 very few companies had a nozzle chamber. To my surprise, I found zero air velocity in the 6-in test duct. The reason was that the remaining air spin past the fan became so strong in the transition cone that the friction path became excessive and used up all the static pressure available. After I put two longi- tudinal cross sheets into the transition cone to prevent the air spin, the proper air-

flow was restored and a fairly normal duct test could be run.


As discussed earlier, a converging cone, as shown in Fig. 1.7, will produce an increased air velocity past the cone, resulting in increased kinetic energy, which is

obtained at the expense of a decreased static pressure. By the same token, a diverg- ing cone, as shown in Fig. 1.9, will produce a decreased air velocity past the cone, resulting in decreased kinetic energy. Will this difference in kinetic energy be lost?

FIGURE 1.9 Air flowing through a diverging cone: It decelerates.


About half of it will be lost, mainly due to turbulence. The other half, believe it or

not, will be regained by an increase in static pressure, as stated by Bernoulli's theo-

rem, provided that the cone angle is small. About 7° or less on one side is recom-

mended. While air normally flows from higher static pressure to lower static

pressure, here is a case where the opposite takes place: Air is flowing from lower

static pressure to higher static pressure.

Using the same dimensions and the same airflow as in Fig. 1 .7, the velocity pres-

sure now will decrease from 1.57 to 0.50 inWC, for a reduction of 1.07 inWC. One- half of this, or about 0.53 inWC. can be expected as a static regain. Such a regain is

sometimes obtained by the use of a diffuser past a fan.


Normally, as we go along with the airflow through a duct system, the static pressure is highest upstream and gradually decreases from there. This is the reason why the air flows. The high static pressure upstream forces the air through the duct, filters, etc. Therefore, it seems hard to believe that the static pressure will increase as the air

passes through a diverging cone. It seems contrary to common sense. It seems para- doxical.

Let us discuss a device that might convince you of the truth of the preceding

statements. It is called an aerodynamic paradox and is shown in Fig. 1.10a. It consists of a circular plate A with a pipe B on top. Another thin, lightweight disk C is sus- pended about 1/2 in below A in such a way that it can easily move up. If we blow into the pipe on top, we would expect that the air stream will blow the lower disk down. Actually, the lower disk will move up. Let me explain why.

The air stream leaving the pipe will turn 90° and move outward, since it has no other way to go. As it moves outward, the cross-sectional area becomes larger (as in


a diverging cone), so the air velocity becomes smaller. As the air stream reaches the outside of the disks, the air velocity will be quite small, and the static pressure at that

point will be close to the atmospheric pressure of the surrounding environment. Fur-

ther inside, however, where the cross-sectional area is much smaller, the velocity of the air flowing outward is larger, and (per Bernoulli) the static pressure, therefore, is

lower than the atmospheric pressure that pushes against the underside of the lower

disk. As a result, the lower disk is lifted up against the upper plate. The moment the lower disk touches the upper plate, the air stream is stopped,

and the lower disk will drop again. The phenomenon then will repeat itself.


As the airflow passes through a system of ducts, converging and diverging cones, etc., the velocity pressure (kinetic energy) may increase or decrease and the static pres- sure (potential energy) also may increase or decrease. These two pressures are mutually convertible. However, the total pressure (total energy), being the sum of velocity pressure and static pressure, will always decrease, since it is gradually used

up by friction and turbulence.


Another example illustrating Bernoulli's principle is a tennis ball moving through the air with top spin, say, from right to left. To analyze the flow conditions, let's exam- ine an equivalent configuration: The ball is spinning in place, and an air stream is moving from left to right, relative to the ball, as shown in Fig. l.K)/?. On top of the ball, the rotation is opposite to the air velocity. This will slow down the velocity of the air passing over the top of the ball. In accordance with Bernoulli's principle, the

slower air velocity will produce a higher pressure in this region. On the other hand, below the ball, the rotation is in the same direction as the air velocity. This will accel- erate the velocity of the air passing over the underside of the ball and will produce a

lower pressure in this region. As a result, the ball is pushed and pulled down. It therefore will drop faster than it would if it were only pulled down by gravity.

High-pressure region


air velocity

\ \

FIGURE 1.106

Low-pressure region

A lennis ball with top spin will drop faster.



When an air stream passing through a round duct of diameter D hits a sharp orifice with a hole of diameter d, a flow pattern as shown in Fig. 1.11 will develop because

the upstream airflow will approach the edge of the opening at an inward angle

rather than in an axial direction. Obviously, this angular velocity will continue past

the orifice. This jet past the orifice will have a minimum diameter of about Q.6d, and

this minimum diameter will occur at a distance of about ().5d past the orifice. After

this point, the airflow will gradually spread out again, but only after a distance of 3d

past the orifice will the airflow fill the duct "evenly," as shown in Fig. 1.3. This con-

traction of the air stream, shown in Fig. 1.12, is called vena contracta (contracted


FIGURE 1.11 Airflow through a sharp orifice.


A similar condition (although somewhat less extreme) exists when an airflow enters a round duct without a venturi inlet, as shown in Fig. 1.1 2. The reason why it is less extreme than in Fig. 1.11 is the upstream flow pattern. In Fig. 1.11. the approaching

air is moving; in Fig. 1.12 it is hardly moving. Nevertheless, even in Fig. 1.12. a vena

contracta exists, even though it is less pronounced. Figure 1.13 shows the improved flow pattern obtained when the duct entrance is

equipped with an inlet bell, also called a venturi inlet. This will reduce the duct resis-

tance and increase the flow (cfm). For best results, the radius should be r = 0.14£) or

more. If due to crowded conditions the radius has to be made smaller, the benefit will be reduced, but it will still be better than no venturi at all.

A venturi inlet is of particular importance at the entrance to an axial fan (as shown in Fig. 1.14) because without the venturi inlet the blade tips would be starved

for air. In a vaneaxial fan, where the blades are short (due to a large hub), we can expect a flow increase of about 15 percent if a venturi inlet (or an inlet duct) is used.


FIGURE 1.12 Airflow entering a round duct.

FIGURE 1.13 Airflow entering a round duct with a venturi inlet at the duct entrance.

In a propeller fan, where the blades are longer (since there is no hub or only a small hub), a flow increase of about 12 percent can be expected. Furthermore, the lack of a venturi inlet (when no inlet duct is used) will result in an increased noise level because the blade tips will operate in extremely turbulent air.

In centrifugal fans without an inlet duct, a venturi inlet will boost the flow by about 6 percent.The improvement here is somewhat less, for the following three rea- sons:

1. The turbulent airflow here will hit the leading edges (not the blade tips), which are moving at lower velocities.

2. Centrifugal fans normally run at lower speeds (rpms) than axial fans.

3. The flow pattern is different in centrifugal fans. The airflow makes a 90° turn before it hits the leading edges of the blades. The airflow ahead of the blades, therefore, contains some turbulence to begin with, and some additional turbu- lence, due to the lack of a venturi inlet, therefore, is less harmful.


FIGURE 1.14 Venturi inlet ahead of a vaneaxial fan.


Let's look at the flow pattern of air passing through an elbow of rectangular cross section, as shown in Fig. 1.15. It is easy to understand that the airflow will tend to crowd on the inside of the outer wall. Call it inertia or centrifugal force, if you will. Obviously, it is mainly the outer wall of the elbow that keeps the air from flowing straight, as it would like to do, due to inertia.

Now let's look at the flow pattern of the air when the inner wall of the elbow has been removed, as shown in Fig. 1.16. The outer wall still keeps the air from flowing straight, and the flow pattern is quite similar to that of Fig. 1.15. Possibly, the air will crowd a little more near the outer wall.


FIGURE 1.16 Airflow along the outer wall only.



FIGURE 1.17 Airflow along the inner wall only.

Now let's go one step further and look at Fig. 1.17, which shows the flow pattern when the outer wall has been removed and only the inner wall has been retained. The air still will not flow straight to the right, as one might expect. It will still attempt to adhere to the inner wall. It may not be 100 percent successful in this attempt, but let's say it will be 70 percent successful. Why does the airflow turn at all, you ask? The reason is that a negative pressure will develop just outside the inner wall, and this negative pressure will tend to keep the airflow fairly close to the inner wall.Thus we can make the following statement:A curved wall will try to keep an airflow on its outside attached to itself. This is an important statement. We will come back to it in Chap. 2, on airfoils.




An airfoil is a streamline shape, such as shown in Fig. 2.1. Its main application is as the cross section of an airplane wing. Another application is as the cross section of a fan blade. This is the application we will discuss now. There are symmetric and asym- metric airfoils. The airfoils used in fan blades are asymmetric. Figures 2.1 and 2.2 show an asymmetric airfoil that has been developed by the National Advisory Com- mittee for Aeronautics (NACA). It is the NACA airfoil no. 65 1 2. Table 2. 1 shows the dimensions (upper and lower cambers) as percentages of the airfoil chord c. Let us

make a list of the features shown in Figs. 2.1 and 2.2:

1. The airfoil has a blunt leading edge and a pointed trailing edge. The distance from leading edge to trailing edge is called the airfoil chord c.

2. The airfoil has a convex upper surface, with a maximum upper camber of 13.3 percent of c, occurring at about 36 percent of the chord c from the lead-

ing edge.

3. The airfoil has a concave lower surface, with a maximum lower camber of 2.4 per- cent of c, occurring at about 64 percent of the distance c from the leading edge. In some airfoils used in fan blades, the lower surface is flat rather than concave.

4. The airfoil has a baseline, from which the upper and lower cambers are mea- sured. The cambers are not profile thicknesses.

5. The angle of attack a is measured between the baseline and the relative air velocity.

6. As the airfoil moves through the air (whether it is an airplane wing or a fan blade), it normally produces positive pressures on the lower surface of the airfoil

and negative pressures (suction) on the upper surface, similar to the phe-

nomenon discussed with Fig. 1.17. While one might expect that the positive pres- sures do most of the work, deflecting the air stream, this is not the case. The suction pressures on the top surface are about twice as large as the positive pres-

sures on the lower surface, but all these positive and negative pressures push and

pull in approximately the same direction and reinforce each other.

The combination of these positive and negative pressures results in a force F. as shown in Fig. 2.1. This force F can be resolved into two components: a lift force L (perpendicular to the relative air velocity) and a drag force D (parallel to the rela- tive air velocity). The lift force L is the useful component. In the case of an airplane wing. acts upward and supports the weight of the airplane. In the case of an axial

fan blade, L (by reaction) deflects the air stream and produces the static pressure of


V Relative air velocity L Lift D Drag F Resultant force a Angle of attack LE Leading edge TE Trailing edge

FIGURE 2.1 Shape of typical airfoil (NACA no. 6512).

Chord c

0 10 ^ 30 40 50 60 70 80 90 100 Distance from leading edge (% of chord)

FIGURE 2.2 Dimensions of NACA airfoil no. 6512.

the fan. The drag D is the resistance to the forward motion of the airfoil. It is the undesirable, power-consuming component. We therefore would like to use airfoil shapes that have not only a high lift L but also a good lift-drag ratio LID. As the

angle of attack changes, lift, drag, and lift-drag ratio all change considerably, as will

be seen in Fig. 2.4


TABLE 2.1 Dimensions of NAC.'A Airfoil No. 6512

Distance from Upper Lower

leading edge camber camber

(% of chord) (% of chord) (% of chord)

0 Z. /


1 .Z5 5.Z5 1.34

Z.MJ o.zi c

/.CHJ \),jZ -7 c 1.7) o.o5

1 1\1U U.IXJ i r I


1 It <Ji\ 1U.89 (J.U5

ZU 1 1 uo1 1.8ft l\ TO

JU 13.IK)

4U 1 3.Z3 Lol

>U IZ. /



60 11.51 2.37

70 9.66 2.29

80 7.18 1.88

90 4.06 1.10

95 2.25 0.58

100 0.24 0


The National Advisory Committee for Aeronautics (NACA) and the Gottingen Aerodynamische Versuchsanstalt have tested many airfoil shapes in wind tunnels in an attempt to find some shapes that will produce high lift forces and at the same time have good lift-drag ratios. These groups found, however, that these are con- flicting requirements. As the cambers increase, the lift normally increases, too, but the lift-drag ratio tends to decrease. Selection of airfoil shapes therefore will

depend on the application. In a high-pressure vaneaxial fan, we will use a high- cambered airfoil, particularly near the blade root. On the other hand, if fan effi- ciency is more important than high static pressure, we will use a low-cambered airfoil shape.

You may wonder why the leading edge of an airfoil is blunt. Wouldn't the drag be smaller if the leading edge were pointed, like the trailing edge? This would indeed be the case, if the relative air velocity at the leading edge were exactly tangential. How- ever, this tangential condition would exist for only one operating condition (i.e., for one flow rate and static pressure). Over most of the performance range, the relative air velocity would deviate from the tangential condition, and this would result in tur- bulence and in an increased drag. Another reason for the blunt leading edge is struc- tural strength.


From the test data for lift L and drag D obtained from wind tunnel tests, we can cal- culate the corresponding coefficients as follows:


Lift coefficient C, - 844Z/

AV 2 for standard air density


844£> (2.2)Drag coefficient CD - AV 2

where L and D are in pounds, A is the area of the tested airfoil plate in square feet, V is the relative air velocity in feet per second, and CL and Cd are dimensionless coefficients. From these formulas we note that CLiCD = LID. In other words, the lift- drag ratio is also the ratio of the corresponding coefficients.

As mentioned earlier, many airfoil shapes have been tested in wind tunnels. The air- foil plates tested by NACA usually have an airfoil chord of 5 in and a length of 30 in, as shown in Fig. 2.3. This is called an aspect ratio of 6.

Figure 2.4 shows the characteristic curves NACA obtained for its airfoil no. 6512 for an aspect ratio of 6. Note that the lift coefficient is much larger than the drag coefficient so that the lift-drag ratios are in the range of about 10 to 20.

For use in fan blades? the characteristic curves have to be converted from an aspect ratio of 6 to an infinite aspect ratio. This conversion further increases the lift-

drag ratio so that the maximum lift-drag ratio will about triple. Table 2.2 shows the calculation for this conversion. Figure 2.5 shows the resulting characteristic curves for an infinite aspect ratio.

Please note the considerable difference between Figs. 2.4 and 2.5. Figure 2.5 (infi- nite aspect ratio) shows lower drag, resulting in higher lift-drag ratios.The reason for this considerable difference is turbulence at the two ends of the airfoil plate, which


FIGURE 2.3 Airfoil plate for wind tunnel test, aspect ratio of 6.


-5 0 5 10 15 20 25 30 Angle of attack a

FIGURE 2A Characteristic curves for NACA airfoil no. 6512 (size of plate: 5 x 30 in; aspect ratio: 6).

occurs only for finite aspect ratios but is eliminated for an infinite aspect ratio and

therefore also for axial-flow fans, where the fan blades are bordered by the fan hub and the housing wall so that no turbulence at the ends can develop. In designing axial-flow fans, therefore, one should use the airfoil's characteristic curves for an

infinite aspect ratio.

Looking again at the characteristic curves of the NACA airfoil no. 6512 for an infinite aspect ratio, as shown in Fig. 2.5, we note the following points:

TABLE 2.2 Calculation for Converting the Characteristic Curves from an Aspect Ratio of 6 to an Infinite Aspect Ratio: NACA Airfoil No. 6512

a6 LID c,. Aa ACo

-5 12.3 0.16 0.013 0.49 -5.5 0.001 0.012 13.3

-1 22.4 0.47 0.021 1.43 -2.4 0.012 0.009 52.2

3 17.8 0.75 0.042 2.28 0.7 0.030 0.012 62.5

7 13.8 1.01 0.073 3.07 3.9 0.054 0.019 53.2

15 9.5 1.52 0.160 4.62 10.4 0.123 0.037 41.1

23 5.9 1.74 0.295 5.29 17.7 0.161 0.134 13.0

29 3.2 1.40 0.43 4.26 24.7 0.104 0.326 4.3

Note: Aa = 18.24C,. x Vt, = 3.04C,: ou = Oft - Aa: AC„ = 0.05305 CI: 0>~ - f- fx. — AC/>.


FIGURE 2.6 At a 5° angle of attack, the airflow is smooth and follows the contours of the air- foil. The direction of the airflow is deflected by 13.5°.


1. The lift coefficient is zero at an angle of attack of about -8°. (If this were a sym-

metric airfoil, the zero lift coefficient would occur at zero angle of attack.)

2. As the angle of attack is increased, the lift coefficient rises, until it reaches a max- imum of about 1.7 at an angle of attack of about 15°. This is the top of the oper- ating range for this airfoil.

3. The lift-drag ratio has its maximum of 62.4 at an angle of attack of l°.The best operating range, then, will be at angles from 1° to about 10°, where the LID ratio is still good (between 62 and 41) and the airflow is smooth, as shown in

Fig. 2.6.

4. For angles of attack from 10° to 15°, the airflow can still follow the contour of the

airfoil, but the fan efficiency will be somewhat impaired because of the lower lift- drag ratios.

5. For angles of attack larger than 15°, the airfoil will stall, resulting in a decrease in

the lift coefficient. At these large angles of attack, the airflow can no longer fol- low the upper contour of the airfoil. It will separate from that contour, as shown in Fig. 2.7.

FIGURE 2.7 At a 16° angle of attack, the airfoil stalls, and separation of airflow takes place at the trailing edge and at the suction side of the airfoil,

with small eddies filling the suction zones. The deflection of the airflow past the trailing edge is close to zero.

0 10 8020 30 40 50 60 70

Distance from leading edge (°k of chord)

FIGURE 2.8 Dimensions of Gottingen single-thickness profile no. 417a.


Please note that angle of attack is not identical with blade angle. The blade angle

of an axial-flow fan is much larger than the angle of attack, as will be discussed in Chap. 4.


It sometimes is desirable to use single-thickness sheet metal blades rather than

airfoil blades. The reason might be a dust-laden airflow or simply lower cost. The

FIGURE 2.9 Performance of Gottingen single- thickness profile no 417a (infinite aspect ratio).


shape and characteristic curves for such a single-thickness profile are given in a

Gottingen publication. Gottingen calls it profile no. 417a. Figure 2.8 shows the

shape and Figure 2.9 shows the characteristic curves, converted to an infinite aspect

ratio. Comparing these with the shape and characteristic curves for the NACA air- foil no. 6512, as shown in Figs. 2.2 and 2.5, we find the following similarities and dif- ferences:

1. The single-thickness profile has a maximum camber of 8 percent, which is about halfway between the upper (13.3 percent at 36 percent) and lower (1.3 percent at

36 percent) cambers of NACA airfoil no. 6512. 2. The maximum camber of the single-thickness profile is located at 38 percent of

the chord from the leading edge, about the same as for NACA airfoil no. 6512. 3. The maximum lift coefficient for profile no. 417a is lower, 1.25 instead of 1.74.

4. The maximum lift-drag ratio is somewhat lower, 57 instead of 62.

5. The angle of attack at which the maximum lift coefficient occurs is considerably lower, 10° instead of 17°. This results in a narrower operating range and particu- larly in a narrower range for good lift-drag ratios. For example, the range for lift- drag ratios of 35 or more is 21° wide for NACA airfoil no. 6512 but only 6° wide for the single-thickness profile.

Despite these disadvantages, single-thickness profiles are often used in fan

blades, especially in propeller fans and in tubeaxial fans.

FIGURE 2.10 Airfoil as used in an axial-flow fan blade.



Let us now examine how the airfoil is used as the cross section of a fan blade. Figure 2.10 shows how an airfoil is used in an axial-flow fan blade. Here the concave side of the airfoil is the pressure side (just like in an airplane wing), so this is a normal con- dition.

FIGURE 2.11 Airfoil as used in a backwardly curved centrifugal fan blade. Note that here the pressure side is on the convex side of the airfoil blade.


FIGURE 2.12 Airfoil as used in a centrifugal fan with radial-tip blades, a design that is rarely used but results in good efficiencies.


Figure 2.11 shows how the airfoil is used in a centrifugal fan with backwardly curved blades. Here the convex side of the airfoil is the pressure side, wbjtfi, of

course, is abnormal. This type of blade does not really have the function of an airfoil,

since there is no lift force in the conventional sense. It is simply a backwardly curved

blade with a blunt leading edge that helps broaden the range of good efficiencies. Figure 2.12 shows how an airfoil could be used in a centrifugal fan with radial-tip

blades. Here the concave side is the pressure side, as it should be. However, this con- figuration is rarely used, partly because of higher cost and partly because radial-tip

blades are often used for handling dust-laden air and this is done better by thick, single-thickness blades.





This book will discuss the following six categories of fans:

1. Axial-flow fans

2. Centrifugal fans

3. Axial-centrifugal fans

4. Roof ventilators

5. Cross-flow blowers

6. Vortex or regenerative blowers


There arc four types of axial-flow fans. Listed in the order of increasing static pres- sure, they are

1. Propeller fans (PFs)

2. Tubeaxial fans (TAFs)

3. Vaneaxial fans (VAFs)

4. Two-stage axial-flow fans

Propeller Fans

The propeller fan, sometimes called the panel fan, is the most commonly used of all fans, it can be found in industrial, commercial, institutional, and residential applica-

tions. It can exhaust hot or contaminated air or corrosive gases from factories, weld-

ing shops, foundries, furnace rooms, laboratories, laundries, stores, or residential

attics or windows.

Sometimes several propeller fans are installed in the walls of a building, oper- ating in parallel and exhausting the air. Figure 3.1 shows two propeller fans with



FIGURE 3.1 Two 21-in propeller fans with direct drive, mounted in a wall and exhausting air from a factory building.

direct drive mounted in the wall of a factory, near the ceiling where the hot air is located.

Figures 3.2 and 3.3 show the general configuration for a propeller fan with a belt drive from an electric motor. The units consist of the following eleven components: a spun venturi housing, a bearing base (plus braces), two bearings, a shaft, a motor base, an electric motor, two pulleys, a belt, and a fan wheel. In Fig. 3.2 the fan wheel

FIGURE 3-2 A 24-in propeller fan with belt drive. Note the location of the motor opposite the rotating blades.



has four blades, and the motor is mounted on a separate, vertical motor base. In Fig. 3.3 the fan wheel has six blades, and a horizontal base supports both the motor and

the two bearings. In both figures, however, the motor is located opposite the rotating

fan blades. This results in good motor cooling but some obstruction to the airflow. Figure 3.4 shows the simpler configuration of a propeller fan with direct drive

from an electric motor. This unit consists of only four components: a spun venturi

housing, a motor base (plus braces), an electric motor, and a four-bladed fan wheel. The belt-drive arrangement has the following three advantages:

h It results in flexibility of performance, since any speed (rpms) can be obtained for the fan wheel by selection of the proper pulley ratio However, when the speed is increased to boost the flow (cfm), the brake horsepower will increase even more, as the third power of the rpm ratio, as will be explained later.

2. In large sizes, belt drive is preferable, since it will keep the speed ofthe fan wheel low or moderate while keeping the motor speed high, for lower cost. (High-speed motors are less expensive than low-speed motors of the same horsepower.)

3. The motor will get good cooling from the air stream passing over it.

The direct-drive arrangement has the following five advantages:

L It has a lower number of components, resulting in lower cost. 2. It requires no maintenance and regular checkups for adjustment of the belt.

3« It has a better fan efficiency, since a belt drive would consume an extra 10 to 15 percent of the brake horsepower.

Fan wheel


Motor pulley

FIGURE33 A 36-in propeller fan with belt drive. Note the large pulley ratio for a low speed of the fan wheel. (Courtesy of Chicago Blower Corporation, Glendale Heights, III.)


Motor base

FIGURE3A A 24-in propeller fan with direct drive. Note the height adjustment ofmotor base for an even tip clearance. -

4 It results in more flow (cfm) because the central location of the motor does not obstruct the airflow.

5. The performance flexibility of the belt-drive arrangement also can be obtained, but at an extra cost, by means of adjustable-pitch blades and by a variation in the number of blades.A 3° increase in the blade angle will result in a 10 to IS percent increase in flow (15 percent in the range of small blade angles, 10 percent for larger blade angles). The static pressure can be boosted by an increase in the number of blades, up to a point

Conclusion: Direct drive is less expensive and more efficient. It is preferable in small sizes. Beit drive is preferable in large sizes and results in better performance flexibility than direct drive, unless adjustable-pitch blades are used.

Figure 3.5 shows a 46-in propeller fan wheel of aluminum with a 13-in-diameter hub and with eight narrow airfoil blades welded to the hub.The hub-tip ratio is 0.28, a good ratio for a propeller fan. This is an efficient but expensive propeller-fan wheel. Most propeller-fan wheels have sheet metal blades riveted to a so-called spi- der, as shown in Fig. 3.6. This is a lightweight, lower-cost construction that is some- what less efficient but adequate in small and medium sizes. Many propeller-fan wheels are plastic molds. In very small sizes, where cost is more important than effi- ciency, one-piece stampings are sometimes used.

Shutters, Most propeller fans are used for exhausting from a space. They are mounted on the inside of a building, with the motor located on the inlet side, inside the building, and the air stream blowing outward.A shutter is mounted on the out- side.There are two types of shutters: automatic shutters and motorized shutters.

Figure 3.7 shows an automatic shutter having three shutter blades linked together and mounted on hinged rods.The shutter will be opened by the air stream on start-up of the fan. It will be closed by the weight of the shutter blades when the fan is turned off. The motorized shutter, used mainly in larger sizes, is opened and


FIGURE 3J Cast-aluminum propeller-fan FIGURE 3.6 A 24-in propeller-fan wheel with wheel with a 46-in o.d., 13-in hub diameter, and four wide steel blades riveted to central spider, eight narrow airfoil blades welded to hub.

closed by a separate small motor mounted on the shutter frame.When the shutter is closed, it will prevent heat losses due to backdraft and keep out wind, rain, and snow.

Screen Guards, Figure 3,8 shows a fan guard, sometimes fitted over the motor side of a propeller fan for safety,whenever the fan is installed less than 7 ft from the floor. It uses a steel mesh, designed for minimum interference with the air stream.

FIGURE 3.7 Automatic shutter having a square FIGURE 3.8 Fan guard, used for safety and frame and three shutter blades linked together for mounted on the inlet side of a propeller fan. simultaneous opening and dosing


FIGURE 3.9 Man cooler mounted on a pedestal far coaling people, products, or pro-

cesses. (Courtesy of Coppus Engineering Divi- sion, Tuthiil Corporation, Military, Mass.)

deeper penetration. Hie nozzle contains vaneaxial fan, to prevent excessive air spL

Man Coolers. Another application for propeller fans (besides exhausting from a space) is for cooling people (as the term man coolers implies) or products or for supplying cool air to certain pro-

cesses. These cooling fans are located in hot places, such as steel mills, foundries,

and forge plants. They are also used for cooling down furnaces for maintenance work, for cooling electrical equipment (such as transformers, circuit breakers,

and control panels), or for drying chem- ical coatings.

Figure 3.9 shows a man cooler mounted on a heavy pedestal for stabil- ity. It has a lug on top so that it can be moved easily to various locations. It has a 30-in propeller-fan wheel with six nar-

row blades and direct drive from a 3-hp, 1740-rpm motor.

Figure 3.10 shows a similar type of man cooler. It has a 30-in propeller-fan wheel with eight narrow blades and

direct drive from a 3-hp, 1150-rpm

motor. It has a bracket for mounting it

on a wall, up high enough so that it can- not be damaged by trucks and the wires cannot be cut by wheels. As a special feature, this unit has a conical discharge

nozzle that boosts the outlet velocity for

some straightening vanes, almost like a . at the narrow end of the cone.

FIGURE 3.10 Man cooler with a bracket for mounting it on a wall, with conical discharge nozzle for deeper penetration. (Courtesy of Bayley Fan Group, Division ofLau Industries, Lebanon, Ind.)


Tubeaxial Fans

Figure 3.11 shows a tubeaxial fan with direct drive from an electric motor. It has a

cylindrical housing and a fan wheel with a 33 percent hub-tip ratio and with ten

blades that may or may not have airfoil cross sections. The best application for tubeaxial fans is for exhausting from an inlet duct.A short outlet duct can be toler- ated, but the friction loss there will be larger than normal because of the air spin. If no inlet duct is used, a venturi inlet is needed to prevent a 10 to 15 percent loss in flow and an increased noise level. Figure 3.11 shows the motor on the inlet side, but it could be located on the outlet side as well.

In case of belt drive, the motor is located outside the cylindrical housing, and a belt guard is needed. Direct drive has fewer parts and therefore lower cost, the same as for propeller fans, and the performance flexibility again can be obtained by means of adjustable-pitch blades.

Vaneaxial Fans

Figure 3.12 shows a vaneaxial fan with belt drive from an electric motor. It has a cylindrical housing (like a tubeaxial fan) and a fan wheel with a 46 percent hub-tip ratio and with nine airfoil blades. It also has eleven guide vanes, neutralizing the air spin, so that the unit can be used for blowing (outlet duct) as well as for exhausting (inlet duct). Again, direct drive is simpler and less expensive than belt drive. Also, performance flexibility for direct drive can be obtained by means of adjustable-pitch blades. Again, a venturi inlet is needed if no inlet duct is used.

Figure 3.13 shows an axial-flow fan wheel with a 42 percent hub-tip ratio and with eight single-thickness steel blades. It could be used in a tubeaxial fan or in a vane- axial fan.

Cylindrical housing

Access door

Fan wheel



FIGURE 3.11 Tubeaxial fan with direct drive from an electric motor on the inlet side (in the background) and with a fan wheel having a 33 percent hub-tip ratio and with ten blades. (Courtesy of General Resource Corporation, Hopkins, Minn,)


Electric motor

Shaft and bearings

FIGURE 3wl2 VaneaxiaJ fan with belt drive from an electric motor and with a fan wheel having a 46 percent hub-tip ratio and nine airfoil blades. (Courtesy of General

Resource Corporation, Hopkins. Minn.)

Figure 3.14 shows a vaneaxial fan wheel with a 64 percent hub-tip ratio and with

five wide airfoil blades. This hub-tip ratio would be too large for a tubeaxial fan but

is quite common in vaneaxial fans.

Two-Stage Axial-Flow Fans

Two-stage axial-flow fans have the configuration oftwo fans in series so that the pres- sures will add up.This is an easy solution when higher static pressures are needed, but excessive tip speeds and noise levels cannot be tolerated. The two fan wheels may rotate in the same direction, with guide vanes between them. Or they may be coun- terrotating, without any guide vanes, as will be explained in more detail in Chap 4.


There are six types of centrifugal fan wheels in common use. Listed in the order of decreasing efficiency, they are

L Centrifugal fans with airfoil (AF) blades 2* Centrifugal fans with backward-curved (BC) blades


FIGURE 3.13 Axial-flow fan wheel with a 42 percent hub-tip ratio and with eight single-

thickness steel blades for use as a tubeaxial fan

wheel or as a vaneaxial fan wheel.

3. Centrifugal fans with backward-

inclined (BI) blades

4. Centrifugal fans wirh radial-tip (RT)


5. Centrifugal fans with forward-curved

(FC) blades

6. Centrifugal fans with radial blades


These six types are used in a variety of

applications, as will be discussed in more

detail in Chap. 7.

Centrifugal Fans with AF Blades

The centrifugal fan with AF blades has the best mechanical efficiency and the

lowest noise level (for comparable tip

speeds) of all centrifugal fans. Figures

3.15 and 3.16 show two constructions

for centrifugal fan wheels with AF blades. Figure 3.15 shows hollow airfoil

blades, normally used in medium and

large sizes. Figure 3.16 shows cast-

aluminum blades, which are often used

in small sizes and for testing and devel-



FIGURE 3.15 Centrifugal fan wheel, SISW, with nine hollow airfoil steel blades welded to back plate and shroud. (Courtesy of General Resource Corporation, Hopkins, Minn.)

FIGURE 3.16 Experimental centrifugal fan wheel, SISW, with eleven cast-aluminum airfoil blades welded to the back plate but not yet welded to the shroud (held above the blades).


opment work, with the shroud held above the airfoil blades prior to welding it to

the blades.

Centrifugal Fans with BC Blades

BC blades are single-thickness steel blades but otherwise are similar to AF blades with respect to construction and performance. They have slightly lower efficiencies

but can handle contaminated air streams because the single-thickness steel blades

can be made of heavier material than can be used for hollow airfoil blades.

Centrifugal Fans with Bl Blades

Figure 3. 17 shows a sketch of an SISW (single inlet, single width) centrifugal fan wheel with Bl blades. These are more economical in production, but they are somewhat lower in structural strength and efficiency. Figure 3.18 shows the same fan wheel in a scroll housing. Figure 3.19 shows a Bl centrifugal fan with scroll housing, noting the

terminology for the various components.

Incidentally, scroll housings are not always used in connection with centrifugal

fan wheels. Centrifugal fan wheels also can be used without a scroll housing, in such

applications as unhoused plug fans, multistage units, and roof ventilators. An excep- tion is FC centrifugal fan wheels. They require a scroll housing for proper function- ing, as will be explained in Chap. 7.

FIGURE 3.17 Centrifugal fan wheel. SISW. with Bl blades welded ti> back plate and shroud.

FIGURE 3.18 Angular view of centrifugal fan with BI wheel inside showing scroll housing, inlet side, and outlet side with cutoff all the way across housing


Centrifugal Fans with RT Blades

RT blades are curved, with good flow conditions at the leading edge. Only the blade tips are radial, as the term radial-tip blades indicates. Figure 3.20 shows a radial-tip

centrifugal fan wheel. These RT wheels are used mainly in large sizes, with wheel diameters from 30 to 60 in, for industrial applications, often with severe conditions

of high temperature and light concentrations of solids.

Centrifugal Fans with FC Blades

FC blades, as the name indicates, are curved forward, i.e., in the direction of the rotation. This results in very large blade angles and in flow rates that are much larger than those of any other centrifugal fan of the same size and speed. Figure 3.21 shows a typical SISW FC fan wheel, with many short blades and a flat shroud with a large inlet diameter for large flows. These fans are used in small furnaces, air

conditioners, and electronic equipment, whenever compactness is more important than efficiency.


Outlet area

FIGURE 3.19 Angular view of centrifugal fan with BI wheel inside showing scroll housing, inlet side (inlet collar and inlet cone), and outlet side with recirculation shield opposite the inlet cone only. (Courtesy ofGeneral Resource Corporation, Hopkins, Minn.)

Centrifugal Fans with Radial Blades

Radial blades (RBs) are rugged and self-cleaning, but they have comparatively low efficiencies because of the nontangential flow conditions at the leading edge. Rgure 3.22 shows an SISW KB fan wheel with a back plate but without a shroud. Some- times even the back plate is omitted (open fan wheel), and reinforcement ribs are added for rigidity.These fans can handle not only corrosive fumes but even abrasive materials from grinding operations.


These fans are also called tubular centrifugal fans, in-line centrifugal fans, or mixed-

flow fans (especially if the fan wheel has a conical back plate). The following two types of fan wheels are used in these fans:

1. A fan wheel with a flat back plate, as shown in Fig. 3.17, i.e., the same type as is used in a scroll housing. When used in an axial-centrifugal fan. h<ww *u ~


FIGURE 301 SISW centrifugal fan wheel with 52 FC blades fastened through corresponding slots in back plate and shroud.


FIGURE 3*22 SISW centrifugal fan wheel with six radial blades welded to a back plate.

stream has to make two 90° turns, which, of course, results in some extra losses, especially if the diffuser ratio (housing i.d./wheel o.d.) is small.

2. A fan wheel with a conical back plate, as shown in Fig. 3.23.This fan wheel is more expensive to build, but the air stream here has to make only two 45° turns, a more efficient arrangement. In either case, the fan wheel usually has BI blades or occa- sionallyAF orBC blades.

The following three types of housings are in common use in axial-centrifugal fans:

L A cylindrical housing, as shown in Figs. 3.23 and 3.24.


FIGURE &23 Mixed-flow fan showing direct motor drive, venturi inlet, fan wheel with conical back plate, and with a 45* diverging air stream discharging into a cylindrical housing.


FIGURE 124 Axial-centrifugal fan showing belt drive from an outside motor, venturi inlet, fan wheel with flat back plate, and outlet guide vanes, all assembled in a cylindrical housing, (Courtesy of General Resource Corporation, Hopkins, Minn.)

2. A square housing, as shown in Rg. 3.25. 3. A barrel-shaped housing, as shown in Figs. 3.26 and 3.27.

Various wheel and housing combinations are possible. Figure 3.27 shows a spe- cial type of barrel-shaped housing which is covered by my U.S. patent no. 3312,386. It has a separate chamber for the motor so that direct drive can be used, even if hot or corrosive gases are handled. This fan has the trade name Axcentrix Birurcator, implying that the air stream is divided into two forks, flowing above and below the motor chamber but never coming into contact with the motor. The various types of axial-centrifugal fans will be discussed further in Chap. 9.


Figure 3.28 shows an exhaust roof ventilator with direct drive, a BI centrifugal fan wheel, and radial discharge using various spinnings. Various other models of roof ventilators are in common use. Some may have belt drive instead of direct drive, some may have axial fan wheels instead of centrifugal fan wheels, and some may be for upblast instead of radial discharge. While most models are for exhausting air from a building,some are for supplying air into a building.The various combinations of these features lead to ten different models, which will be illustrated and described in Chap. 10.


(a) (b)

FIGURE &25 Mixed-flow fan in a square housing. Two models are shown, one for direct drive and one for belt drive. Both models have a venturi inlet, a fan wheel with a conical back plate,and an access door.The squarehousing results in lowercost and allows connection to either square or round ducts. (Courtesy ofFloAire, Inc., Bensalem, PA.)

FIGURE 326 Mixed-flow fan with barrel-shaped spun housing for smaller diameters of inlet and outlet ducts. Direct drive.The fan wheel has a conical back plate. Outlet guide vanes (not shown) prevent excessive air

spin at the small outlet diameter. (Courtesy ofFhAire. Inc., Bensalem, Pa.)


FIGURE 3.27 Mixed-flow fan with barrel shaped housing for smaller diameters of inlets and outlets.The fan wheel with conical back plate is directly driven by a motor in a separate chamber. Outlet vanes (not shown) prevent excessive outlet spin. (Courtesy ofBayley Fan, Division ofLou Industries, Lebanon, Ind)

FIGURE 3.28 Schematic sketch of a centrifugal roof exhauster, direct drive, radial discharge, 15-in wheel diameter, 1 hp, 1725 rpm. (Courtesy of Fto- Aire, Inc., Comwells Heights, Pa.)



A cross-flow blower is a unique type of centrifugal fan in which the airflow passes twice through a fan wheel with FC blading, first inward and then outward, as shown in Fig. 3.29. The main advantage of cross-flow blowers is that they can be made axi- ally wider, in fact to any width desired. This makes them particularly suitable for cer- tain applications such as air curtains, long and narrow heating or cooling coils, and

dry blowers in a car wash. The flow pattern and principle of operation will be explained in Chap. 12.

FIGURE 3.29 Cross-flow blower showing airflow passing twice through the rotating fan wheel.

Stationary housing Rot.

inlet outlet

FIGURE 3.30 Vortex blower showing blades rotating in right half of the torus-shaped housing



The vortex or regenerative blower is another unique type of centrifugal fan. Here the airflow circles around in an annular, torus-shaped space, similar to the shape of

a doughnut. On one side of the torus are rotating fan blades, throwing the air out- ward, as shown in Fig. 3.30. The airflow then is guided back inward by the other side of the torus so that it must reenter the inner portion of the rotating blades. This results in a complicated flow pattern that will be discussed in detail in Chap. 9.


From the preceding we note that there are many different types of fans. Neverthe- less, only two basic operating principles are used in all these fans: deflection of air- flow and centrifugal force.

In axial-flow fans, the operating principle is simply deflection of airflow. Here the pressure is produced exclusively by the lift of the airfoil or of the single-thickness

sheet metal profile used for the cross sections of the blades. Since an airfoil has a bet-

ter lift-drag ratio over a wider range of angles of attack than a single-thickness pro-

file (see Chap. 2), airfoil blades will result in better efficiencies than single-thickness


In centrifugal fans (including mixed-flow, cross-flow, and vortex fans), the operat- ing principle is a combination of airflow deflection plus centrifugal force. This results

in the following two differences between the performances of axial-flow fans and centrifugal fans:

1. Centrifugal fans normally produce more static pressure than axial-flow fans of the same wheel diameter and the same running speed. (Axial-flow fans, on the other hand, have the advantages of greater compactness and of easier installa- tion.)

2. Since in centrifugal fans the airfoil lift contributes only a small portion of the

pressure produced (while most of it is produced by centrifugal action), the improvement in performance due to airfoil blades (over sheet metal blades) is not as pronounced in centrifugal fans as it is in axial-flow fans.



In past years, "fans and blowers" was a common expression. Axial-flow fans were called "fans," and centrifugal fans were called "blowers." Calling centrifugal fans blowers was misleading because centrifugal fans can be used for exhausting as well as for blowing. Axial-flow fans also can be used for either blowing or exhausting.

Today the accepted terms are axial-flow fans and centrifugal fans. The term axial-flow fan indicates that the air (or gas) flows through the fan in an

approximately axial direction, as opposed to centrifugal fans (sometimes called radial-flow fans), where the air flows through the fan wheel approximately in a radi- ally outward direction.


The casual observer in general will not realize that the dimensions of an axial-flow fan (fan wheel outside diameter, hub diameter, number and width of the blades and vanes, blade and vane angles, curvature of blades and vanes, fan speed, and motor horsepower) can be calculated from the requirements (air volume in CFM and static pressure).

Let us dwell on the casual observer for a moment, and let us try to put ourselves in his or her place and look at the problem of fan design as he or she would do. The point of view of the casual observer will be about as follows: Fasten a number of blades somehow symmetrically around a hub. Put this fan wheel, together with a motor, into a housing, and run a test on it. If it does not give you "enough air," make certain changes on the unit. Make the hub diameter larger or smaller, and increase the number of blades, the blade angles, the blade widths, and the famous "scooping curvature." Keep on changing and retesting until you just about get what you want.

This viewpoint of the casual observer naturally is rather primitive. In some respects, it is even incorrect, since it overlooks certain limitations, such as the fact

that the addition of blades or an increase in blade curvature does not always result

in an increased air delivery. However, you cannot blame this casual observer if you stop and realize that many years ago this also was the conception of the fan designer. And you must admit one thing: As primitive as this purely experimental cut-and-try method may be, with a sufficient amount of persistence, time, and money, it often will be possible to obtain the desired air volume and static pressure by the application of this method. There are only three objections to this method:

1. The resulting units often will be larger, run at higher speeds, and consume more brake horsepower than necessary.

2. The method is too expensive because in general it will require the building and testing of three to five samples until the desired air volume and static pressure is obtained.



3. This method will practically always lead to units with uneven and turbulent air-

flow and with stalling effects in certain portions of the blade. As a result of all this,

these units usually will be inefficient and noisy.

Once the necessity of fan design on a theoretical basis has been recognized, the first question is: Is it at all possible to determine the dimensions for a fan unit so that

it will perform in accordance with a certain set of specifications, by pure calculation,

this way completely eliminating the use of any experimental cut-and-try method? The answer to this question is: In most cases, this is possible, and even more than that, it is possible in more than one way. In other words, several designs are possible that will meet a certain set of requirements with respect to air delivery and pressure.

This naturally leads to the next question: If this problem of meeting the require-

ments can be solved by several designs, is it possible to find one optimal design? The answer is yes, providing that a definition for the word optimal can be agreed on.

The question "What is the optimal design?" is rather complex, and the answer to it will vary with the prospective application of the fan unit. In the majority of cases,

it will include, among other things, the call for high efficiency and low sound level, both over the widest possible range of operation. Other requirements may be, for instance, a nonoverloading brake horsepower characteristic: or a flat pressure curve,

which means a large free delivery; or a steep pressure curve, which means little vari- ation in air delivery throughout the operating range; or a large pressure safety mar-

gin; or compactness; or some other supplementary requirement that may be desirable in a certain application. The combination of these requirements often results in interference problems and in conflicting specifications whose relative

importance has to be considered before a decision is made.


Let's be more specific about the statement that air flows through an axial-flow fan in an "approximately axial direction." On the inlet side, as the flow approaches the fan blades, the direction of the flow is axial, i.e., parallel to the axis of rotation, provided

there are no inlet vanes or other restrictions ahead of the fan wheel. The fan blade then deflects the airflow, as shown in Fig. 2.10 (see Chap. 2).

The operating principle of axial-flow fans is simply deflection of airflow, as explained in Chap. 2 (page 2.1 ), on airfoils, and Chap. 3 (page 3.20), on types of fans. Past the blades, therefore, the pattern of the deflected airflow is of helical shape, like

a spiral staircase. This is true for all three types of axial-flow fans: propeller fans,

tubeaxial fans, and vaneaxial fans. Accordingly, the design procedures and design calculations are similar for all three types. As to both construction and performance, however, there are some differences. We may anticipate one statement right here: The sequence propeller, tubeaxial,

and vaneaxial also indicates the general trend of increasing weight, price, hub diam- eter, static pressure, aerodynamical load, and efficiency.

Coming back to the helical pattern of the airflow past the blades of an axial-flow fan. the air velocity there can be resolved into two components: an axial velocity and a tangential or circum ferential velocity.

The axial velocity is the useful component. It moves the air to the location where we want it to go. In a propeller fan, the axial velocity moves the air across a wall or a partition. In a tubeaxial or vaneaxial fan. the axial velocity moves the air through a duct on the inlet side or on the outlet side or both.


The tangential or circumferential velocity component is an energy loss in the case of a propeller fan or a tubeaxial fan. In a vaneaxial fan, however, the tangential com- ponent is not a total loss; some of it is converted into static pressure, as will be explained later. This is the main reason why vaneaxial fans have higher efficiencies than propeller or tubeaxial fans.


For good efficiency, the airflow of an axial-flow fan should be evenly distributed over the working face of the fan wheel.To be more specific, the axial air velocity should be the same from hub (or spider) to tip.The velocity of the rotating blade, on the other hand, is far from evenly distributed: It is low near the center and increases toward the tip. This gradient should be compensated by a twist in the blade, resulting in larger blade angles near the center and smaller blade angles toward the tip.This can be seen clearly in Fig. 4.1, which shows two views of a 30-in tube axial fan wheel with a 13-in hub diameter and eight single-thickness steel blades welded to the hub. Low-cost fan wheels (especially propeller-fan wheels) sometimes do not have this variation of the blade angle from hub to tip.They sometimes have the same blade angle from hub to tip (or worse, a slightly larger blade angle at the tip). This will result in a loss of fan efficiency because most of the airflow men will be produced by the outer portion of the blades, even at tow static pressures, At higher static pressures, the blade twist is even more important, because without it, the inner portion of the blade will stall and permit reversed airflow, which, of course, will seriously affect the fan efficiency.

Thepropellerfan, as shown in Figs. 4.2 through 4.4, is the lightest, least expensive, and most commonly used fan. As mentioned, normally it is installed in a wall or in a partition to exhaust air from a building. This exhausted air, of course, has to be

FIGURE 4,1 TVvo views of a fabricated tube- FIGURE42 Propeller fan with motor on inlet axial fan wheel, 30-in o.d., 45 percent hub-tip side, 25 percent hub-tip ratio, and direct drive,

ratio, and eight die-formed steel blades, with larger blade angles at the hub and smaller blade angles at the tip.


FIGURE 43 Two views of an 18-in propeller fan with direct drive from a Vt-hp, 1150-rpm motor, 40 percent hub-tip ratio, and five cast- aluminum airfoil blades that are backswept for a lower noise level.

FIGURE 44 A 48-in propeller fan with direct drive from a 7V4-hp, 1150-rpm motor, 28 percent hub-tip ratio, and seven cast-aluminum airfoil blades.


replaced by fresh air, coming in through other openings. If these openings are large

enough, the suction pressure needed is small. The propeller fan, therefore, is

designed to operate in the range near free delivery, to move large air volumes

against low static pressures. As can be seen from Figs. 4.2 through 4.4, the unit con-

sists of a narrow mounting ring, a motor support, a motor, and a fan wheel. The

mounting ring normally has a spun inlet bell that often is extended to a square

mounting panel.The mounting panel carries some tubes or braces, which in turn sup-

port the motor base and the motor.

The motor is usually on the inlet side, as shown in Figs. 4.2 through 4.4, but in spe-

cial applications it can be on the outlet side, which means slightly less noise. If the

motor is on the inlet side, it is inside the building, protected from rain and snow by a

shutter on the outside. When the fan is in operation, the shutter will be held open by the air stream. When the fan is not running, the shutter will close and prevent any backdraft from entering the building.

Large-size propeller-fan wheels usually run at low speeds (rpms) and therefore

are belt driven. If the motor horsepower is large, good efficiency is desired, and to accomplish this, the fan wheel has a 20 to 40 percent hub-tip ratio and airfoil blades

with a twist, resulting in blade angles between 30° and 50° at the hub and between 10° and 25° at the tip. As mentioned earlier, the larger blade angles at the hub will compensate for the lower blade velocities there, and this will result in a fairly even

air velocity distribution over the face of the fan. This, as mentioned, is required for

good fan efficiency. Small propeller fans can be built with either direct drive or belt drive. Here the

motor horsepower is small, and fan efficiency, therefore, is of minor importance. Lower cost is more important. These small propeller fans therefore do not have a hub and airfoil blades. Instead, they use a so-called spider with radial extensions to which

sheet metal blades are riveted.The blade angles here are often constant from spider to blade tip. This results in most of the air being moved by the outer half of the blade.The inner portion of the blade produces mostly turbulence and, of course, consumes just as much power (or perhaps more) as it would if it were functioning properly. Sometimes there is some reverse flow near the spider when the static pressure increases. In other words, the tips of the blades move faster and produce say V2 in of static pressure, but the spider portion of the blades moves slower and may produce no static pressure. Therefore, some of the air past the blade tips will flow inward and then backward. Some small, low-cost fan wheels are stamped in one piece to keep the cost down.

The tubeaxial fan, as shown in Figs. 4.1 and 4.5, is a glorified propeller fan. It has a cylindrical housing, about one diameter long, containing a motor support, a motor, and a fan wheel.The motor can be located either upstream or downstream of the fan wheel. An upstream motor has the advantage that the airflow has a chance to smoothen out before it hits the fan blades, but this is only important if the venturi inlet is too small and therefore not effective. If the venturi inlet is adequate, the air- flow will be smooth to begin with and does not need any extra space for smoothing. The upstream motor, on the other hand, has the disadvantage that some turbulence will be produced by the motor support ahead of the fan wheel. This may affect the efficiency and will result in a somewhat increased noise level.

In general, we should keep in mind that a blade, operating in turbulent airflow, will not function properly. Turbulence past the fan wheel, therefore, is not too harm- ful. It just increases the resistance of the system and therefore the static pressure

against which the fan will operate. Turbulence ahead of the fan wheel, however, is

harmful. It not only increases the static pressure required, but it also results in the

blades operating in turbulent airflow and therefore with lower efficiency and a higher noise level.


FIGURE 4.5 Tubeaxial fan with motor on outlet side and with separate venturi inlet. 43 percent hub-

tip ratio, and direct drive.

The fan wheel of a tubeaxial fan can be similar to that of a propeller fan. It often has a medium-sized hub diameter, about 30 to 50 percent of the blade outside diam-

eter, in the case of direct drive preferably not too much different from the motor diameter, for streamline flow conditions. The unit is designed to operate in the range

of moderate static pressures, higher than for a propeller fan (due to the larger hub

diameter) but not as high as for a vaneaxial fan (due to the smaller hub diameter and

the lack of guide vanes, which in a vaneaxial fan convert some of the tangential air velocity into static pressure).

A tubeaxial fan can be connected to an inlet duct or an outlet duct or both. If there is no inlet duct, a spun venturi inlet is required, as shown in Fig. 1 . 1 3, to prevent

vena contracta, as shown in Fig. 1.11. Small tubeaxial fans usually are designed with direct drive; large units are designed with belt drive.

The vaneaxialfan, as shown in Figs. 4.6 and 4.7, is a more elaborate unit. It has the outside appearance of a cylindrical housing at least one diameter long. As in a tubeaxial fan, this housing contains the motor support, the motor, and the fan wheel.

FIGURE 4.6 Vaneaxial fan with outlet vanes around the motor and with separate venturi inlet, 66 percent hub-tip ratio, and direct drive.


[ M J

FIGURE 4.7 Vancaxial fan with inlet vanes around the motor and with separate venturi inlet, 66 percent hub-tip ratio, direct drive, and outlet diffuser and tail piece for static regain.

but the vancaxial fan housing contains in addition a set of guide vanes and some- times an inner ring, a converging tail piece, and an expanding diffuser for static regain (see Chap. Lpage 1.12, on basics).

The guide vanes usually are arranged around the motor. This makes the unit more compact, since the same axial length is used for both the motor and the guide vanes. Motor and guide vanes can be located either past the fan wheel (see Fig. 4.6) or ahead of the fan wheel (see Fig. 4.7).

Hub Diameter d of Vaneaxial Fans

The hub diameter of a vaneaxial fan is larger (than that of a tubeaxial fan), usually between 50 and 80 percent of the wheel diameter, sometimes even slightly larger. The vaneaxial fan is designed to operate in the range of fairly high static pressures, and this requires a larger hub diameter. The customer usually specifies the required air volume, static pressure, fan diameter D, and speed (rpms). In designing a vane- axial fan to meet these requirements, the first step will be to determine the hub diameter d. This can be done from the formula

where d is in inches and SP is in inches of water column. Hub diameter also can be determined from the graph shown in Fig. 4.8.

Suppose the customer requires that the vaneaxial fan should run at 1750 rpm and produce 12,(KK)cfm against 3 in of static pressure at the point of operation. Figure 4.8

indicates that the corresponding minimum hub diameter will be 18.8 in. (Please note that this is the hub diameter, regardless of the requirements for air volume and for wheel diameter D.) If for some reason a somewhat larger hub diameter d is desired, this will be acceptable. (It would merely result in a slight reduction in the annular area and therefore in the air volume, which could be compensated by a slight increase in the blade angles. It would not be critical.) A smaller hub diameter, on the other hand, could be critical. It might result in an inadequate performance of the inner blade portion, i.e., turbulence and possible reversed air flow near the hub. This inadequate performance is called stalling.

</min = (l <MKX)/rpm) \'SP (4.1)


Static pressure (in. WC)

FIGURE 4.8 Minimum hub diameter d of a vaneaxial fan as a function of speed (rpms) and static pressure.

The static pressure SP produced by a vaneaxial fan can be calculated for each radius from the following formula:

SP = 3.43 x 10"" x rpm x zB x C,_ x / x W (4.2) where SP = static pressure, in inches of water column

z B = number of blades Ci = lift coefficient of airfoil at the angle of attack, used at this radius

/ = blade width at this radius, in inches W = air velocity relative to the rotating blade, in feet per minute (fpm) (a

formula for W will be given later)

This formula indicates again that a larger hub diameter will result in a larger static pressure because, for a larger hub diameter.

1. The relative velocity W will be larger (due to the increased blade velocity). 2. There will be more room available for wider blades without overlapping.

For good efficiency, the static pressure produced should be the same for any radius from hub to tip. Since the relative air velocity W is smallest at the hub, this


must be compensated by a larger / x CL at the hub. but both / and CL can be increased only up to a certain limit (/ up to the point where the blades would overlap and CL up to the maximum lift coefficient the airfoil can produce). This is why the hub diameter

d can be slightly larger but not smaller than d = (19,000/rpm)V5P. If d were too

small, /x CL could not be made large enough to compensate for the smaller W at the hub, and stalling would occur near the hub.

Wheel Diameter D of Vaneaxial Fans

After the hub diameter d has been determined, the next step is to check whether the

wheel diameter D requested by the customer is acceptable. Obviously, it is not acceptable if it is smaller than the hub diameter d we just determined, but this is an extreme case that will happen rarely. However, even if D is larger than d, it may not be large enough.

In order to check whether the wheel diameter D requested by the customer is acceptable, we use the formula

£>min = Vd2 + 61 (cfm/rpm) (4.3)

or the graph shown in Fig. 4.9. Using the preceding values of d = 18.8 in and cfm/rpm = 12,000/1750 = 6.86, we find Dmm = 27.2 in. This is the minimum wheel diameter. It would result in a hub-tip ratio of 18.8/27.2 = 0.69, a good hub-tip ratio for a vaneaxial fan. If the customer requested a wheel diameter of 28, 29, or 30 in, we would use this size. If the customer suggested, for example, a wheel diameter of 24 in or anything

else smaller than 27 in, we would point out that this would be risky because the pres- sure safety margin would be too small and that we would prefer a larger wheel diam- eter D.

If the customer cannot accept a larger wheel diameter, a two-stage axial-flow fan

may solve the problem. Then each stage has to produce only about one-half the static pressure, and the hub diameter d as well as the wheel diameter D can be reduced. This two-stage unit will be longer and more expensive. This may or may not be acceptable to the customer. If it is not acceptable, a centrifugal fan may have to be considered instead of a vaneaxial fan.

Summarizing, we found that the hub diameter d is a function of static pressure and speed and that the wheel diameter D is a function of d and of cfm/rpm.

Vaneaxial Fans of Various Designs

Best efficiencies for vaneaxial fans are obtained with airfoil shapes as cross sec-

tions of the blades because airfoils have large lift-drag ratios (see Chap. 2, page 2.9, on airfoils). Airfoils, then, result in higher static pressures (produced by the airfoil lift) and lower power consumption (produced by the airfoil drag) and therefore higher fan efficiencies.

Airfoil blades usually are made as aluminum castings and sometimes as steel castings, incorporating both features, the twist in the blade angles and the airfoil shape in the cross sections. Figures 4. 1 0 through 4.17 show such fan wheels with cast- aluminum airfoil blades. Let us examine the different designs shown in these pic- tures.

Figures 4.10. 4.11, and 4.12 show three experimental fan wheels for the same vaneaxial fan. All three wheels have the same hub diameter and the same blade sec- tion (blade width, airfoil shape, and blade angle) at the hub, but the width at the blade

4.10 ( HAPTKR (OCR

0 10 20 30 40 50 60 70

Hub diameter d (in.)

FKJL' RE 4.9 Minimum fan-wheel diameter D of a vaneaxial fan as a function of efm, rpm, and hub diameter d.

tip is varied. Figure 4.10 shows a wide blade tip. Figure 4.1 1 shows a medium tip, and Figure 4.12 shows a narrow tip. I tested the three wheels and found the following:

1. Wide blade tips result in high pressure, high efficiency, and quiet operation, but they cause in considerable motor overload at the point of no delivery.

2. Medium tips reduce the maximum static pressure and the no-delivery overload.


FIGURE 410 A 31 -in vaneaxial fan wheel, of cast aluminum, with a 63 percent hub-tip ratio, direct drive, and seven airfoil blades with wide tips for high pressure and quiet operation at high efficiency-

FIGURE 4.11 A 31-in vaneaxial fan wheel,of cast aluminum, with a 63 percent hub-tip ratio, direct drive, and seven airfoil blades with medium-wide tips for medium pressure and leas overload at no delivery.

(a) (b)

FIGURE 4.12 A 31-in vaneaxial fan with a 63 percent hub-tip ratio, direct drive from 10-hp, 1750-rpm motor, and seven airfoil blades with narrow tips for nonoverloadiiig brake horsepower characteristics. Twejve outlet vanes go well with seven blades, (a) Inlet side, (b) Outlet side.

3. Narrow tips result in a nonoverloading brake horsepower characteristic, this was the fan wheel I adopted for production, even though the efficiency was slightly lower and the noise level was slightly higher, but still acceptable A compromise between conflicting performance features had to be made.


FIGURE 4.13 Inlet view of a 5-in vaceaxiaJ fan, of cast aluminum, with a 78 percent hub-tip ratio,

seven airfoil blades, eight outlet vanes, belt drive

from Vi*-hp shaded-pole motor, and a fan speed of 3800 rpm.

FIGURE 4.14 A 5-in vaneaxial fan for projector lamp cooling showing belt drive in assembled projector. Fan pulley is part of the fan-wheel casting.

Figures 4.13 and 4.14 are views from the inlet side of a 5-in vaneaxial fan that was designed for cooling a projector lamp, A large 78 percent hub-tip ratio was needed in order to produce sufficient static pressure to force the cooling air through some narrow passages. Not as many guide vanes are needed because of the small size. Eight vanes go well with seven blades.


FIGURE 415 A 34-in tank engine cooling fan wheel, of aluminum casting, 35 hp, 2400 rpm. View from the inlet side showing teeth inside the bub for gear drive from engine.

FIGURE 4.16 Angular side new of the same 34-in cast-aluminum tank engine cooling fan wheel showing the airfoil cross sections of the blades and the variation in the blade angles from hub to tip.

FIGURE 4.17 The same 34-in cast-aluminum tank engine cooling fan wheel, view from the outlet side showing a 63 percent hub-tip ratio

and twelve blades with blade width slightly decreasing from hub to tip.

due to the lesser rigidity of the single-thi come by an increase in the hub-tip rati

An interesting design is shown in Fig- ures 4.15, 4.16, and 4.17. These are three views of a cast-aluminum fan wheel with

airfoil blades. This 34-in fan wheel was designed to cool the engine of an Army tank. It used gear drive from the engine, as shown in Fig. 4.15, a viewfrom the inlet side. Figure 4.16 is an angular side view showing the airfoil sections of the blades and the variation of the blade angles fromhub to tip. Figure 4.17 is a view from the outlet side showing the 63 percent hub-tip ratio and the twelve blades.

Figures 4.18 through 4.20 are three

views of the 34-in fan wheel redesigned in steel, by request of the customer, to replace the cast-aluminum fan wheel. Here, some single-thickness steel blades with slots were welded to a heavy steel disk, even though the aerodynamic con- ditions of the steel disk were not ideal. Furthermore, certain other difficulties

had to be addressed, such as the lower lift coefficients of the single-thickness

blades and the risk of blade vibration ness blades. These difficulties were over- and by using more blades of narrower


width. Various shapes of welding tabs were tried out. The steel blades were subjected to vibrations in order to deter-

mine which shape would have the best resistance to fatigue failure. Figure 4.21 shows the test setup, in which six blades (differing only in the way they were attached to the heavy steel disk) were vibrated several million times by con- necting rods from solenoids. The solenoids were kept cool by air streams from a separate pressure blower.

Figure 4.22 shows another vaneaxial fan wheel with cast-aluminum airfoil blades. Figure 4.23 shows the corre- sponding vaneaxial fan housing.This unit

was designed for grain drying. It had an 18-in wheel outside diameter and direct drive from a 10-hp, 3450-rpm motor.

Single-thickness blades, as shown in Figs, 4.1, 4.18, 4.19, and 4.24 also have the desired variation in the blade angles

from hub to tip. For accuracy, uniformity, and economy of production, single-thickness steel blades are press-formed in a die mat takes care of the springback of the material.

In small sizes, the entire fan wheel can be molded in plastic, as shown in Fig, 4.25. This is a good and efficient fan wheel, even though it has only single-thickness blades, The figure shows a good blade twist from hub to tip.

FIGURE 4.18 A 34-in tank engine cooling fan wheel redesigned in steel, 35 bp, 2400 rpm. View from the inlet side showing the welds on the slots and on the tabs of the blades.


FIGURE 4J9 Angular side view of the same 34-in tank engine cooling fan wheel of welded

steel showing single-thickness steel blades welded to a heavy disk.

FIGURE 42* The same 34-in tank engine cool- ing fan wheel.View from the outlet side showing a 68 percent hub-tip ratio and 16 steel blades of constant width from hub to tip and the heavy disk with mounting holes and balancing holes.


FIGURE 421 Setup for testing single-thickness steel blades for fatigue failure due to vibration of the blades.

inl«t Bell

Figures 4.12 through 4.14 and Fig. 4.24 showed vaneaxial fans with inlet bells attached. In vaneaxial fans, the inlet bell (also called the venturi inlet) is even more important than in propeller fans or tubeaxial fans because, owing to the larger hub diameter the annular area between the hub outside diameter and the housing inside

FIGURE 4J2 Front view of an 18-in vaneaxial fan wheel, of cast aluminum, with a 48 percent hub-tip ratio, six airfoil blades with medium-wide tips for grain drying, and direct drive from 10-hp, 3450-rpm motor.


FIGURE 423 Inlet view of bousing for an 18-in vaneaxial fan, direct drive. Motor and fan wheel were removed so that

the inner ring,motor base,and outlet guide vanes can be seen.

diameter is smaller and the acceleration of the entering air stream is therefore

greater. Without an inlet bell, the vena contracta would be worse and would affect a

larger portion of the blades (to operate in turbulent air and to be starved for air),

particularly if the fan wheel is located near the housing inlet, as in Fig. 4.6.Hie use of

an inlet bell, therefore, will boost the flow rate by 10 to 15 percent. It also will

increase the fan efficiency and reduce the noise level considerably.


tivuiu ho-ui vaneaxiai tan with a 54 percent hub-tip ratio, duett drive from 26-hp, 1150-rpra motor, ten single-thickness steel blades welded to the fabricated hub, and nine outlet

vanes: (a) inlet side; (h) outlet side.


FIGURE 425 Side view of a 10-in vaneaxial fan wheel,plastic mold,with a 64 percent hub-tip ratio, seven single-thickness blades, and blade angles varying from 50° at the hub to 34° at the tip, for a 16° twist (Courtesy of Coppus Engi- neering Division, Tuthill Corporation, Millbury,


For the shape of the inlet bell, an elliptic contour is ideal, but a circular

contour is a good approximation and reduces the depth of the spinning. As mentioned in Chap. 1 (page 1.14), on

basics, the radius of curvature of the

inlet bell should be at least 14 percent of the housing inside diameter. This is

graphically shown in Fig. 4.26. For exam- ple, for a 42-in throat inside diameter (or

housing inside diameter), the radius of

curvature should be at least 57/fc in. Figures 4.27 and 4.28 show an oversize

inlet bell that was built for experimental purposestoconfirm by test that not much would be gained ifthe radius of curvature were made much larger than 14 percent of the housing inside diameter. Figure

427 shows the fan blades stationary. Fig- ure 4.28 shows a curved string being drawn by the air stream while the fan is running; note the shadow of the string, indicating that it did not touch the sur-

face of the inlet bell but followed the cur-

vature of the converging air stream.




S 7

8 4




4 *

10 20 50 60 7030 40

Throat i. d. (in.)

FIGURE 416 Recommended radius of curvature versus throat inside diameter for a venturi inlet


FIGURE 427 An oversize inlet bell shown on a vaneaaal fan with five wide-tip airfoil blades in stationary position.

Direct Drive, Belt Drive, Duct Connection

As in tubeaxial fans, direct drive generally is preferable in small vaneaxial fans so that obstructions to airflow (from the belt housing), belt losses, maintenance, and the

extra expense for bearings, brackets, and belt housings is avoided. Belt drive, on the other hand, is preferable in large sizes so that the running speed can be kept low

without the use of expensive low-speed motors. Belt drive also permits better cover-

FIGURE 428 The same inlet bell with the fas wheel running and a string beingdrawn in by the air stream entering the unit


age of the range so that requirements

for air volume and static pressure can be met more closely. Figure 4.29 shows a belt-driven vaneaxial fan.

The cylindrical shape of vaneaxial fans makes them suitable for straight- line installation. Like tubeaxial fans,

they can be connected to an inlet duct,

an outlet duct, or both.

Guide Vanes

As mentioned previously, the airflow past an axial-flow fan wheel has a helical pattern. This means that the air moves along cylindrical surfaces with practi-

cally no radial component, only an axial and a rotational or circumferential com- ponent.

It is the function of the guide vanes to eliminate or at least reduce the air spin

past the fan blades. There are two ways in which guide vanes can be provided to perform their function of reducing the

rotational energy loss: They can be located on the outlet side or on the inlet side of the fan blades. In the case of outlet vanes, the static pressure is produced partly by

the blades and partly by the vanes. In the case of inlet vanes, the vanes do not pro- duce any static pressure; they merely prepare the airflow for the blades.

In order to study the two types of guide vanes, let us make a schematic sketch of the flow pattern along a cylindrical surface for each type of vane. In order to do this, we have to "unroll" or develop this cylindrical surface into a flat plane. This is done in Fig. 4.30 for outlet vanes and in Fig. 4.31 for inlet vanes. These two figures show not only the shapes of the blades and vanes but also the different v " dties of the air flowing past the blades and vanes.

FIGURE 429 Inlet view of a 54-in vaneaxial fan (930 rpm, belt drive from 25-hp, 1750-rpm motor) with a 45 percent hub-tip ratio, six cast-

aluminum airfoil blades, and the equivalent of eleven outlet vanes. (Courtesy of Ammerman Division, General Resource Corporation, Hop- kins, Minn.)

Outlet Guide Vanes

The function of outlet vanes is easier to understand. Figure 4.30 shows how the air- flow will pass first through the rotating blade section and then through the station- ary guide vane section. The airflow approaches the blades with an air velocity Va of

cfm VQ =VB ^ : (4.4)

annular area

The airflow then gets deflected by the blades and leaves the blades with velocity V,. This velocity V

x has an axial component Va that, of course, has to be retained

for continuity (see Chap. 1, page 1.10, on basics). V] also has a rotational compo- nent Vn resulting in

V, = + (4.5)


Rotating blades



Stationary outlet vanes


l 1 i


FIGURE 430 Function of outlet vanes. They guide the helical airflow, produced by the rotating blades, back to an axial direction, thereby decelerating the aii velocity fromVx to V>

Stationary inlet vanes

1(11 ) ) ) ) )

Rotating blades


V2 =V0 FIGURE 431 Function of inlet vanes. They guide the approaching axial airflow into a helical motion, with the rotational component opposite to the fan rotation. The rotating blades then

deflect the airflow back into a more or less axial direction, thereby decelerating the ait velocity from V\ to V2.


so that V\ is about 20 to 30 percent larger than V(] at the hub. At the tip, the increase will be smaller, about 10 to 15 percent. In other words, the deflection of the airflow

will be largest at the hub, according to the following formula:

w 233 x 10* SPVr = x (4.6) rpm r

where Vr = rotational component of the helical airflow, in fpm SP= static pressure, in inches of water column

r = radius of the rotating blade section, in inches

The airflow then passes through the stationary vane section. These outlet vanes guide the air stream back into the axial direction. The air leaves the guide vane sec- tion with a velocity V2 that again must be the same as VQ for continuity. In the vane section, therefore, the airflow is decelerated from V, to Vn . Some of this difference V

x - V{) is converted to static pressure. This phenomenon of static recovery was

explained in Chap. 1 . Figure 4.24 shows a vaneaxial fan housing with the outlet guide

vanes visible.

Inlet Guide Vanes

The function of inlet vanes is different. It is illustrated in Fig. 4.31. Here, the air is first drawn through the stationary vane section. These inlet vanes are curved in such

a way that they guide the airflow into a helical motion whose rotational component is opposite to the fan rotation. This rotational component should be just sufficient to neutralize the subsequent deflection in the direction of the fan rotation that is

imposed on the airflow by the rotating blades. As a result, the air should leave the blades in an approximately axial direction. In this arrangement, the airflow

approaches the guide vanes with air velocity

V„ =Va = cfm/annular area (4.4)

The airflow then is guided into a spin by the stationary inlet vanes and leaves them with velocity

V, = WlTvl (4.5)

Again, K, is larger than Vn . Finally, as the air passes through the rotating blade sec- tion, it is deflected back into an approximately axial direction and thereby is slowed

down to V„ again, with some of the difference V x - Va converted to static pressure.

Shape of the Guide Vanes

The shape of the guide vanes can be determined from the requirement of tangential conditions at the leading edge of the outlet vanes and the trailing edge of the inlet

vanes.The width and spacing of the vanes follow considerations similar to those gen- erally used in the design of turning vanes in duct elbows. The considerations are, however, not quite the same because the airflow in elbows usually does not change its velocity (except for the direction), whereas outlet vanes operate in decelerated

airflow and inlet vanes in accelerated airflow.

Since the guide vanes are stationary, the relative air velocities and therefore the

losses here are much smaller than in the blade section. From this it might be sur-


mised that the shape of the guide vanes is not quite as critical as that of the blades.

This is actually true for the outlet vanes. The shape of inlet vanes, h