Selecting Centrifugal Pumps Data

7/30/2019 Selecting Centrifugal Pumps Data

1/92

Technical inormation

Selecting Centriugal Pumps


2/92


3/92

Copyright by

KSB Aktiengesellschat

Published by:

KSB Aktiengesellschat,

Communications (V5),

67225 Frankenthal / Germany

All rights reserved. No part o

this publication may be used,

reproduced, stored in or intro-

duced in any kind o retrieval

system or transmitted, in any

orm or by any means (electro-

nic, mechanical, photocopying,

recording or otherwise) without

the prior written permission o

the publisher.

4th completely revised and ex-

panded edition 005

Layout, drawings and

composition:

KSB Aktiengesellschat,

Media Production V5

ISBN -00-0784-4


4/92

4

Contents

1 Nomenclature ..................................................................6

2 Pump Types ................................................................89

3 Selection or Pumping Water..........................................0

3.1 Pump Data...............................................................................0

.. Pump Flow Rate ......................................................................0

.. Developed Head and Developed Pressure o the Pump .............0.. Eciency and Input Power ......................................................0

..4 Speed o Rotation ....................................................................

..5 Specic Speed and Impeller Type ..............................................

..6 Pump Characteristic Curves .....................................................

3.2 System Data .............................................................................6

.. System Head ...........................................................................6

... Bernoullis Equation .................................................................6

... Pressure Loss Due to Flow Resistances .....................................8

.... Head Loss in Straight Pipes ......................................................8

.... Head Loss in Valves and Fittings ..............................................

.. System Characteristic Curve .....................................................6

3.3 Pump Selection.........................................................................8

.. Hydraulic Aspects ....................................................................8

.. Mechanical Aspects ..................................................................9

.. Motor Selection .......................................................................9

... Determining Motor Power .......................................................9

... Motors or Seal-less Pumps ......................................................

... Starting Characteristics ............................................................

3.4 Pump Perormance and Control...............................................4

.4. Operating Point .......................................................................4

.4. Flow Control by Throttling ......................................................4

.4. Variable Speed Flow Contol .....................................................5

.4.4 Parallel Operation o Centriugal Pumps ..................................6.4.5 Series Operation .......................................................................8

.4.6 Turning Down Impellers ..........................................................8

.4.7 Under-ling o Impeller Vanes ..................................................9

.4.8 Pre-swirl Control o the Flow ...................................................9

.4.9 Flow Rate Control or Change by Blade Pitch Adjustment ........9

.4.0 Flow Control Using a Bypass ...................................................40

3.5 Suction and Inlet Conditions....................................................4

.5. The NPSH Value o the System: NPSHa ..................................4

.5.. NPSHa or Suction Lit Operation ...........................................4

.5.. NPSHa or Suction Head Operation .........................................44

.5. The NPSH Value o the Pump: NPSHr .....................................44

.5. Corrective Measures ................................................................45

3.6 Eect o Entrained Solids .........................................................47

4 Special Issues when Pumping Viscous Fluids ..................48

4.1 The Shear Curve ......................................................................48

4.2 Newtonian Fluids .....................................................................50

4.. Infuence on the Pump Characteristics ......................................50

4.. Infuence on the System Characteristics ....................................54

4.3 Non-Newtonian Fluids ............................................................54

4.. Infuence on the Pump Characteristics ......................................54

4.. Infuence on the System Characteristics ....................................55

Table o contents


5/92

5

5 Special Issues when Pumping Gas-laden Fluids ..............56

6 Special Issues When Pumping Solids-laden Fluids ..........57

6. Settling Speed ...........................................................................57

6. Infuence on the Pump Characteristics ......................................58

6. Infuence on the System Characteristics ....................................596.4 Operating Perormance ............................................................59

6.5 Stringy, Fibrous Solids ..............................................................59

7 The Periphery ................................................................6

7. Pump Installation Arrangements ..............................................6

7. Pump Intake Structures ............................................................6

7.. Pump Sump ..............................................................................6

7.. Suction Piping ..........................................................................6

7.. Intake Structures or Tubular Casing Pumps ............................64

7..4 Priming Devices .......................................................................65

7. Arrangement o Measurement Points .......................................67

7.4 Shat Couplings ........................................................................68

7.5 Pump Nozzle Loading ..............................................................69

7.6 National and International Standards and Codes .....................69

8 Calculation Examples

(or all equations numbered in bold typeace) ................7

9 Additional Literature .....................................................79

10 Technical Annex (Tables, Diagrams, Charts) .................80

Tab. : Centrigal pump classication ...................................................8

Tab. : Reerence speeds o rotation ....................................................

Tab. : Approximate average roughness height k or pipes ..................0

Tab. 4: Inside diameter d and wall thickness s in mm and weight otypical commercial steel pipes and their water content ...........0

Tab. 5: Loss coecients or various types o valves and ttings ........Tab. 6: Loss coecients in elbows and bends ....................................4

Tab. 7: Loss coecients or ttings ..............................................4/5Tab. 8: Loss coecients or adapters .................................................5

Tab. 9: Types o enclosure or electric motors to EN 60 59 and

DIN/VDE 050, Part 5 ............................................................0

Tab. 0: Permissible requency o starts Z per hour or electric motors ..0

Tab. : Starting methods or asynchronous motors ..............................

Tab. : Vapour pressure, density and kinematic viscosity o water at

saturation conditions as a unction o the temperature .............4

Tab. : Infuence o the altitude above mean sea level on the annual

average atmospheric pressure and on the corresponding

boiling point .........................................................................4

Tab. 4: Minimum values or undisturbed straight lengths o piping

at measurement points in multiples o the pipe diameter D ......67

Contents

Tables


6/92

6

A m Area

A m Distance between measuring point and pump

fange

a m, mm Width o a rectangular elbow

B m, mm Vertical distance rom suction pipe to foor

Cv gpm Flow coecient or valves, dened as the fowo water at 60 F in US gallons/minute at a

pressure drop o lb/in across the valve

cD Resistance coecient o a sphere in water fow

cT (%) Solids content in the fow

D m (mm) Outside diameter; maximum diameter

DN (mm) Nominal diameter

d m (mm) Inside diameter; minimum diameter

ds m (mm) Grain size o solids

d50 m (mm) Mean grain size o solids

F N Force

Throttling coecient o an orice

H Conversion actor or head (KSB system)

Q Conversion actor or fow rate (KSB system)

Conversion actor or eciency (KSB system)

g m/s Gravitational constant = 9.8 m/s

H m Head; discharge head

Hgeo m Geodetic head

Hs m Suction lit

Hs geo m Vertical distance between water level and pump

reerence plane or suction lit operation

Hz geo m Vertical distance between pump reerence planeand water level or positive inlet pressure

operation

HL m Head loss

H0 m Shut-o head (at Q = 0)

I A Electric current (amperage)

K Dimensionless specic speed, type number

k mm, m Mean absolute roughness

k Conversion actors kQ, kH, k (HI method)

kv m/h Metric fow actor or valves, dened as the

fow o water at 0 C in cubic metres per hour

at a pressure drop o bar

L m Length o pipe

Ls m Straight length o air-lled pipe

M Nm Moment

NPSHr m NPSH required by the pump

NPSHa m NPSH available

Ns Specic speed in US units

n min (rpm) Speed o rotation

s (rev/s)

nq min Specic speed in metric units

P kW (W) Power; input power

1

1Nomenclature

Nomenclature


7/92

7

pe Pressure in suction or inlet tank

PN (bar) Nominal pressure

p bar (Pa) Pressure rise in the pump; pressure dierential(Pa N/m)

p bar (Pa) Pressure (Pa N/m = 05 bar)

pb mbar (Pa) Atmospheric pressure (barometric)pL bar (Pa) Pressure loss

pv bar (Pa) Vapour pressure o fuid pumped

Q m/s, m/h Flow rate / capacity (also in litre/s)

qair % Air or gas content in the fuid pumped

Qo m/h Flow rate at switch-o pressure

Qon m/h Flow rate at start-up pressure

R m (mm) Radius

Re Reynolds number

S m Submergence (fuid level above pump);

immersion depth

s mm Wall thickness

s m Dierence o height between centre o pump im-

peller inlet and centre o pump suction nozzle

T Nm Torque

t C Temperature

U m Length o undisturbed fow

U m Wetted perimeter o a fow section

VB m Suction tank volume

VN m Useul volume o pump sump

v m/s Flow velocity

w m/s Settling velocity o solidsy mm Travel o gate valve; distance to wall

Z /h Switching cycle (requency o starts)

z Number o stages

zs,d m Height dierence between pump discharge and

suction nozzles

Angle o change in fow direction; opening angle Angle o inclination Loss coecient (%) Eciency

Pa s Dynamic viscosity Pipe riction actor m/s Kinematic viscosity

kg/m Density

N/m Shear stress N/m Shear stress at yield point Temperature actor; opening angle o a butter-

fy valve; cos : power actor o asynchronousmotors

Head coecient (dimensionless head generatedby impeller)

1

Indices, Subscripts

a At outlet cross-section o

the system; branching o

Bl Reerring to orice bore

d On discharge side; at dis-

charge nozzle; fowing

through

dyn Denoting dynamic com-

ponent

E At the narrowest cross-

section o valves (Table 5)

E At suction pipe or bell-

mouth inlet

e At inlet cross-section o

system, e. g. in suction

or inlet tank Reerring to carrier fuid

H Horizontal

in Reerring to inlet fow

K Reerring to curvature

L Reerring to losses

m Mean value

max Maximum value

min Minimum value

N Nominal value

opt Optimum value; at best

eciency point (BEP)P Reerring to pump

p Reerring to pressure

r Reduced, or cutdown im-

peller or impeller vanes

s On suction side; at suc-

tion nozzle

s Reerring to solids

stat Static component

sys Reerring to system /

installation

t Reerring to impeller

prior to trimming

V Vertical

w Reerring to water

z Reerring to viscous fuid

0 Basic position, reerred

to individual sphere

, , Consecutive numbers;

items

I, II Number o pumps oper-

ated

Nomenclature


8/92

8

Table 1: Centriugal pump classication

Number o stages Single stage Multistage

Shat position Horizontal Vertical Horiz. Vertic.

Casing design Radial Axial Radial Axial Stage casing

Impeller entries

Motor type, Fig. ..Dry (standardized)

motor a b c d e g h

Magnetic drive i

Submerged dry rotor

motor (See ..) j k l m

Wet rotor motor

(See ..) n o p

2Pump Types

Typical selection criteria or

centriugal pumps are their

design data (fow rate or capac-

ity Q, discharge head H, speed

o rotation n and NPSH), the

properties o the fuid pumped,

the application, the place o

installation and the applicable

regulations, specications, laws

and codes. KSB oers a broad

range o pump types to meet the

most varied requirements.

Main design eatures or classi-

cation are:

the number o stages (single-

stage / multistage),

2

the position o the shat (hori-

zontal / vertical),

the pump casing (radial, e. g.

volute casing / axial, e. g.

tubular casing),

the number o impeller entries

(single entry / double entry),

the type o motor (dry mo-

tor / dry rotor motor, e. g.

submerged motor / wet rotor

motor, e. g. canned motor,

submersible motor).

These eatures usually determine

what a pump type or series

looks like. An overview o typi-cal designs according to classi-

cation eatures is given below

(Table and Figs. a to p).

Other pump classication

eatures include:

the mode o installation, which

is dealt with in section 7.,

the nominal diameter (or thepump size, as a unction o

the fow rate),

the rated pressure (or the

wall thickness o casings and

fanges),

the temperature (or example

or the selection o cooling

equipment or shat seals),

the fuid pumped (abrasive,

aggressive, toxic fuids),

the type o impeller (radial

fow / axial fow depending on

the specic speed),

the sel-priming ability,

the casing partition, the posi-

tion o the pump nozzles, an

outer casing, etc.

a

b

Pump Types (Examples)


9/92

9

2

Fig 1 (a to p):

Centriugal pump classication

acc. to Table 1

hg

kji

ml

po

edc

n

Pump Types (Examples)


10/92

0

3 Flow Rate Head Eciency Input Power

3Selection or Pumping Water

This section applies mainly to

pumping water; the particulari-

ties o pump selection or other

media are treated in sections 4,

5 and 6.

3.1Pump Data

3.1.1Pump Flow Rate

The pump fow rate or capacity

Q is the useul volume o fuid

delivered to the pump discharge

nozzle in a unit time in m/s

(l/s and m/h are also used in

practice, as are GPM in the US).

The fow rate changes propor-

tionally to the pump speed o

rotation. Leakage fow as well

as the internal clearance fows

are not considered part o the

pump fow rate.

3.1.2Developed Head andDeveloped Pressure othe Pump

The total developed head H o

a pump is the useul mechani-

cal energy in Nm transerred

by the pump to the fow, per

weight o fuid in N, expressed

in Nm/N = m (also used to be

called metres o fuid)). The

head develops proportionally

to the square o the impellers

speed o rotation and is inde-

pendent o the density o the

fuid being pumped. A given

centriugal pump will impart the

same head H to various fuids

(with the same kinematic viscos-

ity ) regardless o their density

. This statement applies to all

centriugal pumps.

The total developed head H

maniests itsel according to

Bernoullis equation (see section

...) as

the pressure head Hp propor-

tional to the pressure dier-

ence between discharge and

suction nozzles o the pump,

the geodetic head zs,d (Figs. 8

and 9), i.e., the dierence in

height between discharge and

suction nozzles o the pump

and

the dierence o the kinetic

energy head (vd-vs

)/g be-

tween the discharge and suc-

tion nozzles o the pump.

The pressure rise p in thepump (considering the location

o the pressure measurement

taps according to section 7.!)

is determined solely by the pres-

sure head Hp along with the

fuid density according to the

equation

p = g[H-zs,d-(vd-vs)/g] (1)

where

Density o the fuid beingpumped in kg/m

g Gravitational constant

9.8 m/s

H Total developed head o the

pump in m

zs,d Height dierence between

pump discharge and suction

nozzles in m (see Figs. 8

and 9)

vd Flow velocity in the discharge

nozzle = 4 Q/pdd in m/s

vs Flow velocity in the suction

nozzle = 4 Q/pds in m/s

Q Flow rate o the pump atthe respective nozzle in m/s

d Inside diameter o the re-

spective pump nozzle in m

p Pressure rise in N/m (orconversion to bar: bar =

00 000 N/m)

High-density fuids thereore

increase the pressure rise and

the pump discharge pressure.

The discharge pressure is thesum o the pressure rise and the

inlet pressure and is limited by

the strength o the pump casing.

The eect o temperature on the

pumps strength limits must also

be considered.

3.1.3Eciency and Input Power

The input power P o a pump(also called brake horsepower)

is the mechanical power in

kW or W taken by the shat or

coupling. It is proportional to

the third power o the speed o

rotation and is given by one o

the ollowing equations:

) In the US, the corresponding units are

t-lb/lbm, i. e. oot head = oot-

pound-orce per pound mass; the

numerical value o head and specic

work are identical.


11/92

P =gQH

in W =gQH

in kW =QH

in kW 000 67

(2)

3

where

Density in kg/m

in kg/m

in kg/dm

Q Flow rate in m/s in m/s in m/h

g Gravitational constant = 9.8 m/s

H Total developed head in m

Eciency between 0 and


12/92

(IGHHEAD

IMPELLER

UPTO

-EDIUMHEAD

IMPELLER

UPTO

,OWHEAD

IMPELLER

UPTO

-IXEDFLOW

IMPELLER

UPTO

!XIALFLOW

PROPELLER

TONQ RPM

3

nq = n = n (3)

where Qopt in m/s Qopt in m

/s = Flow rate at optHopt in m Hopt in m = Developed head at optn in rpm n in rev/s = Pump speed

nq in metric units nq Dimensionless parameter

g Gravitational constant 9.8m/s

Approximate reerence values:

nq up to approx. 5 Radial high head impellerup to approx. 40 Radial medium head impeller

up to approx. 70 Radial low head impeller

up to approx. 60 Mixed fow impeller

approx. rom 40 to 400 Axial fow impeller (propeller)

Fig. 2: Eect o the specic speed nq on the design o centriugal

pump impellers. The diuser elements (casings) o single stage

pumps are outlined.

Qopt/

(Hopt/)/4

Qopt(g Hopt)/4

Specic Speed

dimensionless characteristic

parameter while retaining the

same numerical value by using

the denition in the right-hand

version o the ollowing equa-

tion []:

For multistage pumps the devel-

oped head Hopt at best eciency

or a single stage and or double-

entry impellers, the optimum

fow rate Qopt or only one im-

peller hal are to be used.

As the specic speed nq in-

creases, there is a continuous

change rom the originally

radial exits o the impellers to

mixed fow (diagonal) and

eventually axial exits (see Fig. ).

The diuser elements o radial

pump casings (e.g. volutes) be-

come more voluminous as long

as the fow can be carried o

radially. Finally only an

axial exit o the fow is possible

(e.g. as in a tubular casing).

Using Fig. it is possible to de-

termine nq graphically. Further

types o impellers are shown in

Fig. 4: Star impellers are used in

sel-priming pumps. Periph-

eral impellers extend the speci-c speed range to lower values

down to approximately nq = 5

(peripheral pumps can be de-

signed with up to three stages).

For even lower specic speeds,

rotary (or example progressive

cavity pumps with nq = 0. to )

or reciprocating positive dis-

placement pumps (piston

pumps) are to be preerred.

The value o the specic speed

is one o the infuencing para-

meters required when convert-

ing the pump characteristic

curves or pumping viscous or

solids-laden media (see sections

4 and 6).

In English-language pump lit-

erature the true dimensionless

specic speed is sometimes des-

ignated as the type number K.

In the US, the term Ns is used,

which is calculated using gal-

lons/min (GPM), eet and rpm.

The conversion actors are:

K = nq / 5.9

Ns = nq / 5.6(4)


13/92

Radial double-entry impeller*)

Star impeller or side channel pump(sel-priming)

Peripheral pump impeller or very lowspeciic speed (nq 5 to 0)

*) Plan view shown without ront shroud

Closed (shrouded) mixed low impeller *)

Open (unshrouded) mixed low impeller

Axial low propeller

Radial impeller *)

3

Fig. 3: Nomograph to determine specic speed nq (enlarged view on p. 80)Example: Qopt= 66 m

3/h = 18.3 l/s; n = 1450 rpm, Hopt= 17.5 m. Found: nq = 23 (metric units).

Fig. 4:

Impeller types or clear liquids

Specic Speed Impeller Types Characteristic Curves

3.1.6Pump Characteristic Curves

Unlike positive displacement

pumps (such as piston pumps),

centriugal pumps deliver a var-

iable fow rate Q (increasing

with decreasing head H) when

operating at constant speed.

They are thereore able to ac-

commodate changes in the

system curve (see section ..).

The input power P and hence

the eciency as well as theNPSHr (see section .5.4) are

dependent on the fow rate.


14/92

4

90

80

70

60

50

40

20

30

80

70

60

50

40

0

5

10

20

30

100 20 40 60 80 100 120

Flow rate Q [m3/h] Flow rate Q [m3/h] Flow rate Q [m3/h]

140 160 0 100 200 300 400 500 0 500 15001000 2000 2500 3000550

HeadH[m]

NPSHr[m]

PowerP[kW]

EfficiencyH[

%]

2422

18

14

10

20

16

12

86

90

30

80

70

60

50

40

0

5

1510

15

14

16

17

13

HeadH[m]

NPSHr[m]

PowerP[kW]

EfficiencyH[

%]

2

4

18

14

10

20

16

12

8

6

90

30

80

70

60

50

40

5

15

10

60

40

20

80

100

0

HeadH[m]

NPSHr[m]

PowerP[kW]

EfficiencyH[

%]

n = 2900 min1 n = 1450 min1 n = 980 min1

Operating limit

> L V

11OPT

H

HOPT

11OP

0

0OPT

300

25

11OPT

(

(OPT

Operating limit forlow input power

for highinput power

25

25

300

300

150

150

70

70

40

40

11OP

NPSHrNPSHr opt

25

25

300

300

150

70

40

150

300

7040

25

3

Fig. 5: Eect o specic speed nq on centriugal pump characteristic

curves (Not drawn to scale! For NPSHr, see section 3.5.4).

Fig. 6: Three examples o characteristic curves or pumps o diering specic speeds.

a: radial impeller, nq 20; b: mixed fow impeller, nq 80; c: axial fow impeller, nq 200.(For NPSH

rsee section 3.5.4)

Characteristic Curves

The relationship o these val-

ues is shown graphically in the

pump characteristic curves,

whose shape is infuenced by

the specic speed nq and which

document the perormance o acentriugal pump (see Fig. 5 or

a comparison o characteristics

and Fig. 6 or examples). The

head curve o the pump is also

reerred to as the H/Q curve.

The H/Q curve can be steep or

fat. For a steep curve, the fow

rate Q changes less or a given

change o developed head H

than or a fat curve (Fig. 7).This can be advantageous when

controlling the fow rate.


15/92

5

(EAD

(

1STEEP

1FLAT

&LOWRATE1

5NSTABLE

REGION

3TEEPCHARACTERISTICCURVE

&LATCHARACTERISTICCURVE

-AXIMUM

(

Fig. 7: Steep, fat or unstable characteristic curve

3Characteristic Curves

H/Q characteristics normally

have a stable curve, which

means that the developed head

alls as the fow rate Q in-

creases. For low specic speeds,

the head H may in the low

fow range drop as the fow

rate Q decreases, i. e., the curve

is unstable (shown by the dash

line in Fig. 7). This type o

pump characteristic curve need

only be avoided when two

intersections with the system

curve could result, in particularwhen the pump is to be used or

parallel operation at low fow

rates (see section .4.4) or when

it is pumping into a vessel which

can store energy (accumulator

lled with gas or steam). In all

other cases the unstable curve is

just as good as the stable charac-

teristic.

Unless noted otherwise, thecharacteristic curves apply or

the density and the kinematic

viscosity o cold, deaerated

water.


16/92

6

A VA

EVE

EVE

VD

VS

AVA

(GEO

(SGEO

ZSD

A VA

PA PA

! " #

PE

$ %

3

Fig. 8: Centriugal pump system with variously designed vessels in suction lit operation

A = Open tank with pipe ending below the water level

B = Closed pressure vessel with ree fow rom the pipe ending above the water level

C = Closed pressure vessel with pipe ending below the water level

D = Open suction/inlet tank

E = Closed suction/inlet tank

va and ve are the (usually negligible) fow velocities at position a in tanks A and C and at position e

in tanks D and E. In case B, va is the non-negligible exit velocity rom the pipe end at a .

System Head Bernoulli

3.2System Data

3.2.1System Head

3.2.1.1Bernoullis Equation

Bernoullis equation expresses

the equivalence o energy in

geodetic (potential) energy,

static pressure and kinetic

energy orm. The system head

Hsys or an assumed rictionless,

inviscid fow is composed o the

ollowing three parts (see Figs.

8 and 9):

Hgeo (geodetic head) is the

dierence in height between

the liquid level on the inlet

and discharge sides. I the dis-

charge pipe ends above the

liquid level, the centre o the

exit plane is used as reerence

or the height (see Figs 8B and

9B).

(pa - pe)/( g) is the pressure

head dierence between the

inlet and outlet tank, applic-


17/92

7

AVA

E VEE VE

VD

VS

A VA

(GEO

(ZGEO

ZSD

A VA

PA PA

! " #

PE

$ %

3

Fig. 9: Centriugal pump system with variously designed vessels in suction head (positive inlet pressure)

operation. Legend as in Fig. 8.

System Head Bernoulli

able when at least one o the

tanks is closed as or B, C or

E (see Figs. 8B, C, E, 9B, C,

E).

(va-ve

)/g is the dierence in

the velocity heads between the

tanks.

For a physically real fow, the

riction losses (pressure head

losses) must be added to these

components:

HL is the sum o the head

losses (fow resistance in the

piping, valves, ttings, etc in

the suction and discharge lines

as well as the entrance and exit

losses, see section ...), and

is reerred to as the system pres-

sure loss.

The sum o all our componentsyields the system head Hsys:

Hsys = Hgeo + (pape)/( g) + (va-ve)/g + HL (5)

where

all the heads H are in m,

all the pressures p are in Pa ( bar = 00 000 Pa),

all velocities v are in m/s,

the density is in kg/m,

the gravitational constant is g = 9.8 m/s

.


18/92

8

Fig. 10: Pipe riction actor as a unction o the Reynolds number Re and the relative roughness d/k(enlarged view on p. 81)

3 System Head Pressure Loss Head Loss

DK

&ULLYROUGHK

LAMINAR TURBULENT

2ECRIT

,IMITINGCURVE

(YDRAULICALLYSMOOTHK

L

2E

2EYNOLDSNUMBER2E

0IPE

FRIC

TION

FACTORL

The dierence o the velocity

heads can oten be neglected in

practice. When at least one tank

is closed as or B, C or E (see

Figs. 8B, C, E, 9B, C, E), Eq. 5

can be simplied as

HsysHgeo+(pape)/(g)+HL (6)

and urther simplied when

both tanks are open as or A

and D (see Figs. 8A, D and 9A,

D) as

HsysHgeo+HL (7)

3.2.1.2Pressure Loss Due to FlowResistances

The pressure loss pL is caused

by wall riction in the pipes and

fow resistances in valves, t-tings, etc. It can be calculated

rom the head loss HL, which is

independent o the density ,

using the ollowing equation:

pL = g HL (8)

where

Density in kg/m


9.8 m/s

HL Head loss in m

pL Pressure loss in Pa

( bar = 00 000 Pa)

3.2.1.2.1Head Loss in Straight Pipes

The head loss or fow in

straight pipes with circular

cross-sections is given in general

by

HL = L

v (9) d g

where

Pipe riction actor accordingto Eqs. () to (4)

L Length o pipe in m

d Pipe inside diameter in m

v Flow velocity in m/s(= 4Q/pd or Q in m/s)


9.8 m/s


19/92

9

3Head Loss in Straight Pipes

For pipes with non-circular

cross-sections the ollowing

applies:

d = 4A/U (10)

where

A Cross-sectional fow area

in m

U Wetted perimeter o the

cross-section A in m; or

open channels the ree fuid

surace is not counted as part

o the perimeter.

Recommended fow velocities

or cold waterInlet piping 0.7.5 m/sDischarge piping .0.0 m/sor hot water

Inlet piping 0.5.0 m/sDischarge piping .5.5 m/sThe pipe riction actor hasbeen determined experimentally

and is shown in Fig. 0. It varies

with the fow conditions o the

liquid and the relative rough-ness d/k o the pipe surace. The

fow conditions are expressed

according to the anity laws

(dimensional analysis) using the

Reynolds number Re. For cir-

cular pipes, this is:

Re = v d/ (11)

where

v Flow velocity in m/s

(= 4Q/pd or Q in m/s)d Pipe inside diameter in m

Kinematic viscosity in m/s

(or water at 0 C exactly

.00 (0)6 m/s).

For non-circular pipes, Eq. 0 is

to be applied or determining d.

For hydraulically smooth pipes

(or example drawn steel tubing

or plastic pipes made o poly-

ethylene (PE) or polyvinyl chlor-

ide (PVC)) or or laminar fow,

can be calculated:

In the laminar fow region

(Re 0)

the test results can be repre-

sented by the ollowing empiri-

cal relationship dened by Eck

(up to Re < 08 the errors are

smaller than %):

=0.09

(lgRe

)(13)

7

In Fig. 0 it can be seen that

the pipe riction actor depends

on another dimensionless para-

meter, the relative roughness o

the pipe inner surace d/k; k is

the average absolute roughnesso the pipe inner surace, or

which approximate values are

given in Table . Note: both d

and k must be expressed in the

same units, or example mm!

As shown in Fig. 0, above a

limiting curve, is dependentonly on the relative roughness

d/k. The ollowing empirical

equation by Moody can be used

in this region:

=0.0055+0.5/ (d/k) (14)

For practical use, the head

losses HL per 00 m o straight

steel pipe are shown in Fig.

as a unction o the fow rate Q

and pipe inside diameter d. The

values are valid only or cold,

clean water or or fuids with

the same kinematic viscosity, or

completely lled pipes and or an

absolute roughness o the pipe

inner surace o k = 0.05 mm,

i.e., or new seamless or longi-tudinally welded pipes. (For

the pipe inside diameters, see

Table 4).

The eect o an increased

surace roughness k will be de-

monstrated in the ollowing

or a requently used region in

Fig. (nominal diameter 50

to 00 mm, fow velocity 0.8

to .0 m/s). The dark-shadedregion in Fig. 11 corresponds

to the similarly marked region

in Fig. 10 or an absolute rough-

ness k = 0.05 mm. For a rough-

ness increased by a actor 6

(slightly incrusted old steel pipe

with k = 0.0 mm), the pipe

riction actor(proportionalto the head loss HL) in the

lightly shaded region in Fig. 0

is only 25% to 60% higher thanbeore.

For sewage pipes the increased

roughness caused by soiling

must be taken into considera-

tion (see section .6). For pipes

with a large degree o incrusta-

tion, the actual head loss can

only be determined experimen-

tally. Deviations rom the no-

minal diameter change the headloss considerably, since the pipe

inside diameter enters Eq. (9) to

the 5th power! (For example, a

5% reduction in the inside dia-

meter changes the head loss by

0%). Thereore the nominal

diameter may not be used as

the pipe inside diameter or the

calculations!


20/92

0

*i>i> `vii

-ii i] i>i >V`Vi>i` }>>i`

}` >i`i`] Li` }>>i` Viii`

ii`

i`] `i>i }V> i>V> >viVi>}

LiVii ii>V> >>}i i

Vii i] vi` v-VVii i] vi` v,ivVi VVii i] vVVii i`] v

i>i `>>]>V,LLiL} i] iLi`7` i >vi}ii>i>

x x x x {

x x x x

3

Table 4: Inside diameter d and wall thickness s in mm and weight o typical commercial steel pipes and their

water content in kg/m to ENV 10 220 (ormerly DIN ISO 4200). D = outside diameter, s = wall thickness

All dimensions in mm Seamless pipe Welded pipeSeamless Welded weight in kg/m weight in kg/m

DN D s * d s ** d Pipe Water Pipe Water

5 . .0 7. .8 7.7 0.95 0.5 0.866 0.460 6.9 .0 .9 .8 . . 0.4 . 0.465 .7 . 9. .0 9.7 .78 0.665 .56 0.69 4.4 .6 7. . 7.8 .55 .09 .7 .40 48. .6 4. . 4.7 .9 .46 .6 .5050 60. .9 54.5 . 55.7 4. . .9 .4465 76. .9 70. .6 70.9 4.7 .88 5.4 .9580 88.9 . 8.5 .9 8. 6.76 5.4 6.5 5.4

00 4. .6 07. . 07.9 9.8 9.00 8.77 9.45 9.7 4.0 .7 .6 .5 .4 .6 . .850 68. 4.5 59. 4.0 60. 8. 9.9 6. 0.00 9. 6. 06.5 4.5 0. . .5 .8 4.750 7.0 6. 60.4 5.0 6.0 4.4 5. .0 54.00 .9 7. 09.7 5.6 .7 55.5 75. 44.0 76.850 55.6 8.0 9.6 5.6 44.4 68.6 90.5 48. 9.400 406.4 8.8 88.8 6. 9.8 86. 8.7 6. .7500 508.0 .0 486.0 6. 495.4 5 85.4 77.9 9.7600 60.0 .5 585.0 6. 597.4 84 68.6 9.8 80.

* above nominal diameter DN identical to DIN 448 ** above nominal diameter DN 5 identical to DIN 458

Table 3: Approximate average roughness height k (absolute rough-

ness) or pipes

Head Loss in Straight Pipes Dimensions and Weights o Steel Pipes


21/92


22/92

3

Fig. 13: Schematic representation o the valve designs listed in

Table 5

4 5

6 7 8 9 0

4 5

6 7 8 9

Head Loss in Straight Pipes Valves and Fittings

The head losses HL in plastic

(or example PE or PVC) pipes

or smooth drawn metal pip-

ing are very low thanks to the

smooth pipe surace. They are

shown in Fig. and valid orwater at 0 C. At other tem-

peratures, the loss or plastic

pipes must be multiplied with

a temperature correction actor

indicated in Fig. to account

or their large thermal expan-

sion. For sewage or other un-

treated water, an additional

00% head loss should betaken into consideration or po-tential deposits (see section .6).

3.2.1.2.2Head Loss in Valves andFittings

The head loss in valves and t-

tings is given by

HL = v/g (15)

where

Loss coecientv Flow velocity in a charac-

teristic cross-section A (or

example the fange) in m/s


9.8 m/s

Tables 5 to 8 and Figures to5 contain inormation about

the various loss coecients orvalves and ttings or operation

with cold water.

The minimum and maximum

in Table 5 bracket the values

given in the most important

technical literature and apply to

valves which have a steady ap-

proach fow and which are ully

open. The losses attributable to

straightening o the fow dis-

turbances over a length o pipe

equivalent to x DN down-

stream o the valve are included

in the value in accordancewith VDI/VDE 7 guidelines.

Depending on the inlet and

exit fow conditions, the valve

models used and the develop-

ment objectives (i.e. inexpensivevs. energy-saving valves), the

loss values can vary dramatically.


23/92


24/92

4

/iZ>ivi}iciL`Li`Li`]LLii`Liv>V`V>i`L>iiiV>i`LiVL>viL>i`

/>Li\VivvViZvv}

L>viLi`-iL\

>\iii>}ìi Zzi>LiLi` Zzni>i`Vi>LiLi` ZziiVi>LiLi`Zz{

iiv}\

ii`}i-> Zz x vDrxc c {xc

>vii` Zz x xx x Zz n

n{

D

V>}iiv}\

Zzì>v>>ì>ii}v>}i>> >iviV LiiViVZz v>ii}>iV`L]vi>ii`>i ì>v>iv}>>i]iV

iì>}i

A xc c {xc c c

-v>Vi -v>Vi -v>Vi -v>Vi -v>Vi

} } } } }

Zv ,r { x x x x Zv ,r` q q { { { x

Zv ,r` q q {

Zv ,x` q q n

LivVVvii>i` q q q q q q q

Z q q q q x q q x q

,ìL

7iì`Li`

,A

A

`

Table 6: Loss coecients in elbows and bends

3 Head Loss in Valves and Fittings Loss Coecients or Fittings

Note: For the branch ttings

in Table 7 and the adapters o

Table 8, one must dierentiate

between the irreversible pressure

loss (reduction in pressure)

pL = v/ (16)

where

pL Pressure loss in Pa

Loss coecient Density in kg/m

v Flow velocity in m/s

and the reversible pressure

change o the rictionless fow

according to Bernoullis equa-

tion (see ...):

pp = (vv)/ (17)

For accelerated fows (or ex-

ample a reduction in the pipe

diameter), pp is alwaysnegative, or decelerated fows

(e.g. pipe expansion) it is always

positive. When calculating the

net pressure change as the arith-

metic sum o pL and pp, thepressure losses rom Eq. 6 are

always to be subtracted.

Oten the so-called kv value is

used instead o the loss coe-


25/92

5

A

A

2+

2+

,OSSC

OEFFICIENTZ

%LBOWRADIUS2+$UCTWIDTHA

2ADIUSONINNERCORNER7ITHTURNING

VANECASCADE

2ADIUSONOUTERCORNER

3

Flow meters:

Short Venturi tube = 0

is reerred to the velocity v at diameter D.

Diameter ratio d/D = 0.0 0.40 0.50 0.60 0.70 0.80

Area ratio m = (d/D) = 0.09 0.6 0.5 0.6 0.49 0.64

Short Venturi tube 6 0.7 0. 0.Standard orice 00 85 0 4.5

Water meters (volume meters) 0For domestic water meters, a max. pressure drop o bar is specied or the

rated load. In practice, the actual pressure loss is seldom higher.

Branch ttings (o equal diameter)

Note:

The loss coecients a or the branched-o fow Qa or d or the main fowQd = Q Qa reer to the velocity o the total fow Q in the branch. On the

basis o this denition, a or d may have negative values; in this case, theyare indicative o a pressure gain instead o a pressure loss. This is not to be

conused with the reversible pressure changes according to Bernoullis equa-

tion (see notes to Tables 7 and 8).Qa/Q = 0. 0.4 0.6 0.8

a 0.4 0.08 0.47 0.7 0.9 d 0.7 0.0 0.4 0.5

a 0.88 0.89 0.95 .0 .8 d 0.08 0.05 0.07 0.

a 0.8 0 0. 0.7 0.7 d 0.7 0.9 0.09 0.7

a 0.68 0.50 0.8 0.5 0.48 d 0.06 0.04 0.07 0.0

Table 8: Loss coecients or adapters

Expansion Contraction

Type I II III IV

Type d/D 0.5 0.6 0.7 0.8 0.9

I 0.56 0.4 0.6 0. 0.04 = 8 0.07 0.05 0.0 0.0 0.0II or = 5 0.5 0. 0.07 0.0 0.0 = 0 0. 0.7 0. 0.05 0.0III 4.80 .0 0.88 0.4 0.IV or 0 < < 40 0. 0.0 0.05 0.0 0.0

Ddv Dd

v D dv D d

v

Fig. 14: Eect o rounding o

the inner and outer side o el-

bows in square ducts on the loss

coecient

Head Loss in Valves and Fittings Loss Coecients or Fittings and Flow Meters

D d D D

Standard oriice

Dv

dv

Qd

Qd

Qd

Qa

Q

Qa

Q

Qd

Qa

Q45

45

Qa

Q

cient when calculating thepressure loss or water in valves:

pL = (Q / kv) . /000 (18)

where

Q Volume rate o fow in m/h (!)

Density o water in kg/m

pL Pressure loss in bar (!)

The kv value is the fow rate in

m/h which would result rom a

pressure drop o bar through

the valve or cold water. It cor-

relates with the pressure loss pL

in bar with the fow rate Q in

m/h. The notation kvs is usedor a ully open valve.

Conversion or cold water:

6 d4/kv (19)

where

d Reerence (nominal) diameter

o the valve in cm (!)

Table 7 (continued)


26/92

6

3YSTEMH

EAD(SYS

3YSTEMHEADCHARACTERISTIC( SYS$YNAMICCOMPONENT(,

VAnVE

G

3TATICCOMPONENT(GEOPAnPEqG

Fig. 16: System characteristic curve Hsys with static and dynamic

components

3

3.2.2System Characteristic Curve

The system characteristic curve

plots the head Hsys required by

the system as a unction o the

fow rate Q. It is composed o

the so-called static and

dynamic components (see

Fig. 6).

The static component consists

o the geodetic head Hgeo andthe pressure head dierence

(pa-pe)/(g) between the inlet

J

J

J

V V

Y

V

10

A

,OSS

COEFFICIENTZ

2ELATIVEOPENINGANGLEJnJJ Degree of opening y/a or relative lift y/DN

Fig. 15:

Loss coecients

o butterfyvalves, globe

valves and gate

valves as aunction o the

opening angle

or degree o

opening (The

numbers desig-

nate the types

illustrated in

Fig. 13)

Head Loss in Valves System Characteristic Curve

One must be careul to distinguishbetween this use o static anddynamic components and the pre-cisely dened static head and dy-namic head used in fuid dynamics,since the dynamic component othe system head curve consists o bothstatic head (i.e. pressure losses)and dynamic head (i.e. velocity or

kinetic energy head).


27/92

7

1M

H

1LS

(

M

n

n

n

n

n n n n n

n

n n

n

n

n n n

n n

n

n n n

nn

n

n

3

and outlet tanks, which are in-

dependent o the fow rate. The

pressure head dierence is zero

when both tanks are open to the

atmosphere.

The dynamic component con-

sists o the head loss HL, which

increases as the square o the

fow rate Q (see section ...),

and o the change in velocity

head (va-ve)/g between theinlet and outlet cross-sections o

the system. Two points are su-

cient to calculate this parabola,

one at Q = 0 and one at any

point Q > 0.

For pipe systems connected

one ater the other (series con-

nection) the individual system

curves Hsys, Hsys etc. are

plotted as unctions o Q, and

Fig. 17: Selection chart or a volute casing pump series or n = 2900 rpm

(First number = nominal diameter o the discharge nozzle, second number = nominal impeller diameter)

the heads or each fow rate are

added to obtain the total system

curve Hsys = (Q).

For branched piping systems the

system curves Hsys, Hsys, etc.

o the individual branches be-

tween the fow dividers are each

calculated as unctions o Q.

The fow rates Q, Q, etc. o

all branches in parallel or each

given head Hsys are then addedto determine the total system

curve Hsys = (Q) or all the

branches together. The sections

beore and ater the fow

dividers must be added as or a

series connection.

System Characteristic Curve Selection Chart


28/92

8

100

50

40

30

20

10

61 2

0.3 0.4 0.5 1 2

3 4 5 10 20Q m3/h

Q l/s

3

10987

65

4

3

2

987

6

5

4

3

2

76

5

3

10

4

2

3

4

2

Pump size 1 Pump size 2 Pump size 3 Pump size 4

H

m

3 4 5

,S

MH

&LOWRATE

MH

K7

(EAD

M

M

.03(

0OWER

MM)MPELLER

H

3

3.3Pump Selection

3.3.1Hydraulic Aspects

The data required or selecting

a pump size, i.e. the fow rate Q

and the head H o the desired

operating point are assumed

to be known rom the system

characteristic curve; the electric

mains requency is also given.

With these values it is possible

to choose the pump size, the

speed o rotation and, i neces-

sary, the number o stages, rom

the selection chart in the salesliterature (see Figs. 7 and 9).

Further details o the chosen

pump such as the eciency ,the input power P, the required

NPSHr (see section .5.4) and

the reduced impeller diameter

Dr can then be determined rom

Fig. 19: Selection chart or a multistage pump series or n = 2900 rpm

Fig. 18: Complete characteristics

o a centriugal pump

Hydraulic Aspects o Pump Selection


29/92

9

3

the individual characteristic

curve (or example see Fig. 8).

I there are no specic reasons

or doing otherwise, a pump

should be selected so that the

operating point lies near its

best eciency point Qopt (=

fow rate at which eciency is

highest, BEP). The limits Qmin

and Qmax (or example due to

vibration behaviour, noise emis-

sion as well as radial and axial

orces) are given in the product

literature or can be determined

by inquiry [].

To conclude the selection, theNPSH conditions must be

checked as described in section

.5.

A multistage pump is chosen us-

ing the same general procedure;

its selection chart shows the

number o stages in addition to

the pump size (Fig. 9).

For pumps operating in series

(one ater the other) the devel-oped heads H, H, etc. o the

individual characteristic curves

must be added (ater subtracting

any head losses which occur be-

tween them) to obtain the total

characteristic H = (Q).

For pumps operating in parallel,

the individual characteristics H,

H, etc. = (Q) are rst reduced

by the head losses occurring up

to the common node (head loss

HL calculation according to sec-

tion ...) and plotted versus

Q. Then the fow rates Q o

the reduced characteristics are

added to produce the eective

characteristic curve o a vir-

tual pump. This characteristic

interacts with the system curve

Hsys or the rest o the system

through the common node.

3.3.2Mechanical Aspects

When selecting a pump the me-

chanical aspects require atten-

tion in addition to the hydrau-

lics. Several examples are:

the eects o the maximum

discharge pressure and tem-

perature o the fuid pumpedon the operating limits,

the choice o the best shat

sealing method and cooling

requirements,

the vibration and noise emis-

sions,

the choice o the materials o

construction to avoid corro-

sion and wear while keeping

in mind their strength andtemperature limits.

These and other similar re-

quirements are oten specic to

certain industries and even to

individual customers and must

be addressed using the product

literature [] or by consulting

the design department.

3.3.3Motor Selection

3.3.3.1Determining Motor Power

Operation o a centriugal pump

is subject to deviations rom

rated speed and fuctuations

in the fow volume handled,

and, consequently, changes in

the operating point (see section

.4.). In particular i steep

power curves are involved (see

Figs. 5 and 6), this may result

in a higher required pump input

power P than originally speci-

ed. For practical purposes, a

saety allowance is thereore

added when the appropriate

motor size is selected. Saety

allowances may be speciedby the purchaser, or laid down

in technical codes, see Fig. 0.

The saety allowances stipulated

by individual associations are

shown in the relevant type series

literature [] or the customers

specication.

When energy-saving control

methods are used (e. g., speed

control systems), the maximum

Fig. 20: Drive power as a unction o rated pump input power at the

operationg point

Example as per ISO 9905, 5199 and 9908 (Class I, II and III)

K7

Drivepowerrelativetopumpinpu

tpowerunderrated

conditionsin%

0UMPINPUTPOWERUNDERRATEDCONDITIONS

Hydraulic Aspects o Pump Selection Motor Selection


30/92

0

3

power peaks which may pos-

sibly occur must be taken into

account.

I a pump is selected or a

product with a density lower

than that o water, the motor

power required may have to be

determined on the basis o the

density o water (or example,

during the perormance test or

acceptance test in the test bay).

Typical eciencies and poweractors cos o standardized IP54 motors at 50 Hz are shown

in Fig. , and the curves o e-

ciency and power actorcos as a unction o relativemotor load P/PN in Fig. .

Table 9 lists types o enclosure

that provide protection o elec-

tric motors against ingress o

oreign objects or water, and

o persons against accidental

contact.

The specic heat build-up in

both electric motors and fexi-

ble couplings during start-up

as well as the risk o premature

contactor wear limit the requen-

cy o starts. Reerence values

or the maximum permissible

number o starts Z are given in

table 0, unless otherwise speci-

ed.

Submersible motor pumps (Figs.

j to m) are ready-assembled

pump units whose motors need

not be selected individually [7].

Their electrical characteristics

are given in the type series

literature. The motor is lled

with air and can be operated

submerged in the product han-

dled thanks to a in most cases

double-acting shat seal with a

paran oil barrier.

Table 9: Types o enclosure or electric motors to EN 60 59 andDIN/VDE 050, Part 5

The type o protective enclosure is indicated by the IP code as ollows:Code letters (International Protection) IPFirst digit (0 to 6 or X i not applicable) XSecond digit (0 to 8 or X i not applicable) X

Alternatively letters A, B, C, D and H, M, S, W or special purposes only.Key to Protection o electrical Protection o persons againstdigits: equipment against ingress o accidental contact by

solid objects

First 0 (not protected) (not protected)digit > 50 mm in dia. back o the hand

> .5 mm in dia. nger > .5 mm in dia. tool4 > .0 mm in dia. wire5 protected against dust (limited wire

ingress permitted, no harmuldeposits)

6 totally protected against dust wire

Protection against ingress o water with harmul consequences

Second 0 (not protected)digit vertical dripwater

dripwater up to 5 rom the vertical sprays (60 rom the vertical)4 sprays (all directions)5 low-pressure jets o water6 strong jets o water (heavy sea)7 temporary fooding8 permanent fooding

K7

H

COSJ

2ATEDPOWER0.

%FFICIENCYH

0OWERFACTORCOSJ

POLES

POLES

Fig. 21: Typical eciencies and power actors cos o standard-ized motors, IP 54 enclosure, at 50 Hz as a unction o motor

power PN

Table 0: Permissible requency o starts Z per hour or electric motors

Motor installation Dry Wet (submersible motors)

Motors up to 4 kW 5 0Motors up to 7.5 kW 5 5Motors up to kW 5Motors up to 0 kW 0Motors above 0 kW 0 0

Motor Selection


31/92

3

K7

H

K7

K7

K7]

COSJ

%FFICIENCYH

0OWERFACTORCOSJ

2ATEDPOWER00.

POLESPOLES

Fig. 22: Curve o eciency and power actor cos o standardizedIP 54 motors plotted over relative motor power P/PN

Submersible borehole pumps,

which are mostly used or ex-

tracting water rom wells, are

another type o ready-assembled

units whose motors need not be

selected individually (Fig. p).

On these pumps, the rotor and

the windings are immersed in

water [7]. Their electrical char-acteristics and permissible re-

quency o starts are indicated in

the type series literature [].

3.3.3.2Motors or Seal-less Pumps

Seal-less pumps are requently

used or handling aggressive,

toxic, highly volatile or valu-

able fuids in the chemical and

petrochemical industries. They

include magnetic-drive pumps

(Fig. ) and canned motor

pumps (Figs. n and o). A

mag-drive pump is driven by a

primary magnetic eld rotating

outside its fameproo enclosure

and running in synchronization

with the secondary magnets in-

side the enclosure [].

The primary component in

turn is coupled to a commercial

dry driver. The impeller o a

canned motor pump is mounted

directly on the motor shat, so

that the rotor is surrounded by

the fuid pumped. It is separated

rom the stator windings by the

can [7].

Seal-less pump sets are generally

selected with the help o compu-

terized selection programs, tak-

ing into account the ollowing

considerations:

The rotor is surrounded by

the fuid pumped, whose kine-

matic viscosity (see section4.) must be known, as it

infuences riction losses andthereore the motor power

required.

Metal cans or containment

shrouds (or example made

o .460) cause eddy current

losses, resulting in an increase

in the motor power required.

Non-metal shrouds in mag-

drive pumps do not have this

eect.

The vapour pressure o the

fuid pumped must be known,

so as to avoid bearing damage

caused by dry running when

the fuid has evaporated. It is

advisable to install monitoringequipment signalling dry run-

ning conditions, i any.

Data on specic fuid proper-

ties such as its solids content

and any tendency to solidiy

or polymerize or orm incrus-

tations and deposits, need to

be available at the time o se-

lection.

3.3.3.3Starting Characteristics

The pump torque Tp transmit-

ted by the shat coupling is

directly related to the power P

and speed o rotation n. Dur-

ing pump start-up, this torque

ollows an almost parabolical

curve as a unction o the speed

o rotation [0], as shown inFig. . The torque provided

by the asynchronous motor

must, however, be higher so as

to enable the rotor to run up to

duty speed. Together with the

voltage, this motor torque has a

direct eect on the motors cur-

rent input, and the latter in turn

on heat build-up in the motor

windings. The aim, thereore, is

to prevent unwanted heat build-up in the motor by limiting the

run-up period and/or current

inrush [] (see also Table ).

Motors or Seal-less Pumps Starting Characteristics


32/92

Table : Starting methods or asynchronous motors

Starting Type o Current Run-up Heat build- Mechani- Hydraulic Cost Recommended Commentsmethod equipment input time up in motor cal loading loading relation motor designs

(mains load) duringstart-up

D. o. l. Contactor 48 IN Approx. High Very high Very high All Mostly limited to(mecha- 0.55 s 4 kW by energynical) supply companies

Star- Contactor / o d. o. l. Approx. High Very high Very high .5 All; canned mo- Usually stipu-delta combi- values 0 s tors and sub- lated or motors

nation mersible motors >4 kW by(mecha- subject to a energy supplynical) major drop in companies

speed duringswitchover

Reduced Autotrans- 0.49 times Approx. High High High 55 All No currentlessvoltage ormer, the d. o. l. 0 s phase during

mostly values switchover70% tap- (gradually re-ping placed by sot

starters)Sot Sot starter Continuous- Approx. High Low Low 55 All Run-up and run-start (power ly variable; 00 s down continu-

electro- typically ously variablenics) IN via ramps or

each individualload appllication;no hydraulicsurges

Fre- Frequency IN 060 s Low Low Low Approx. All Too expensive toquency inverter 0 use solely or run-inverter (power up and run-down

electro- purposes; betternics) suited or open-

or closed-loopcontrol

3

In the case o d.o.l. starting

(where the ull mains voltage

is instantly applied to the mo-

tor once it is switched on), the

ull starting torque is instantly

available and the unit runs up

to its duty speed in a very short

period o time. For the motor

itsel, this is the most avour-

able starting method. But at up

to 4 8 times the rated current,

the starting current o the d.o.l.

method places a high load on

the electricity supply mains,

particularly i large motors are

involved, and may cause prob-

lematic voltage drops in electri-

cal equipment in their vicinity.

For motor operation on public

low-voltage grids (80 V), the

regulations laid down by the en-

ergy supply companies or d.o.l.

starting o motors o 5.5 kW

and above must be complied

with. I the grid is not suitable

or d.o.l starting, the motor can

be started up with reduced volt-

ages, using one o the ollowing

methods:

Star-delta starting is the most

requent, since most inexpen-

sive, way o reducing the start-

ing current. During normal

operation, the motor runs in

delta, so that the ull mains

voltage (or example 400 V) is

applied to the motor windings.

For start-up, however, the wind-

ings are star-connected, so that

the voltage at the windings is

reduced by a actor o 0.58 rela-

tive to the mains voltage. This

reduces the starting current and

torque to one third o the values

o d.o.l. starting, resulting in a

longer start-up process.

The motor runs up in star con-

nection beyond pull-out torque

up to the maximum speed o

rotation at point B in Fig. .

Then, switchover to delta is e-

ected and the motor continues

to accelerate up to rated speed.

During the switchover period o

approx. 0. s, the current sup-

ply to the motor is interrupted

Starting Methods


33/92

Fig. 23: Starting curve or current I and torque T o squirrel-cage

motors in star-delta connection

( = star connection; = delta connection; P = pump)

3

"gg

$g

$gg

$

)

)

4

4

40

"g

"

OFNSYNCHR

#URRENT)

4ORQUE4

-OTORSPEEDN

and the speed drops. On pump

sets with a low moment o in-

ertia (canned motors and sub-

mersible motors), this speed re-

duction may be so pronounced

that switchover to delta mayresult in almost the ull starting

current being applied ater all,

same as with d.o.l. starting.

An autotransormer also serves

to reduce voltage at the motor

windings and unlike star-delta

starting allows selection o

the actual voltage reduction. A

70% tapping o the transormer,or instance, will bring down

the start-up torque and current

supplied by the mains to 49%

o the values or d.o.l. starting.

The act that current supply is

never interrupted is another ad-

vantage o autotransormers.

Sot starters are used or elec-

tronic continuous variation o

the voltage at the motor wind-

ings in accordance with the

dimmer principle. This means

that the start-up time and start-

ing current can be reely selected

within the motors permissible

operating limits (heat losses due

to slip!). Special constraints re-

garding the requency o starts

(contrary to Table 0) have tobe heeded [].

Frequency inverters (usually or

open- or closed-loop control)

provide a sot starting option

without the need or any addi-

tional equipment. For this pur-

pose, the output requency and

voltage o the requency inverter

(see section .4.) are increased

continuously rom a minimum

value to the required value,

without exceeding the motors

rated current.

Starting Methods


34/92

4

3

3.4Pump Perormance andControl [4], [6], [8]

3.4.1Operating Point

The operating point o a centri-ugal pump, also called its duty

point, is given by the intersec-

tion o the pump characteristic

curve (see section ..6) with

the system characteristic curve

(see section ..). The fow

rate Q and the developed head

H are both determined by the

intersection. To change the op-

erating point either the system

curve or the pump curve must

be changed.

A system characteristic curve

or pumping water can only be

changed

by changing the fow resist-

ance (or example, by chang-

ing the setting o a throttling

device, by installing an orice

or a bypass line, by rebuildingthe piping or by its becoming

incrusted) and/or

by changing the static head

component (or example, with

a dierent water level or tank

pressure).

A pump characteristic curve can

be changed

by changing the speed o rota-

tion (see section .4.),

by starting or stopping pumps

operated in series or parallel

(see sections .4.4 or .4.5),

or pumps with radial impel-

lers, by changing the impel-

lers outside diameter (see sec-

tion .4.6),

or pumps with mixed fow

impellers, by installing or

changing the setting o in-

stalled pre-swirl control

equipment (see section .4.8),

or axial fow (propeller)

pumps, by changing the blade

pitch setting (see section

.4.9).

Please note: the eect o these

measures or changing the char-

acteristic curve can only be pre-

dicted or non-cavitating opera-

tion (see section .5).

3.4.2Flow Control by Throttling

Changing the fow rate Q by

operating a throttle valve is the

simplest fow control method

not only or a single adjustment

o the fow rate but also or

its continuous control, since it

requires the least investment.

But it is also the most energy

wasting method, since the fow

energy is converted irreversibly

to heat.

Fig. 4 illustrates this process:

by intentionally increasing the

system resistance (or exampleby throttling a valve on the

1;=

1;=

(;=

"

"

0

0

0;=

0UMPCHARAC

TERISTICCURVE

4HROTTLING

3URPLUSHEAD

3YSTEMHEAD

REQUIREMENT

3YSTEMCURVE( SYS

3YSTEMCURVE(SYS

0OWERSAVED

Fig. 24: Change o the operating point and power saved by

throttling a pump whose power curve has a positive slope

Pump Perormance Operating Point Throttling


35/92

5

3

pump discharge side) the origi-nal system curve Hsys becomes

steeper and transorms into

Hsys. For a constant pump

speed, the operating point B on

the pump characteristic moves

to B at a lower fow rate. The

pump develops a larger head

than would be necessary or

the system; this surplus head is

eliminated in the throttle valve.

The hydraulic energy is irrevers-

ibly converted into heat which

is transported away by the fow.

This loss is acceptable when

the control range is small or

when such control is only sel-

dom needed. The power saved

is shown in the lower part o

the gure; it is only moderate

compared with the large surplus

head produced.

D"L

D

!REARATIOD"LD

4HROTTLINGC

OEFFICIENTF

Fig. 25: Orice plate and its throttling coecient

The same is principally trueo the installation o a xed,

sharp-edged orifce plate in the

discharge piping, which can be

justied or low power or short

operating periods. The neces-

sary hole diameter dBl o the

orice is calculated rom the

head dierence to be throttled

H, using the ollowing equa-tion:

dBl = Q/ g H (20)

where

dBl Hole diameter o the orice

in mm

Throttling or pressure drop

coecient acc. to Fig. 5

Q Flow rate in m/h


9.8 m/s

H Head dierence to bethrottled in m

Since the area ratio (dBl/d)

must be estimated in advance,

an iterative calculation is neces-

sary (plotting the calculated vs.

the estimated diameter dBl is

recommended so that ater two

iterations the correct value can

be directly interpolated, see

example calculation 8.0).

3.4.3

Variable Speed Flow Control

At various speeds o rotation n,

a centriugal pump has dier-

ent characteristic curves, which

are related to each other by the

anity laws. I the characteris-

tics H and P as unctions o Q

are known or a speed n, then

all points on the characteristic

curve or n can be calculated

by the ollowing equations:

Q = Q . n/n (21)

H = H (n/n) (22)

P = P (n/n) (23)

Eq. () is valid only as long

as the eciency does not de-crease as the speed n is reduced.

With a change o speed, the op-

erating point is also shited (see

section .4.). Fig. 6 shows the

H/Q curves or several speeds o

rotation; each curve has an in-

tersection with the system char-

acteristic Hsys. The operating

point B moves along this system

curve to smaller fow rates when

the speed o rotation is reduced.

Orice Plate Variable Speed


36/92

6

I the system curve is a parabola

through the origin as or Hsys

in the example, the developed

head H according to Eq. () is

reduced to one ourth its value

and the required driving power

in Eq. () to one eighth itsvalue when the speed is halved.

The lower part o Fig. 6 shows

the extent o the savings Pcompared with simple

throttling.

I the system curve is a parabola

with a large static head compo-

nent as or Hsys, it is possible

that the pump characteristic at

reduced speed has no intersec-

Fig. 26: Operation o a variable speed pump or dierent system

characteristic curves Hsys1 and Hsys2(Power savings P1 andP2 at hal load each compared with simplethrottling)

1;=

1;=

(;=

(!

(!

(!

0

0

(!

"

0;=

]

N

N

N

0

0

0

(EADREQUIRED

(!STAT

Powersaved

3

tion with it and hence, that no

operating point results; the low-

er speed range is then o no use

and could be eliminated. The

potential power savings Pat a given fow rate Q are less

than or the system curve Hsysas shown in the lower part o

the diagram [4]. The improve-

ment compared with throttling

decreases as the static head

component Hsys,stat increases

(i.e., or a lower dynamic head

component Hsys,dyn).

Variation o the speed usually

means varying the electrical

driving requency, which must

be considered when choosing

the motor. The expenditure or

variable speed drives is not low,

but it is amortized quickly or

pumps which are used oten and

which are requently requiredto run at reduced fows with

small static head component

Hsys,stat [8]. This is particularly

the case or pumps in heating

systems.

3.4.4Parallel Operation o Centri-ugal Pumps

Where one pump is unable todeliver the required fow Q at

the operating point, it is possi-

ble to have two or more pumps

working in parallel in the same

piping system, each with its

own non-return valve (Fig. 7).

Parallel operation o pumps is

easier when their shuto heads

H0 are all equal, which is the

case or identical pumps. I the

shuto heads H0 dier, the low-

est shuto head marks the point

on the common H/Q curve or

the minimum fow rate Qmin,

below which no parallel opera-

tion is possible, since the non-

return valve o the pump with

smaller shuto head will be held

shut by the other pump(s).

During parallel pumping it

must be kept in mind that aterstopping one o two identical

centriugal pumps (Fig. 7),

the fow rate Qsingle o the re-

maining pump does not all

to hal o Qparallel, but rather

increases to more than hal. The

remaining pump might then

immediately run at an operat-

ing point Bsingle above its design

point, which must be considered

Variable Speed Parallel Operation


37/92

7

3

when checking the NPSH values

(see section .5) and the drive

power (see section ..). The

reason or this behaviour is the

parabolic shape o the system

characteristic Hsys. For the samereason, the reverse procedure o

taking a second identical pump

on line does not double the fow

rate Qsingle o the pump that

was already running, but rather

increases the fow rate less than

that:

Qparallel


38/92

8

3

3.4.5

Series Operation

In series operation, the pumps

are connected one ater the

other so that the developed

heads can be added or a given

fow rate. This means that the

discharge pressure o the rst

pump is the inlet pressure or

the second pump, which must

be considered or the choice o

shat seal and or the strength

o the casing. For this reason

multistage pumps are usually

used or such applications (ex-

cept or the hydraulic transporto solids, see section 6). They

do not pose these shat sealing

problems.

3.4.6

Turning Down Impellers

I the fow rate or developed

head o a radial or mixed fow

centriugal pump are to be

reduced permanently, the out-

side diameter D o the impeller

should be reduced. The reduc-

tion should be limited to the

value or which the impeller

vanes still overlap when viewed

radially. The documentation o

the pump characteristics (Fig.

8) usually shows curves or

several diameters D (in mm).

Impellers made rom hard ma-terials, such as those used or

solids-handling pumps, or rom

stainless steel sheet metal, as

well as single vane impellers

(Fig. 4) and star or periph-

eral pump impellers cannot be

turned down. (The same is true

or under-ling as described in

section .4.7). For multistage

pumps, usually only the vanes

Dt

Dr

D1

but not the shrouds o the im-

pellers are cut back. It is some-

times possible to simply remove

the impeller and diuser o one

stage o a multistage pump and

replace them with a blind stage(two concentric cylindrical cas-

ings to guide the fow) instead

o cutting back the impeller

vanes. Impellers with a non-

cylindrical exit section are either

turned down or have only their

blades cut back as specied in

the characteristic curve litera-

ture (or example, as shown in

Fig. 9).

I the impeller diameter only

needs to be reduced slightly, a

rule o thumb can be applied.

An exact calculation cannot be

made, since the geometrical sim-

ilarity o the vane angle and exit

width are not preserved when

turning down the impeller. The

ollowing approximate relation-

ship exists between Q, H and

the impeller diameter D to beound (averaged, i required):

(Dt/Dr) Qt/Qr Ht/Hr (26)

where subscript t designates the

condition beore the reduction

Fig. 29: Contour or cutting

back the vanes o an impeller

with a mixed fow exit

o the impeller outer diameter

and index r the condition ater

the reduction. The required

(average) reduced diameter re-

sults as:

DrDt (Qr/Qt)Dt (Hr/Ht) (27)

The parameters needed to deter-

mine the reduced diameter can

be ound as shown in Fig. 0:

in the H/Q curve (linear scales

required!) a line is drawn con-

necting the origin (careul: some

scales do not start at zero!) and

the new operating point Br . The

extension o the line intersects

the characteristic curve or ull

diameter Dt at the point Bt. In

this way the values o Q and H

with the subscripts t and r can

be ound, which are used with

Eq. (7) to nd the desired re-

duced diameter Dr.

The ISO 9906 method is more

accurate, but also more involvedthrough the consideration o

the average diameter D o the

impeller leading edge (sub-

script ), valid or nq < 79 and

or a change o diameter < 5%,

as long as the vane angle and

the impeller width remain con-

stant. Thus using the nomen-

clature o Figs. 9 and 0:

Series Operation Impeller Diameter Reduction


39/92

9

3

"T

1T

"R

1R

(T

$T

(R

$R

usin

gEq

.28

using

Eq.26

Totaldevelopedhe

adH

Flow rate Q

Fig. 30:

Determination

o the reduced

impeller dia-

meter Dr

A solution is only possible when

D is known and when a para-

bola H ~ Q is drawn through

the reduced operating point Br

(with Hr and Qr), not a line as

n

(DrD)/(DtD) = Hr/Ht = (Qr/Qt) (28)

in Fig. 0, which intersects the

base H/Q curve or diameter Dt

at a dierent point Bt (with di-

erent Ht and Qt).

Fig. 31: Under-led vanes o a

radial impeller

Fig. 32: Characteristic curve set o a centriugal pump with pre-swirl

control equipment, nq 160

HHOPT

/PERATINGLIMIT

0RESWIRLCONTROLSETTING

2ELATIVEFLOWRATE11OPT

2ELATIVEDE

VELOPED

HEAD

((OPT

3.4.8Pre-Swirl Control o the Flow

For tubular casing pumps with

mixed fow impellers, the pump

characteristic can be infuenced

by changing the pre-rotation

in the impeller inlet fow. Such

pre-swirl control equipment is

oten tted to control the fow

rate. The various characteristic

curves are then shown in the

product literature labelled with

the control setting (Fig. ).

3.4.9

Flow Rate Control or Changeby Blade Pitch Adjustment

The characteristic curves o ax-

ial fow (propeller) pumps can be

altered by changing the setting

o the propeller blade pitch. The

setting can be xed and rmly

bolted or a device to change the

blade pitch during operation

can be used to control the fow

rate. The blade pitch angles are

3.4.7Under-ling o ImpellerVanes

A small, permanent increase o

the developed head at the best

eciency point (up to 4 6%)

can be achieved or radial im-

pellers by ling the back sides o

the backward-curved vanes, i.e.,

by sharpening the vanes on the

concave side, as shown in Fig.

. The shuto head does not

change. This method is suitable

or minor nal corrections.

Impeller Diameter Reduction Under-ling Pre-swirl Blade Pitch Adjustment


40/92

40

3

shown in the product literature

with their respective characteris-

tic curves (see Fig. ).

3.4.10Flow Control Using a Bypass

The system characteristic curve

can be made steeper by closing

a throttle valve, but it can also

be made fatter by opening a

bypass in the discharge piping

as shown in Fig. 4. The pump

operating point moves rom B

to a larger fow rate B. The

bypass fow rate is controlled

and can be ed back into theinlet tank without being used

directly. From the point o view

o saving energy, this type o

control only makes sense when

the power curve alls or in-

creasing pump fow rates (P >

P), which is the case or high

specic speeds (mixed and axial

fow propeller pumps).

For these types o pumps, con-trolling the fow by pre-swirl

control or by changing the blade

pitch is even more economical,

however. The expenditure or a

bypass and control valve is not

small [4]. This method is also

suitable or preventing pumps

rom operating at unacceptably

low fow rates (see operating

limits in Figs. 5 and 6c as well

as in Figs. and ).

HHOPT

/PERATINGLIMIT

"LADEPITCHSETTING

2ELATIVEFLOWRATE11OPT

2ELATIVEDEVELOPED

HEA

D

((OPT

1;=

1;=

(;=

"

"

0

0

0;=

-

"YPASSFLOWRATE5SEFULFLOWRATE

3URPLUSHEAD

(EADREQUIRED

BYTHESYSTEM

3YSTEMCURVE

WITHOUTBYPASS0UMP

CHARACTERISTIC

3YSTEMCURVE

WITHBYPASS

0OWERSAVED

Fig. 33: Characteristic curve set o an axial fow pump with blade

pitch adjustment, nq 200

Fig. 34: Characteristic curves and operating points o a pump with

a alling power curve and fow control using a bypass. (For a radial

fow pump the power curve would increase towards the right and

this type o control would cause an increase in power input, see

Fig. 5).

Blade Pitch Adjustment Bypass


41/92

4

3

3.5Suction and Inlet Conditions[3]

NPSH = Net Positive Suction

Head

3.5.1

The NPSH Value o theSystem: NPSHa

The NPSHa value is the dier-

ence between the total pressure

in the centre o the pump inlet

and the vapour pressure pv,

expressed as a head dierence

in m. It is in certain respects a

measure o the probability o

vaporization at that location

and it is determined only by

the operating data o the sys-

tem and the type o fuid. The

vapour pressure o water and

other liquids are shown in Table

and in Fig. 5 as a unction

o the temperature.

Carbon

disulph

ide

!CETON

E"E

NZOL

N"UT

ANE

%TH

ANE

%TH

ANOL

$IETH

YLETH

ER

0ROPANE

I"

UTA

NE

"ENZOL

0HE

NOL

4OLU

OL

!NILIN

E

-ETH

ANOL

!CETO

NE

!CETICACID&

ORMIC

ACID

'LY

CERIN

#ARB

ON

DIS

ULPHID

E

!MM

ONI

A

3ULP

HURDIOXI

DE

#ARBO

NTETRACHL

ORID

E

"ENZOL

n #

BAR

4EMPERATURET

6APOURPRESSURE

P6

Fig. 35: Vapour pressure pv o various fuids as a unction o the

temperature t (or an enlarged view see page 84)

Suction and Inlet Conditions NPSH Available


42/92

4

Table 12: Vapour pressure pv, density and kinematic viscosity o water at saturation conditions as a

unction o the temperature t

3 NPSH Available Data or Water

t pv C bar kg/m mm/s

0 0.006 999.8 .79 0.00656 999.9 0.00705 999.9 0.00757 000.04 0.008 000.05 0.0087 000.06 0.0095 999.97 0.000 999.98 0.007 999.89 0.046 999.7

0 0.07 999.6 .07

0.0 999.5 0.040 999.4 0.0496 999.4 0.0597 999.5 0.070 999.06 0.086 998.8

7 0.096 998.78 0.006 998.59 0.096 998.40 0.07 998. .004

0.0485 997.9 0.064 997.7 0.0808 997.54 0.098 997.5 0.067 997.06 0.060 996.77 0.0564 996.48 0.0779 996.

9 0.04004 995.80 0.044 995.6 0.80

0.0449 995.

0.0475 994.9 0.0509 994.64 0.058 994.5 0.056 99.96 0.05940 99.57 0.0674 99.8 0.0664 99.99 0.0699 99.640 0.0775 99. 0.658

4 0.07777 99.84 0.0898 99.44 0.0869 99.044 0.0900 990.645 0.0958 990.46 0.0085 989.8

47 0.06 989.48 0.6 988.949 0.76 988.550 0.5 988.0 0.55

5 0.960 987.75 0.6 987.5 0.49 986.754 0.500 986.55 0.574 985.756 0.6509 985.57 0.7 984.758 0.846 984.59 0.905 98.760 0.990 98. 0.474


6 0.086 98.66 0.84 98.6 0.85 98.664 0.9 98.65 0.50 980.566 0.64 980.067 0.7 979.468 0.856 978.869 0.98 978.70 0.6 977.7 0.4

7 0.5 977.7 0.96 976.67 0.54 976.074 0.696 975.475 0.855 974.876 0.409 974.77 0.489 97.7

78 0.465 97.079 0.4547 97.580 0.476 97.8 0.65

8 0.49 97.8 0.5 970.68 0.54 969.984 0.5557 969.485 0.5780 968.786 0.600 968.87 0.649 967.488 0.6495 966.789 0.6749 966.090 0.70 965. 0.6

9 0.78 964.79 0.756 964.09 0.7849 96.94 0.846 96.695 0.845 96.996 0.8769 96.97 0.9095 960.498 0.940 959.899 0.9776 959.0

00 .0 958. 0.95

0 .0878 956.804 .668 955.506 .504 954.008 .90 95.60 .47 95.0

.56 949.6

4 .66 948.06 .7465 946.48 .868 944.80 .9854 94. 0.460

.44 94.54 .50 99.86 .9 98.8 .544 96.50 .70 94.8

.8668 9.4 .040 9.46 .4 99.68 .47 97.940 .64 96. 0.60


45 4.55 9.750 4.760 96.9

55 5.4 9.60 6.80 907.4 0.890

65 7.008 90.470 7.90 897.

75 8.95 89.80 0.07 886.9 0.697

85 .4 88.490 .55 876.0

95 .989 870.00 5.550 864.7 0.579

05 7.45 858.70 9.080 85.8

5 .06 846.60 .0 840. 0.488

5 5.504 84.00 7.979 87.

5 0.65 80.640 .480 8.6 0.40

Documents

Selecting Centrifugal Pumps Data