Mechanical Engineer''''s Reference Book 2011 Part 13 pot

Vul 11s usually considered as zero in design flow conditions, SO Radial Figure 15.1 Flow paths used in rotodynamic machines Compressors, fans and pumps 1513 Figure 15.2 A simple radi

Dynamics of floating systems 14/31 conditions Now, the solution for scattered wave potential due

to the stationary floating body, subjected to incident waves of potential, ~ $ 2 , is identical to that described in Section 14.5 for fixed structures A set of linear simultaneous equations are obtained by equating the flow due to the local source plus the additional flow due to all other sources to the negative of the flow due to the undisturbed wave for each facet on the body surface Solutions of these equations yields the unknown source strengths and, therefore, the velocity potential, bs,

which is used to derive pressures and wave forces by integra- tion over the body surface Thus the wave force vector, F, of equation (14.46) may be obtained for an incident wave of specified frequency and direction

The velocity potentials, + f > are obtained in a way similar to that above except for the use of a different boundary condition which reflects the fact that bf arises from body motions in otherwise still water Thus, at all facets, the source strengths, + f i > are such that the flow due to the local source plus the flow due to all other sources equals the velocity component of the body along the facet normal This velocity component will depend on the mode of motion (surge, sway, heave and SO on)

in which the body is moving All of this can be represented by equating the normal velocity of the fluid and of the jth facet for the vessel moving in its kth mode of motion This yields the equation

(14.53) where v,k is the normal velocity of the jth facet with the vessel

moving in its kth mode of motion Furthermore, nj is the

normal to the jth facet, a+,lanj is the normal fluid velocity at

the jth facet due to a unit source at the itb facet, and utk are the unknown source strengths required in the kth mode Application of equation (14.53) for all facets produces a system of complex equations to be solved for the source strengths Once these are known, the pressures at the facets are evaluated and their effects integrated over the vessel surface to yield forces in each mode of motion to unit motion

in the kth mode

These forces may be written as a complex square matrix, G(w) which can be decomposed into its real and imaginary parts through the equation

G(w) = w 2 MA (w) - i~ BJw) (14.54)

to yield frequency-dependent added mass and damping ma- trices M A ( w ) and Bp(w) which are required for equation

(14.46)

The inclusion of physical mass, hydrostatic and mooring

stiffness matrices, M , K and K, completes derivation of all of

the coefficient matrices of equation (14.46) The hydrodyna- mic coefficient matrices are, however, frequency dependent and require carrying out a diffraction analysis at all frequen- cies at which motions are required Equation (14.46) is linear

and can readily be solved to yield the displacement vector X

The exciting force vector F(w) and the coefficient matrices

M A ( w ) and BJw) can also be derived using finite-element

methods in a way analogous to that for the boundary-integral approach described above

There is one further point of interest regarding the relation- ship between the scattered and forced wave potentials (rnS and

rnf) for a floating vessel problem The use of equations called Haskind relations (see Newman3') enables the scattered wave potential, rnS, to be expressed in terms of the incident and forced wave potentials, I$,, and + f Thus, once 6f is calculated, need not be computed by diffraction analysis but can

?stead be derived using the Haskind relations

$ (Tik = Vjk

linearity around resonance with the heave response amplitude

per unit wave amplitude reducing from 4.88 mim at 1 m wave

amplitude to 1.26 mim ai 6 m wave amplitude The vessel

motion response away from resonance is not significantly

affected, although there is some increase in response around

16-19 s due to the corresponding increase in wave force

amplitude at these periods The large change in the unit heave

response at and around resonance is to be expected, since the

damping force in a vibratory system is dominant at resonance

14.6.6 Diffraction theory

Calculations of wave-induced motions of a large non-space

frame structure in gravity waves requires a solution of the

wave problem with no flow boundary conditions at the moving

body surface in addition to the free surface and sea-bed

boundary conditions The solution can be split into two related

problems - the scattering wave problem defines wave forces

on a floating body when fixed in space and with waves incident

on it in an identical manner to the technique for computing

wave forces on a fixed body described in Section 14.4 The

radiation wave problem is concerned with defining forces on

the body (added mass and damping) due to its oscillation in

otherwise still water These oscillations will induce wave

potentials such that the total wave potential in the fluid is the

sum of the incident, +,,,> scattered, &, and forced wave

where V,q is the velocity of the body surface in the direction n

normal to the surface This boundary condition can be applied

at the mean body surface since the theory is applied for small

motions +> together with its three components It must also

satisfy thle Laplace equation and the free surface and sea-bed

boundary conditions Furthermore, and $f must satisfy the

radiation conditions

Boundary conditions for the scattering and radiation wave

problem:j can be split up from equation (14.51) as

respectively, both being applied on the body surface The

scattering problem is identical to the application of diffraction

theory on fixed structures as described in Section 14.4 The

radiation problem can also be solved by using either

boundary-integral or boundary-element techniques For brev-

ity, only the solution using boundary-integral techniques is

describesd here As in Section 14 4, the analysis assumes

inviscid, irrotational flow and that wave amplitudes are small

The unsteady flow around the floating vessel is calculated by

introducing oscillating sources of unknown velocity potential

on the vessel's submerged surface that is discretized by a mesh

of facets with an oscillating source on the surface of each facet

A Green's function is used to represent the velocity poten-

tial of each source which, because of the form of the Green's

function satisfies Laplace's equation, zero flow at the hori-

zontal sea bed, the free surface and radiation boundary

(14.52)

~ _ _ - "in

14/32 Offshore engineering

0.6 -

Figure 14.32 Facet discretization of a submerged ship hull for diffraction theory

-0.6

Typical results of a boundary integral diffraction analysis for

a ship-shaped hull are shown in Figure 14.33 The discretiza-

tion of the submerged hull geometry is shown in Figure 14.32

using 277 triangular facets on the ship half-hull The vessel is

of 263.7 m overall length, 40.8 m beam and 145 937 t

displacement with 14.80 m draught floating in deep water

Figure 14.33(a) presents the variation of added mass and

radiation damping coefficients with frequency for heave and

pitch motions Note that the variation in added mass is

relatively small but the radiation damping shows large changes

with very small values at some wave periods Wave-induced

heave force and pitching moments and the resultant motion

responses for head seas are presented in Figures 14.33(b) and

14.33(c)

14.7 Design considerations and certification

It is important to appreciate that the design procedures for

jacket structures outlined in the previous three sections are

Wave period ( 5 )

(C)

Figure 14.33 Variations of heave and pitch added masses, wave-excitation forces and motion response with wave period for ship hull

1 Bids, evaluations, contractors, selection I

only a small part of the total design process In order to

illustrate this point, Figure 14.34 presents a flow chart showing

the design procedures that need to be followed, from the

initial specification through to commencing operation of a

typical offshore structure The jacket has to have sufficient

strength, as it is assembled during the fabrication stage and

loaded lout of the yard It has also to meet the naval architec-

tural an,d structural requirements of tow-out, up-ending and

installation as well as surviving for a 20-40-year life Some of

the supplementary design tasks not covered ifi this chapter

include the response of the structure to earthquakes, the

provision of corrosion protection and in-service structural monitoring The design procedure for iarge jackets invariably contains a model test phase for critical operations such as up-ending during installation The documentation of the material, structural and welding details of the design during its certification, fabrication and service life pose an engineering management problem

Certifying authorities play a key role in the design proced- ure for an offshore structure The major certifying authorities

in the United Kingdom, Norway and the United States have built up extensive codes of practice which reflect research

14/34 Offshore engineering

w o r k , in-service experience a n d t h e results of failure investi-

gations o v e r m a n y years of operation (see Lloyd’s Register of

S h i ~ p i n g , ~ ’ D e p a r t m e n t of Energy,j‘ D e t Norske V e r i t a ~ , ~ ~

a n d A m e r i c a n B u r e a u of Shipping36) Certifying authorities

also provide an i n d e p e n d e n t check of m a n y of t h e calculations

a n d decisions that n e e d t o be made during a typical design

T h e r e tends to be close technical collaboration b e t w e e n

research establishments, designers a n d t h e o p e r a t o r s of off-

s h o r e structures

References

1 Department of Energy, Offshore Installations, Guidance on

design and construction, Part 11, Section 4.3, HMSO, London

(1986)

2 American Petroleum Institute, Basic Petroleum Databook,

Volume VI, No 3, September API, 1220 L Street NW,

Washington, D C 20005, USA (1986)

3 Lee G C., ‘Recent advances in design and construction of

deep water platforms, Part l ’ , Ocean Industry, November,

71-80 (1980)

platforms: design and application’, Engineering Structures, 3,

July, 140-152 (1980)

5 Thornton, D., ‘A general review of future problems and their

solution‘, Proceedings of the Second International Conference

on Behaviour of Offshore Sfructures, 28-31 August, Paper 88,

BHRA Fluid Engineering, Craufield, Bedford, UK (1979)

6 Hamilton, J and Perrett, G R , ‘Deep water tension leg

platform designs’, Proceedings of the Royal Institution of Naval

Architects International Svmuosium on Develooments in Deeoer

4 Fumes, 0 and Loset, O., ‘Shell structures in offshore

Waters, 6-7 October, Paier‘no 10 (1986)

Meteorological Office Meteorology for mariners, 3rd edition,

HMSO London (1986)

Strahler, A N and Strahler, A.H., Modern Physical

Geography, Wiley, New York (1978)

Airy, Sir G B ‘Tides and waves’, Encyc Metrop., Art 192,

Morrison, J R , O’Brien, M P , Johnson, J W and Schaaf,

S A , ‘The forces exerted by surface waves on piles’,

Petroleum Transactions, 189, T P 2846, 149 (1950)

Sarpkaya, T.; ‘In line and transverse forces on smooth and

sand roughened cylinders in oscillatory flow at high Reynolds

numbers’, Report No NPS-69SL76062, Naval Postgraduate

School, Monterey, California (1976)

Sarpkaya, T and Isaacson, M., Mechanics of Wave Forces on

Offshore Structures, Van Nostrand Reinhold, New York (1981)

Sommerfield, A , , Partial Differential Equations in Physics,

Academic Press: New York (1949)

Stoker, J J., Water Waves, Interscience, New York (1957)

MacCamy, R C and Fuchs, R A , , ‘Wave forces on piles, a

diffraction theory’, US Army Corps of Engineers, Beach

Erosion Board, Tech Memo No 69 (1954)

Garrison C J and Chow, P Y , ‘Wave forces on submerged

bodies’, Journal of Waterways, Harbours and Coastal Division,

International Journal for Numerical methods in Engineering,

Zienkiewicz 0 C., Bettes, P and Kelly D W., ‘The finite element method of determining fluid loading on rigid structures - two and three dimensional formulations’: in Zienkiewicz, 0 C Lewis, P and Stass, K G (eds)

Numerical Methods in Offshore Engineering Wiley, Chichester

( 1978) Penzien, J and Tseng, W S , ‘Three dimensional dynamic analysis of fixed offshore platforms’ in Zienkiewicz, 0 C et

al (eds) Numerical Methods in Offshore Engineering, Wiley,

Chichester (1978) Bathe, K J and Wilson, E L., ‘Solution methods for eigen-value problems in engineering‘, International Journal for

Numerical Methods in Engineering, 6, 213-216

Malhotra A K and Penzien, J., ‘Nondeterministic analysis of offshore tower structures’, Journal of Engineering Mechanics

Division, American Society of Civil Engineers, 96 No EM6 985-1003 (1970)

Poulos, H G and Davis, E H., Pile Foundation Analysis and Design, Wiley, New York (1980)

Reese, L C , ‘Laterally loaded pile; program documentation‘,

Journal of the Geotechnical Engineering Division, American

Society of Civil Engineers 103, No GT4, 287-305 (1977) Focht, J A , Jr and Kock, K J., ‘Rational analysis of the lateral performance of offshore pile groups’, Proceedings of the Offshore Technology Conference OTC 1896 (1973)

O’Neill, M W., Ghazzaly, 0 I and Ho, Boo Ha, ‘Analysis of three-dimensional pile groups with nonlinear soil response and pile-soil-pile interaction’ Proceedings of the Offshore Technology Conference OTC 2838 (1977)

American Petroleum Institute, Recommended practice for planning, designing and constructing fired offshore platforms,

Dallas, Texas, Rpt No API-RP-2A (revised annually) (1987) British Standards Institution, Code of practice for fixed offshore structures, BS 6235: 1982, BSI, 2 Park Street London, W I A 2BS

Dover, W D and Connolly, M P ‘Fatigue fracture mechanics assessment of tubular welded Y and K joints’, Paper

No C141186 Institution of Mechanical Engineers London

(1986) Dover, W D and Wilson, T J., ‘Corrosion fatigue of tubular welded T-joints’, Paper No C136186; Institution of Mechanical Engineers, London (1986)

Warburton, G B., The Dynamical Behaviour of Structures,

2nd edition, Pergamon Press, Oxford (1976) Newman J N., ‘The exciting forces on fixed bodies in waves’,

Journal of Ship Research, 6, 10-17 (1962)

Lloyd’s Register of Shipping, Rules and regulations for the classification of mobile offshore units, January, Part IV,

Chapter 1, Sections 2, 3, 4 and 5 , Lloyd’s Register of Shipping,

71 Fenchurch Street, London EC3 4BS (1986) Department of Energy, Development of the oil and gas resources of the United Kingdom Appendix 15, Department of

36 American Bureau of Shipping, Rules for building and classing

mobile offshore drilling units, ABS, 45 Eisenhower Drive, PO

Box 910, Paramus, New Jersey, USA (1987)

15.1.2 Machine selection 15/13 15.3.5 Waste-heat boilers 15/84

15.1.3 Performance monitoring and prediction 15/14 15.3.6 Economizers 15/84

ct requirement for chimneys and

15.3.1 Types of boilers 15/75

15.4 Heating, ventilation and air conditioning 15191 15.9.3 Sound power 151139

15.4.1 Heating 15/91 15.9.4 Addition and subtraction of decibels 15/139 15.4.2 Ventilation 15/97 15.9.5 Addition of decibels: graph method 151139 15.4.3 Air conditioning 151106 15.9.6 The relationship between SPL, SIL and 15.5 Refrigeration 151114 15.9.7 Frequency weighting and the human response SWL 151139

15.5.2 Pressure-enthalpy chart 151115 15.9.8 Noise indices 151140

15.5.3 Gas refrigeration cycle 151115 15.9.9 Noise-rating curves 15/141

15.9.10 Community noise units 15/141

15.6.1 The energy manager 15/116 15.9.12 Air traffic 151142

15.6.2 Energy surveys and audits 151116 15.9.13 Railway noise 151142

15.6.3 Applications 1511 18 15.9.14 Noise from demolition and construction

15.6.5 Control systems 151123 15.9.15 Noise from industrial premises 151142

15.7.1 Preventive maintenance 151124 15.9.19 Digital signal analysis 151143

15.7.2 Predictive preventive maintenance 151124 15.9.20 Noise control 15/143

15.7.3 Condition monitoring 151125 15.9.21 Noise nuisance 151143

15.7.5 Vibration monitoring for machine 15.9.23 Damage to plant/machinery/building

15.7.6 Vibration analysis techniques 151126 15.9.24 Legislation concerning the control of

15.9.17 Microphones 15/142

noise 151144 15.8 Vibration isolation and limits 151129 15.9.25 British Standard 4142: 1990 151145

15.8.3 Multi-degree of freedom systems 151130 15.9.28 The Health and Safety at Work etc Act

15.8.5 Shock isolation 151131 15.9.29 The Noise at Work Regulations 1989 151146 15.8.6 Vibration attenuation 151132 15.9.30 Noise control engineering 151147

15.8.7 Measurement of vibration 151133 15.9.31 Noise-reduction principles 151147

15.9.33 Vibration isolation 151148

15.9.1 Introduction - basic acoustics 151138

15.9.2 Sound intensity 151139 References 151150

15.1 Compressors, fans and pumps

15.1.1 Design principles

15.1 I .1 General

Compressors, fans and pumps are all devices for increasing the

pressure energy of the fluid involved Two basic types are

used: rotodynamic, where flow is continuous, and positive

displacement where fluid is worked on in discrete packages

defined by machine geometry Compressors, fans and pumps

may be rotodynamic, and compressors and pumps positive

displacement In general, the positive displacement machines

give low mass flow and high pressure rise

15.1.1.2 Rotodynamic machine principles

These can be discussed together as the Euler equation applies

to all types, differences being due to the fluid involved and the

flow path Figure 15.1 illustrates flow path differences

15.1.1.3 Forms of the Euler equation

Standard turbomachinery textbooks (see Turton') derive this

equation, so it will be applied here to centrifugal and axial

machines Considering Figure 15.2 (a simple centrifugal

pump) the specific energy increase is given by the Euler

equation

gH = 112vu2 - U l V U , (15.1)

where u,, u2 are peripheral velocities (=wr) V uz, V u , are the

peripheral components of the absolute velocities V2 and V,,

respectively (see Figure 15.3)

Vul 11s usually considered as zero in design flow conditions,

SO

Radial

Figure 15.1 Flow paths used in rotodynamic machines

Compressors, fans and pumps 1513

Figure 15.2 A simple radial outflow machine

Inlet velocity

v

0

(b)

with K2 depending on pz Figure 15.3 shows how varying p 2

affects both velocity diagrams and the gH to Q plot of

performance plots, compressors being affected at lower flows

by surge as discussed later

A simple axial machine is shown in Figure 15.4, with typical

general velocity diagrams, which define the geometry and

15.1.1.5 Reaction

This is defined for a compressor as:

Energy change due to or resulting from static pressure change in the rotor Total change in the stage

Compressors, fans and pumps 15/5

If a simple pump is considered, it is possible to state that

there must be a working relation between the power input P ,

the flow rate 0 , energy rise g H , fluid properties p and p , and

size of the machine D If a dimensional analysis is performed it

can be shown that a working relation may exist between a

group of non-dimensional quantities in the following equation:

Term (1) is a power coefficient which does not carry any

conventional symbol Term (2) can easily be shown to have

the shape V/Uand is called a flow coefficient, the usual symbol

being 8 Term (3) similarly can be shown to be gH/U2 and is

usually k.nown as a head coefficieat (or specific coefficient) 4

Term (4) is effectively a Reynolds number with the velocity

the peripheral speed w D and the characteristic dimension

being usually the maximum impeller diameter Term (5) is

effectively a Mach number, since K is the fluid modulus

Since these groups in the SI system are non-dimensional

they can be used to present the results of tests of pumps in a

family of pumps that are geometrically similar and dyna-

mically similar This may be done as shown in Figures 15.6 and

15.7 and Figure 15.8 shows how the effect of changing speed

or diameter of a pump impeller may be predicted using the

In Figure 15.8 points A define the energy rise gHand power

PI at a flow rate 01, when the pump is driven at speed w, If equations (15.17) are applied, D and p being the same

Q J w l D 3 = Q2/w2D3; hence Q2 gHJw{D2 = gH2/w$D2; hence gH2 PJpw:D5 = Pdpw2Ds; hence P2

This approximate approach needs to be modified in practice to give accurate results, for using model tests to predict full size power, as discussed by codes such as the American Hydraulic Institute standard^.'^

The classical approach to the problem of characterizing the performance of a pump without including its dimensions was discussed by A d d i ~ o n , ~ who proposed that a pump of standardized size will deliver energy at the rate of one horsepower when generating a head of one foot when it is driven at a speed called the Specific Speed:

minute as well a metres or feet Plots of efficiency against specific speed are in all textbooks based upon the classic Worthington plot, and Figure 15.9, based on this information, has been prepared using a non-dimensional statement known

as the characteristic number

(15.20) This is based on the flow and specific energy produced by the pump at its best efficiency point of performance following the approach stated by Wisli~enus:~ ‘Any fixed value of the specific speed describes a combination of operating conditions that permits similar flow conditions in geometrically similar hydrodynamic machines.’

Figure 15.10 presents, on the basis of the Characteristic number, the typical impeller profiles, velocity triangle shapes and characteristic curves to be expected from the machine flow paths shown In the figure the characteristic ordinates are

tested (The reader is referred to Karassik et aL5)

For compressors equation (15.16) could be employed but convention generally uses:

(15.22)

Radial M i x e d flow Axial

Figure 15.9 The variation of overall efficiency with

non-dimensional characteristic number k, for pumps (Turton’)

are the ratios of actual head/design head and actual

flow/design flow This indicates the use of the number as a

design tool for the pump engineer

The scaling laws (equation (15.17)) may be used to predict

the performance from change of speed as indicated in Figure

15.8 In many cases the pump engineer may wish to modify the

performance of the pump by a small amount and Figure 15.11

illustrates how small changes in impeller diameter can affect

the performance The diagram in its original form appeared in

the handbook by Karrasik et aL5 and has been modified to

The temperature and pressure statements are conventionally stagnation values Most compressor manufacturers use a dimensional form, and state the gas involved, so that equation (15.22) becomes:

(15.23) Figure 15.12 presents a typical compressor plot

15.1.1.6 Positive displacement machine principles

Whether the machine is of reciprocating or rotary design, fluid

is transferred from inlet to outlet in discrete quantities defined

by the geometry of the machine For example, in a single- acting piston design (Figure 15.13) the swept volume created

by piston movement is the quantity delivered by the pump for each piston stroke, and the total flow is related to the number

Compressors, fans and pumps 15/7

(adapted from Karrasik e t a/.?

Pump scaling laws applied to diameter change

- P a 2

Po 1

Lines of constant efficiency

Suction

Figure 15.13 A plunger pump (or piston pump)

of strokes per unit time Similarly, the spur-gear device (Figure 15.14) traps a fixed quantity in the space between adjacent teeth and the casing, and total flow rate is related to the rotational speed of the gear wheels

PI, and PL are defined in Figure 15.15 Table 15.1 gives typical

values of 7" and T J ~ for a number of pump types

Since discrete quantities are trapped and transferred, the

delivery pressure and flow vary as shown in Figure 15.16:

which also illustrates how increasing the number of cylinders

in a reciprocating pump reduces fluctuations In the case of

lobe and gear pumps the fluctuations are minimized by speed

of rotation and increasing tooth number, but where, for

control or process reasons, the ripple in pressure is still

excessive a means of damping pulsations must be fitted Often

a damper to cope with this and pressure pulses due to valve

closure is fitted, two types being shown in Figure 15.17 The

capacity of the accumulator is important, and one formula

based on experience for sudden valve closure is

QP2(0.016 L - T )

Here QA is the accumulator volume (m3); Q is flow rate

(m3/s); L is pipe length (m); Tis valve closure time (seconds);

Table 15.1 Some values of 17" and 7o for positive displacement

15.1.1.7 Limitations on performance

For pumps, performance is limited by cavitation, viscosity effects, gas entrainment and recirculation Cavitation occurs

in the suction zone of a pump due to the local pressure falling

to around vapour pressure as Figure 15.18 illustrates

Figure 15.18 Pressure changes on a stream surface in the

suction zone of a rotodynamic pump

Tme Three cranks 120" out of phase

Compressors, fans and pumps 1519

can be used for the duty flow required Equation (15.27) is used for reciprocating and rotary positive displacement machines, but allowance is made for acceleration effects

In reciprocators hf is calculated at peak instantaneous flow including maximum loss through a dirty filter, and an addi- tional head ‘loss’ to allow for pulsation acceleration is used:

The pump flow range is reduced as suction pressure

reduces Cavitation also causes considerable damage as

bubbles of gas form and then collapse Two criteria are used to

judge whether a pump is in trouble from cavitation or not: one

is the concept of NPSH (net positive suction head) and the

other is the noise generated

Net positive suction head is the margin of head at a point

above the vapour pressure head Two statements are used:

NPSH available and NPSH required:

NPSHA := Total head at suction flange - vapour pressure

Figure 15.19 illustrates how system NPSH or NPSHavaiiable is

calculated for the usual suction systems shown

For a centrifugal pump, the basic NPSH is calculated from

head

(15.27) where

h, = static suction head at the pump suction (rn)

hf = flow losses in suction system (m)

B = minimum barometric pressure (mbar)

(use 0.94 of mean barometer reading)

P, = minimum pressure on free surface (bar gauge)

P, = vapour pressure at maximum working temperature

(bar absolute)

In the process industries hf is calculated for the maximum

flow rate and the NPSH at normal flow allowed for by using

the formula

This gives a ‘target’ value to the pump supplier that is ‘worst’

condition In general, for cold-water duties equation (15.28)

head falls by x% (3% is often used)

For the centrifugal pump two terms are in common use: the

Thoma cavitation number u and the suction specific speed SN:

(15.33) NPSHR is defined as in Figure 15.22 This figure gives a typical

plot of u against k , that may be used as a first ‘design’ estimate

of NPSHR, but in many applications test data are required:

at design rotational speed

where K is a constant = 175 if g = 9.81 m s-*, Q is in l/s, Nin

revolutions/second, and NPSHR is m of liquid A ‘good’ value

of SN for a centrifugal pump is around 10 000

For reciprocating metering pumps NPSHR is related to

valve loading as shown in Figure 15.23:

(15.35) where dv = nominal valve size (mm) for single valves, and

PQ*

A = - 8ovQp + 15 x 105-

for double valves It is recommended that for hydraulically

operated diaphragm pumps the extra losses imposed by the

diaphragm and support plate are treated as a single unloaded

Compressors, fans and pumps 1511 1 Figure 15.27 Figure 15.25 indicates that in a positive displace- ment pump the volumetric efficiency improves and power requirement increases (with increasing viscosity)

Table 1.5.2 summarizes the effects of liquid changes (effect- ively, viscosity and density changes) on pump performance and Figure 15.26 presents material by Sterling6 which illus- trates how efficiency falls away with viscosity for two pumps working at the same duty point, graphically illustrating the rapid decay of efficiency as p increases in a centrifugai pump Figure 15.27 demonstrates a well-known method of correct- ing for fluid change from water for a centrifugal pump This allows an engineer to predict change in performance if the kinematic viscosity of the liquid to be pumped is known and the water test data are available

Recirculation effects at low flow rates are now well docu- mented, and can cause vibration and, in some cases, severe

QP

b

AP

Figure 15.25

displacemlent pump performance

Effect of viscosity increase on positive

Table 15.2 The effect of viscosity - a comparison

Type of pump Significant Effect of viscosity level Treatment and/or notes

Above 100

u p to 100

Above 100 Above 1000

- Above 100 None

Internal gear None

Up to 500 Above 500

Lowering of H-Q curve

increase in input hp Marked loss of head

Marked loss of performance

Little

Performance maintained but power input increased Flow through valves may become critical factor

-

Sliding action impaired:

slip increased Power input and heat generated increases with increasing viscosity

Power input and heat generated increases with increasing viscosity None

Cavitation may occur

-

Little or none Increasing power input required

Performance maintained similar to water performance General lowering of efficiency but may be acceptable Considerable reduction in eificiency, but high

Performance generally maintained Some reduction in speed may be advisable to reduce power input required

Speed is generally reduced to avoid excessive power inputs and fluid heating

Larger pump size selection run at reduced speed - e.g 3 X size at 1000 centistokes running at one-third speed Modification of valve design may

be desirable for higher viscosities For very high-pressure deliveries only Not generally suitable for use with other than light May be suitable for handling viscosities up to 25 000 viscosity fluids

centistokes without modification For high viscosities:

(a) Clearances may be increased (b) Speed reduced

(c) Number of gear teeth reduced For higher viscosities:

(a) Speed may be reduced (b) Number of gear teeth reduced (c) Lobe-shaped gears employed (a) Speed may have to be reduced

(b) Modified rotor form may be preferred

Nitrile rubber stator used with oil fluids Speed may be reduced to improve efficiency

-

Figure 15.26 Comparison of efficiency reduction with viscosity

increase for a screw pump and a centrifugal pump of similar duty

cavitation damage Papers given at a recent conference’ indicate the magnitude of the problem

Gas content is another important effect It is well known that centrifugal pumps will not pump high gas content mix- tures, as flow breaks down (the pump loses ‘prime’) when the gas/liquid ratio rises beyond 15% Figure 15.28 clearly shows how a centrifugal pump is affected particularly at low flow rates, and the behaviour is typical of conventional centrifugal pumps Figures 15.29 and 15.30 present well-known informa- tion on the effects of dissolved and entrained gas on the volumetric efficiency of a positive displacement pump Fans are often used in near-ambient conditions, and density change is not significant, so that inlet density is used in power calculations Care is needed in air-conditioning systems to correct for the temperature at the fan inlet Axial fan perfor- mance is affected by blade stall as in compressors

A compressor characteristic is shown in Figure 15.31 Flow

is limited at the high mass flow end of the curve at any speed when local velocity in a passage (usually the last stage outlet guide vanes in an axial machine and the diffuser vane ring in a radial compressor) reaches sonic velocity and thus mass flow cannot increase further The phenomenon of surge is more complicated as it is caused by flow instability Its effects can be limited by reducing the pressure rise in an axial stage but not eliminated Rotating stall occurs in both radial and axial

machines and its action is shown in Figure 15.32 A vane stalls and affects flow round an adjacent vane which in turn stalls This effect thus propagates round the blade row, in the opposite direction to rotation, at about half the rotational speed Reference 2 gives more detailed discussion

Figure 15.31 shows the total limitations on the compressor surge line and mass flow rate of stall and choking For detailed discussion, textbooks such as those by Horlock’ and Balje” may be consulted

Figure 15.27 A method of correction for viscosity (adapted from

American Hydraulic standard^'^) Example: The pump is t o

handle 750 USGPM of 1000 SSU Liquid against a head of 30 m

From the diagram, C, = 0.64; Ca = 0.95; C, = 0.92 at duty point

(1.0 x QN) To test on water needs tests at a flow rate of 789.5

USGPM and 32.6 m : if t h e test efficiency ’1 is 75%, oil

x

/ increasing air content

Figure 15.28 Effect of gas content on centrifugal pump performance

1 l O r Gas solubility, % by volume

Compressors, fans and pumps 15/13

Stall cell movement

Figure 15.30 ( a ) Effect of entrained gas on liquid displacement

for a positive displacement pump; (b) solubility of air in oil

by voluini? p u m p capacity is reduced to 84% of theoretical

Although Baljex and Csanady9 have proposed a common basis

of performance presentations using a non-dimensionalized

number resembling specific speed, each type of machine will

be discussed separately Engineers employed in water supply,

the process industries and other spheres of activity have a

formidable task when selecting equipment If the equipment

movement

Figure 15.32 Rotating stall in a n axial blade row

they select does not come up to specification the maximum claim on the supplier is the price paid The cost to their company is that of plant downtime and lost production which

is likely to exceed equipment costs by many times 'Buyer beware' is thus a normal rule To assist the buyer there are BS and I S 0 specifications and codes of practice such as the American Petroleum Industry (API) standards, but in many

areas there are no such aids, and the buyer has to rely on advice, experience and, ultimately, engineering common sense

Any pump, fan or compressor selected must fulfil the specified duty (or duties) and be capable of operating safely and economically with a minimum of maintenance and down- time The selector has therefore a challenging task The first essential task is to prepare the technical brief which will become the tender document This brief must state the entire operating envelope of the machine, with complete details of temperature, humidity, fluid properties and site variations, and detail the standards and codes which will apply, e.g API 610" for refinery and petrochemical centrifugal pumps This covers materials, bearing and seal systems, pressure testing of casings vibration and noise limits, hydraulic performance, draft documents, shipping and installation in over 100 pages

In short it is a comprehensive document of mutual understand- ing between customer and supplier

The project engineer needs data to decide which type of machine, likely size, rotational speed and drive system before submitting a detailed tender document Some basic charts will therefore be discussed

The principles of the two groups of pumps (rotodynamic and positive displacement) have been discussed, and Figures 15.33 and 15.34 illustrate the main types A universal index of flow path and size for centrifugal pumps is the specific speed referred to above (Figure 15.10) which indicates the flow path shapes and probable characteristics A useful pressure to flow rate envelope is shown in Figure 15.35 Once a type is decided, manufacturers' data may be consulted Usually these are test

data when pumping water for rotodynamic machines, and an

approximate idea of performance can be obtained by convert- ing water data using a conversion chart as described earlier When considering the selection of positive displacement pumps, Figure 15.36 is a useful range guide Fan selection devolves into the choice of an axial or a centrifugal machine, and whether a single- or double-stage machine is required, but choice is usually determined by flow rate and pressure rise needed, and in some cases by the space available in which a machine will need to be installed Table 15.3 gives a working basis for fan selection

Figure 15.37 outlines the main types of compressor, and it must be said that when selecting turbo compressors the choice

of machines is a function of delivery pressure and flow rate (as Figure 15.38 indicates) To extend consideration further,

Figure 15.39, based on an article in a Sulzer Technical Review,

15/14 Plant engineering

Figure 15.33 Some typical pump layouts (a) A monobloc design with the impeller fixed on the motor shaft; (b) a modern back pull-out design; (c) a double-entry pump; (d) a multistage pump design

covers plant supplied by the company of both turbomachine

and positive displacement, screw, vane, or diaphragm types

Most makers offer oil-lubricated and non-lubricated

machines

As Figures 15.38 and 15.39 indicate, the selection of com-

pressor type depends on the pressure rise and flow rate

required Large compressors are supplied for a number of

duties About 20% are used for air compression, for factory

services where usage is typically around 87 bar for energy

storage, for other industrial duties, or in bottles Industrial

usage in the field of oxygen, nitrogen and medical gases

accounts for a large sector Natural gas transmission is also a

substantial field of application The industrial and process

processes work on pressure up to about 400 bar Polyethylene

processes demand pressures up to 3500 bar As the discussion

on pumps indicated, when the duty could be met by a number

of types, choice is often determined by experience in service,

complexity or cost The only positive attitude is probably to

choose a turbomachine unless company policy dictates a

positive displacement one One factor with compressors is the

temperature rise (over 150°C with a reciprocator) Cooling

causes water and water vapour to accumulate with the conse-

quent need for careful after cooling, intercooling between

stages in multi-stage machines and water collection to prevent

tools, instruments or equipment being damaged

Oil injection is often used in screw and rotary vane machines

to cool and to help eliminate water Wear is also reduced Oil

injection at the rate of up to 20 mg m-3 of gas is used, and then filtered well below the contamination limit for factory air (5 mg ~ n - ~ ) (An efficiency of recovery of 99.9999% has been claimed.)

Selection methods should reflect operating experience as well as being based on intelligent use of manufacturers’ data, satisfactory performance results from rigorous adherence to company specifications as well as good selection

15.1.3 Performance monitoring and prediction Any pump, fan or compressor is supplied against a contract duty In the case of many small pumps and fans which are quantity or batch produced makers will often quote against a typical performance which they check by routine testing, and will only do a full works test if a customer requires this With larger pumps, fans and compressors, a full works test (usually witnessed) is required, and often check tests when installed in the systems will be needed; this latter point will be discussed together with routine monitoring

15.1.3 1 Works tests

For back-pull out pumps I S 0 519911 covers all aspects, includ- ing testing, seals, bearings, noise and vibration, and lists all

the relevant I S 0 and related BS 5316 standards, among which

Part 1 (for general-duty class C pumps) and Part 2 (for class B

fans and

CISCHARGE DISCHARGE

Figure 15.34 Some positive displacement pump designs (a) Single plunger pump; (b) simple diaphragm pump; (c) mono pump; (d) twin-screw pump; (e) steam reciprocating pump; (f) gear pump; ( 9 ) lobe pump; (h) vane pump

15/16 Plant engineering

Figure 15.35 Range chart for rotodynamic pumps (after data

published by Nederlandse Aardolie MIJ BV) Figure 15.36 Range chart for positive displacement pumps (after Nederlandse Aardolie MIJ BV)

Table 15.3 An aid t o fan selection

Type Pressure Industry Normal Applica-

H and V

Dust and fume

Vee and direct

Vee and direct

Will only handle clean air Compact and quiet running Used on heating, ventilation and air-conditioning work

General dust and fume High-pressure systems and on dust-collector plants Will handle some dusty air

Furnace blowing, cooling, conveying and where there is

a need for high pressures Blowers

Compressors, fans and pumps 15/17

Actual suction volume ( m 3 / h )

Figure 15.39 An example of a manufacturer’s range chart (based on a Solzer Review article17) AI - reciprocating compressor, lubricated

and non-lubricated cylinders; A, - reciprocating compressors, lubricating compressors; B - screw compressors, dry or oil-flooded rotors;

C - liquid ring compressor; D - rotary (Roots type); E - centrifugal compressors; F - axial compressors

pumps) detail test arrangements and procedures as well as

instrumentation for pressure, flow, torque power and speed

Permissible bands of readings are specified as are alternative

cavitation tests In the case of BS 5316 it is stated in an annex

that for mass-produced pumps the manufacturers, if they state

that the standard is being satisfied, must be able to ensure that

performance for any pump does not diverge from the pub-

lished curve by more than +6% for total head, +8% for flow

rate and t8% for input power This allows customers to have

confidence in the published curves Similar provisions will be

found in the American Hydraulic Institute Standards.13 If the

pump is to follow API 610 these standards must be satisfied

Where the liquid to be pumped is not water it is common

practice to test on cold water and to predict the performance

to he expected by using a chart such as Figure 15.36, which

gives an example of how water test duty may be obtained if the

duty is known

For fans, standards also specify instrumentation and test rig

layout BS 848: Part li4 gives methods of standardized testing

and also of prediction when models are used and of allowance

for compressibility Since fan noise is important in ventilation

systems BS 848: Part 214 lays down noise-testing techniques

and gives details of test chambers and site provisions The two

parts form an essential item of fan test provision, and give all

the necessary equations required for test data presentation as

well as for prediction of probable performance from model

tests, and for correction for non-standard situations and air

conditions

A similar standard, BS 2009,” covers acceptance tests for

turbo-type compressors and exhausters This also states provi-

sions for standardized rig layout and instrumentation and

methods of presenting data in a standardized way Corrections

for compressibility and methods of performance prediction are all given

BS 1571: Part 116 lays down provisions for testing positive

displacement compressors of all the common types in use, both in packaged form and other installations

All the standards give lists of British Standards which are relevant and quote I S 0 Standards which correspond The reader is referred to the literature listed if test procedures and equipment are being planned and where standardized me- thods of performance are being sought for contract purposes

Performance prediction is covered in the standards and fol- lows broadly the dimensionless quantities described here

15.2 Seals and sealing

15.2.1 Compression packing

15.2.1.1 Introduction

Compared to the deterministic qualities of ferrous metals, for example, the essentially deformable nature of sealing ma- terials has introduced a measure of variability that causes many commentators to regard fluid sealing technology as an art rather than a science Seen as an anachronism in a period

of high technological achievement, compression packings show no signs of losing significant ground in terms of produc- tion quantities as new and improved types proliferate in both Europe and elsewhere To understand this situation requires some appreciation of the fundamental mode of operation of the adjustable gland or stuffing box shown in Figure 15.40

Seals and sealing 15/19

0 Frequent ability to cater for adverse conditions without elaborate precautions

Valves If any doubt exists regarding selection on pumps then

a much more obvious choice of soft packing applies to the valve scene The relative lack of movement, ease of fitting and, in this case, lack of leakage requirement for lubrication purposes (plus the most decisive advantage of low cost) are factors which ideally relate to compression packings There are areas where moulded elastomeric seals present a reasonable alternative but even the most exotic compounds would seldom be used above 250°C - unless reinforced by asbestos fabric

Compressive force System

pressure

c

Y

Figure 15.40 Compression packing

This may be filled with split packing rings chosen from a

variety of materials and constructions, described elsewhere,

which art: persuaded to react against a shaft, whether rotary or

reciprocating, to the extent that the radial force developed

exceeds the pressure to be sealed Packings in this category

used for rotating or reciprocating equipment rely on a con-

trolled leakage for long-term lubrication purposes if they are

to survive for an adequate period The continued justification

for the icompression packing might appear obscure against

such a background but there can be no doubt that certain areas

of application exist where no reasonable substitute is avail-

able

Pumps Many reasoned and well-researched papers have

been published to support mechanical seals against soft pack-

ing, and vice versa There is no doubt that the former have

supplanted packed glands as original equipment on the major-

ity of rotodynamic pumps for a variety of process and service

fluids, but there are operating parameters and cost considera-

tions which will frequently dictate the choice of soft packing

Table 158.4 compares the relative attributes of the two con-

tenders in basic terms

In general, it may be said that, unless zero leakage is an

absolute priority, compression packings will retain an impor-

tant position wherever regular maintenance is available and

the following considerations apply:

@ Simplicity in gland design and ancillary equipment

@ Ease o f fitting

0 Flexibility of supply and spares for plant utilizing many

different types and sizes of pump handling a wide variety of

To increase density and dissipate heat, soft packings inva- riably contain lubricants, loss of which, through excessive compression or overheating in service, will result in packing volume loss with subsequent reduction in the effective sealing reaction and correspondingly increasing leakage rates By limiting compression to a point where slight controlled leakage

is obtained, adequate lubrication of the dynamic surfaces is ensured and overcompression of the packing avoided However, where lubrication is a problem - or a degree of gland cooling is required - a lantern ring can be incorporated into the gland area for the distribution of additional lubri-

Of the order of 10:l in favour of soft packing depending on size and application Ample warning of impending failure with

possibilities for correction Essentially simple - requiring no special skills if

correct procedure adopted

Facility for stocking length form material or complete pre-formed sets at relatively low cost Can be considerable; shaft sleeves reduce replacement costs

Friction losses slightly higher with soft packing Leakage losses zero with mechanical seals but positive with soft packing as lubrication of sealing rings is essential

APPROXIMATELY EQUAL

Little or no warning of end of useful life with possibility of sudden complete failure Skilled fitting required - precisely defined environment and assembly

Spare seal components must be available - cost can be substantial

Nil

cant/coolant (Figure 15.41(a)) The position of a lantern ring

will depend on the nature of the application but, since the

packing rings nearest to the gland spigot do most of the work,

the additional fluid should usually be introduced near to that

area

If it is essential that the fluid being pumped does not escape

to atmosphere (e.g a toxic medium), the lantern ring may

serve to introduce a barrier fluid at a pressure of 0.5-1 bar

above that to be sealed (Figure 15.41(b)) Similarly, where

there is a risk of severe abrasive wear to the packing, a

flushing fluid may be introduced through the lantern ring

(Figure 15.41(c)) For application with negative pump press-

ures (i.e suction) a supply of the medium being sealed can be

made through the lantern ring to prevent air-drawing (Figure

15.41(d))

If extreme temperatures are to be encountered it is unlikely

that cooling through the lantern ring will be sufficient and

recourse must be made to internal cooling of the gland housing

and shaft to reduce the temperature at the gland to a value

within the packing’s capabilities Conversely, when dealing with media which crystallize or congeal when cool (e.g sugars, tars, etc.), the packing will face rapid destruction unless gland heaters or a steam-jacketed arrangement are employed to restore the fluid state before starting up

It should always be remembered that the inclusion of a lantern ring into the gland area invariably complicates as- sembly and can provide a possible source of shaft scoring They should, therefore, only be considered when the nature of the application absolutely demands their presence

15.2.1.3 Gland design

At this juncture, few international standards exist to define housing design for soft packings but the dimensions shown in Table 15.5 should be satisfactory for most applications Hous- ing depths will vary with individual circumstances, such as the inclusion of a lantern ring, but five rings of square-section

Table 15.5 Suggested housing widths in relation to shaft diameters (all dimensions in millimetres)

Seals and sealing 15/21 modern packing materials, bevelled glands are seldom an advantage and can actually promote movement of the sealing ring on the spigot side into the live clearance

5 The need to avoid excessive shaft misalignment or whip

6 The provision of adequate shaft support The packing must not be used as a bearing

at m i n i m u m

to prevent extrusion under hydraulic or compressive load

Figure 15.42

packings are usually recommended for the average, uncompli-

cated duty

Other design considerations worthy of note but often

overlook.ed, are summarized as follows (see Figure 15.42):

1 The provision of an adequate tapered 'lead in' at the

mouth of the giand to facilitate entry of the packing and to

obviate the risk of damage in the assembly operation A

minimum of 15" X 6.5 mm usually represents good prac-

tice

2 The provision of a reasonable surface finish on adjacent

metal parts - particularly the dynamic surface The better

the finish, the less wear will occur; 0.4 pm (16 p in) R, on

the shaft and 1.6 pm (64 p in) R, on the stuffing box bore

should be ideal for most applications The use of shaft

sleeves can give considerable maintenance advantage

when considering the question of surface finish

3 The danger of extreme running clearances at the gland -

particularly on the spigot side In those exceptional cases

where excessive clearance is unavoidable the packing

should be protected by an independent ring of suitably

robust material or construction which reduces the clear-

ance to a miaimum

4 An allowance for entry of the gland spigot well into the

glancl area; certainly to an extent that exceeds substantially

the depth of the tapered lead in The length of spigot

selected must also cater for packing compression, resulting

from gland adjustment Typical entry lengths should be at

least two times packing section For packings of softer

construction, maximum length should be provided With

Wet

graphite mica talc molybdenum disulphide tallow

castor oil straight mineral lubricating oil petrolatum

solid fractions paraffin wax soaps silicone grease PTFE dispersions Metals: lead foil and wire

aluminium foil copper foil and wire brass wire

monel wire inconel wire stainless steel wire Elastomers: natural and synthetic The principal forms of constructions for fibrous compres- sion packings are (see Figure 15.43):

e Braided Individual yarns are braided tube over tube and

squared off The density of this type of construction is high and ideal for many valve applications

e Plaited Multiple yarns are interwoven in plaited bundles

in such a way that the direction of fibre follows the periphery of the packing ring The natural characteristic of this construction is more suited to centrifugal pump applica- tions than valve service although the inherent flexibility of the form is popular with some users

e Cross-plait All the yarns are interlocking and pass dia-

gonally through the packing to provide a firm construction

of consistent density and shape Used extensively for syn- thetic yarn packings for valves and pumps

e Composite asbestos plastic This packing category is of

fairly recent origin and includes those types based upon a braided and reinforced asbestos jacket enclosing a 'plastic' core Although inaccurate in lthe scientific sense, the term 'plastic' conveniently describes those many mixtures of

15/22 Plant engineering

Cross-plait

Figure 15.43 Basic packing constructions

Composite

asbestos fibre and lubricant, both mineral and solid, from

which readily deformable packing materials may be made,

This packing is widely accepted for difficult valve-sealing

duties

All the fibre-based constructions described here are fre-

quently reinforced with metal This applies particularly to

asbestos based products where the use of metal wire in the

yarn can extend the service capability of the packing to 800°C

and beyond But for this feature, even the best quality

non-metallic asbestos yarn packing would be restricted to

temperatures of about 315°C maximum

All the lubricants described above are used in conjunction

with fibre packings of different sorts and are applied by

dipping, coating, soaking, vacuum impregnation, dusting, etc

The prime object is maximum lubricant retention Frequently,

several treatments and repeat processes are employed to

achieve this end

In the field of metallic packings there are many construc-

tions available the three most popular being foil-wrapped

resilient asbestos core, foil crinkled and folded upon itself, and

corrugated foil, concertina wound (see Figure 15.44) These

useful types are suitable for both reciprocating and rotary

shafts and are widely used on pumps, valves, turbines, com-

pressors and refrigeration plant All are normally lubricated

with mineral oil and graphite Lead and aluminium alloys are

the most popular foil materials

In the field of compression packings, elastomers are not

widely used although some braided packings do employ yarns

that are treated with a rubber proofing to render them more

suited to difficult wet applications such as condensate duty

Rings of square or rectangular section compression packing,

manufactured from folded, rolled or laminated elastomer

proofed cloth, are still popular for relatively slow-moving,

lower-pressure reciprocating pumps handling water or LP

steam One particular design, with a moulded, double-

bevelled section, made from semimetallic rubberized yarn, is

particularly effective on rotary applications dealing with vis-

cous media which solidify when the pump is idle and cause

damage to conventional plaited packings on restarting from

cold This moulded packing is also suitable for duties involving

solids and abrasives (see Figure 15.45)

Foil-wrapped deformable Foil crinkled, twisted and asbestos core folded upon itself

Corrugated foil, concertina wound

Figure 15.44 Typical metal foil-based packing construction

Figure 15.45 Double-bevelled, elastomer-proofed fabric packing for abrasive duties

15.2.1.5 Type of wear

Although a typical set will be five rings, experience has shown that there is not a linear pressure drop through/across the five rings The majority of the effective sealing is done by the rings adjacent to the gland spigot, Le on the atmosphere side of the set - as indicated in Figure 15.46

The two rings on the pressure side will often be virtually uncompressed and still contain ample lubricant In contrast, the rings at the gland/atmosphere side will invariably have lost all lubricant, be very hard and probably have suffered around 50% compression - it is these rings which have been doing most of the sealing In desperate situations replacement of these two - or possibly three - rings can often restore perfor- mance and leakage to an acceptable level

The quantity of packing to use and its size for a given application relies largely on the experience of the user/manufacturer in the type of duty being performed, or in liaison with a packing supplier at the design stage The latter course of action is always to be favoured if any doubt exists, since an exact knowledge of the capabilities and limitations of the material employed can be found only with those specialists responsible for compounding and production

Seals and sealing 15/23

Five rings of square section packing are often accepted as a

sufficient number for the average uncomplicated duty but

there are many pump applications where the presence of a

lantern ring or similar consideration may dictate a greater

quantity

The appropriate packing section to use in relation to

diameter is open to a degree of individual preference but

broad recommendations are shown in Table 15.5 To give an

idea of the capabilities of the various materials and construc-

tions of soft packings which are readily available, reference

may be made to Table 15.6 (suitability in different

media/speed and temperature limits) Table 15.7 (comparative

speed performance), Table 15.8 (comparative temperature

performance) and Table 15.9 (comparative cost indication)

(Note: The statement of speed and temperature limits for a

given material should not be construed as meaning that a

packing will be suitable for duties where such maxima are

jointly encountered )

15.2.1.6 Reciprocating pumps' duties

Much of the above information also relates to reciprocating

pumps 'While many years ago compression packings were

used on such pumps, many engineers today would automa-

tically think of using a more moldern multi-lip type seal

Howevei , for the more difficult applications compression

packings are now being used again

The most important difference from rotary applications is

that for these reciprocating duties the packing is generally

subjected to far higher operating pr'essures, and so requires a

proportionately greater degree of gland spigot load-

inghompression in order to develop sufficient sealing force

This extra compressive force combined with the frictional drag caused by the reciprocating movement means that extrusion of the packing into the gland bush clearance is often the main cause of seal failure While the new synthetic yarns are extremely strong they are generally very small and so tend to extrude more easily Once extrusion has occurred this will invariably stop any leakage completely and excessive tempera- tures are developed very quickly

Often the operating clearance between the gland spigot and the ram is too wide to prevent extrusion and so some form of anti-extrusion element must be incorporated into the sealing arrangement One very simple and very effective method is to use a hard fabric ring as is normally employed as the support ring of a typical set of multi-lip seals - as shown in Figure 15.47(a) Because the lip makes intimate contact with the ram and the wall of the stuffing box this contains the packing perfectly and prevents extrusion

Many pumps are still fitted with multi-lip seals of the chevron type Generally, these will operate well providing the conditions are within their capability On many of the larger high-speed pumps, however, the frictional heat generated by the pressure and rubbing speed causes the rubber proofing of the seals to carbonize and soon the flexibility is completely lost and the set becomes virtually a solid mass which will not then respond to further gland adjustment It is for these more

arduous pump duties that a change back to compression packing is now occurring and is proving very successful

On modern pumps, particularly of the smaller sizes, there is

a trend towards non-adjustable glands Compression packings cannot be considered for such applications unless some form

of spring loading is incorporated In this respect the arrange- ment as shown in Figure 15.47(b) has proved very successful

RECIPROCATING, ROTARY

PUMPS AND VALVES

Lubricated aluminium foil

Lubricated braided asbestos

Lubricated plaited asbestos

Plaited, lubricated asbestos impregnated with PTFE dispersion

Plaited lubricated asbestos impregnated with PTFE

PTFE impregnated asbestos and glass fibre yarns with

PTFE yarn impregnated with PTFE dispersion and inert

Soft lead-based foil wrapped round lubricated asbestos core

Hydrocarbon-resistant lubricated plated asbestos

Lubricated plaited cotton

Cross-plait aramid fibre yarns

Pure graphite foil with no volatile additives

RECIPROCATING PUMPS AND VALVES

Monel wire reinforced asbestos cover with plastic core *

Synthetic rubber bonded braided asbestos with brass wire

Lubricated plaited flax

VALVES ONLY

Constructed from a jacket of asbestos reinforced with inconel

Lubricated braided asbestos with rnonel wire reinforcement *;

Self-lubricating fibrous asbestos with flake graphite or mica

Lubricated braided asbestos and brass wire reinforced *

PTFE yarn impregnated with PTFE dispersion and inert

Unsintered PTFE cord gland seal for rapid valve packing

dispersion but with no additional lubricant

suitable lubricant

lubricant

reinforcement

wire braided over a resilient asbestos core *

lubricant but with no additional lubricant

Seals and sealing 15/25

Table 15.7 Maximum rotary speeds for pump packings

Table 15.8 Maximum service temperatures of pump packings

ted ramie yarn (Plaited)

ted PTFE yarn (Plaited)

ted lead foil (Foil-wrapped

I-purpose lubricated asbest,

ted aluminium foil (All ty

0 100 200 300 400 500 600

15/26 Plant engineering

Table 15.9 Typical relative costs of pump packings

I

Figure 15.47

anti-extrusion ring; (b) spring-loaded packing configuration for

non-adjustable glands

(a) Packing configuration with lip-profiled fabric

Although the springs provide the basic force on the packing,

during the pressure stroke the fluid acts over the full radial

width of the metal ring and this then provides the extra

‘hydraulic’ force with which the conventional adjustable gland

would have to be developed by appropriate gland loading

15.2.1.7 Fitting

It is often assumed that unskilled labour can be used to repack pump glands but this is true only so long as ‘unskilled’ is not equated with ‘unaware’ Familiarity with the following ideal procedure will be more than repaid in terms of trouble-free packing performance:

Where length form is used:

1 Spirally wrap the material around a rod of diameter

2 Cut the required number of rings cleanly to obtain good

3 Proceed as for pre-formed split packing rings

equivalent to the pump shaft

butt-joins (see Figure 15.48)

Seals and sealing 15/27 tained Approximately 15 minutes should be :eft between successive adjustments Do not overtighten

8 Where loose-form material is used for valves proceed as (1) and (2) and tamp packing into a dense homogeneous mass, progressively filling the housing to the required degree

9 Because of the danger of corrosion through electrolytic action, packings containing graphite should be avoided on valves or pumps with stainless or chrome steel stems This risk is most acute when the packing remains in the gland during storage and is particularly aggravated by the pres- ence of moisture

iere pre-formed split packing rings are used:

Carefully remove old packing (including, where

appropriate, the packing on the far side of a lantern ring)

Thoroughly clean all surfaces that will contact the packing

and, where permitted, smear with oil Gland and neck

bushes, shaft surface and bearings should also be checked

for signs of wear and rectified as necessary

Place first ring over the shaft by opening to an ‘S’

confnguration to ensure that bending effects are spread

over the whole ring (see Figure 15.49)

Insert firs; ring into stuffing box and lightly bed in with a

split (wooden) distance piece and gland spigot With

plaited packing the ‘v’ formation on the outside diameter

of the ring should be pointing in the direction of shaft

rotation (see Figure 15.50)

Repeat ( 3 ) and (4) with remainder of rings ensuring that

each ring is firmly seated and that the butt joins are

staggered by at least 90” (Note: The rings must be fitted

individually and under no circumstances should complete

sets be fitted as a unit.)

When the requisite number of rings have been fitted,

tighten gland nuts until the shaft or spindle torque

increases Then slack off gland and pull up to finger

tightness only (If pump is to be stored before use leave

g l a d slack so that packing resilience is not impaired.)

Running-in pumps

Prime casing and run pump up to operating speed for

IC15 minutes If pump is not fitted with gland cooling, a

cold-water spray over the gland housing will avoid excess-

ive heat build-up during this stage If no leakage occurs,

stop pump, vent casing pressure and slacken gland further

Repeat until leakage starts

The controlled leakage, essential for lubrication pur-

poses, can then be obtained by running the pump and

evenly tightening the gland nuts in increments of two flats

until approximately one drop every few seconds is ob-

15.2.1.8 Fault finding

A major advantage of compression packings is that breakdown

is rarely sudden or catastrophic but rather a matter of a gradual build-up of leakage until an unacceptable level is reached Normally, considerable life can be achieved by controlling leakage with further tightening of the gland nuts

(Note: the seepage of fluid which acts as a lubricant for the packing on rotary applications should not be confused with leakage and the rate of one drop every few seconds should be maintained.) However, if other than routine maintenance or just plain ‘fair wear and tear’ are suspected as the cause of leakage and the need for repacking, then the following hints could well prove useful:

1 Confirm that the packing is rated as suitable for the application

2 If one or more rings are missing from the set, check for excessive neck bush clearance allowing extrusion of rings into the system If the top ring has extruded between the gland follower, anti-extrusion rings could avoid replace- ment of metal parts

3 If the packing’s radial thickness appears diminished in one

or more places, check for an undersize shaft or badly worn bearings which could cause shaft whip or spindle wobble

4 If radial section of packing directly beneath the shaft is reduced or premature leakage occurs along the top of the shaft, check for misalignment of shaft centre to stuffing box bore

5 If the packing is worn on the outer diameter, check for loose rings or rings rotating with the shaft due to insuffi- cient gland load

6 If the packing rings have bulges on their radial faces the adjacent ring was probably cut too short, causing packing under pressure to be forced into the gap at the joint

7 If the packing nearest the gland spigot shows excessive deformation while other rings are in fair condition, the set was probably incorrectly installed and subjected to excess- ive gland tightening (Note: Overtightening is usually the greatest single cause of premature packing failure.)

8 If the cause of your particular problem is still not apparent, give equal attention to i.d and o.d leakage and check for

a rough stuffing box bore before seeking specialist advice

15.2.1.9 Standardization

In the interests of stock control it is clearly an advantage to rationalize the variety of packings used in any plant to that minimum number which will effectively cater for all the conditions likely to be encountered If cost is no object, then there are single, sophisticated materials and constructions that will go some way towards satisfying most demands, but it is doubtful if cost effectiveness could be justified Far better to compromise on a small number of reputable products deve- loped for the areas in question, e.g pumps, valves, etc

15/28 Plant engineering

There are few standards applying to compression packings

on a national or international basis, although many individual

companies and organizations have domestic standards which

have, in many cases been the subject of collaboration be-

tween user and packing manufacturer BS 4371: 1968 specifies

minimum standards for lubricated plaited cotton, lubricated

plaited flax, lubricated plaited or braided asbestos, dry white

nonmetallic plaited or braided asbestos, plaited or braided

asbestos, metallic wire reinforced, indurated asbestos, and

lubricated fibrous asbestos and gives guidance on limiting

operating parameters for these constructions There are other

pump and valve British Standards which specify common

stuffing box dimensions such as BS 1414: 1975, BS 1873: 1975

and BS 3808: 1964

Where packings are required for service with potable water

in the water authority distribution system (which covers

reservoir to tap), only those materials which have gained a

National Water Council Approval may be used Such products

have been tested to establish that they produce no colour,

taste or turbidity, are non-toxic and will not support microbial

growth

Statutory Instruments 1978, No 1927, The Materials and

Articles in Contact with Food Regulations 1978 required that

compression packing materials, for example,

‘ do not transfer their constituents to foods with which

they are, or likely to be, in contact, in quantities which

(1) endanger human health or

(2) bring about a deterioration in the organoleptic [sensory]

quality of such food or an unacceptable change in its

nature, substance or quality.’

Such regulations inevitably restrict the range of available

materials and lubricants Consultation with the supplier is

recommended to establish preferred grades

could -

15.2.1.10 Compression packings: material developments

Introduction The route to improve performance in most

areas of sealing is generally in the use of new/improved

materials - this is without doubt the case with compression

packings In recent years what can loosely be termed ‘synthe-

tic materials’ have become available which have changed the

overall attitude of the engineer regarding the use of the

packed gland These have three distinct advantages over the

old-established materials:

1 The performance capabilities have been extended in terms

of pressures, speed and temperature limitations

2 Combined with these the resultant life has also been

significantly increased

3 The age-old problem of repeated gland adjustments has

been reduced to the extent that on many applications it has

been virtually eliminated - the ‘non-adjustable’ packed

gland is now a reality

Four materials have achieved these improvements and al-

though not exactly new, PTFE is included, since by compari-

son with the natural fibres of asbestos and cotton, etc it is

relatively new and does occupy an important place in today’s

range of compression packings

PTFE Polytetrafluoroethylene yarns provide soft packings

for services where corrosive media are being handled or

freedom from contamination is an essential requirement A

semi-rigid fluorocarbon plastic, PTFE is unique in possessing

almost complete chemical resistance within its temperature

range which, in this field of application, spans the cryogenic

area to 250°C Another major advantage refers to its very low coefficient of friction Lubricated plaited PTFE yarn packings are suitable for rotary surface speeds up to 8 m s-l and are also finding increasing acceptance on high-speed, high- pressure, multi-ram reciprocating pumps In solid form this material is not acceptable as a compression packing due to poor creep properties and lack of resilience However, solid junk rings or spacers in PTFE are often used to enhance packing performance on arduous pump duties

Aramid fibre Packings made from aramid fibre, usually of a distinctive yellow colour, are becoming increasingly popular for a variety of pump and valve services hitherto satisfied by PTFE-lubricated asbestos packings It has a high tensile strength, excellent resilience and thermal stability up to 250°C and is resistant to a wide range of chemicals Experience has shown that while it is extremely resistant to abrasives, it is also extremely tough such that it can sometimes severely dam- agelwear the surface of a rotating pump shaft However, the latest quality yarns combined with an improved lubricant have reduced this problem and speeds of 20 m s-l have now been successfully achieved It has also proved to be successful in high-pressure reciprocating pump applications, being used typically with arrangements as shown in Figures 15.47(a) and (b)

Graphite Without doubt, this material more than any other

is responsible for extending the performance of the packed gland It is available in various forms, each having advantages and, hence, specific areas of use:

Graphite yarnlfilaments Yarn packings in this material are a development for rotary pump applications and provide possibilities for extending the range of the packed gland beyond boundaries hitherto estab-

lished A high coefficient of thermal conductivity, low friction

and resistance to chemical attack are the useful characteristics

of this material Temperatures up to 400°C may be considered

If a good performance is to be obtained, then close attention must be paid to mechanical conditions such as shaft run-out and finish Care in fitting and running-in is also mandatory

Expanded graphite foil is the most recent and significant

application of graphite, particularly in the context of valve applications Expanded graphite materials combine the well- established thermal and friction characteristics, long asso- ciated with the correctly developed use of carbon-based products, with a unique flexibility and resilience The at- tributes of this exfoliated form of graphite bear recording Excellent resistance to compression set resulting in little loss of radial gland force or flange seating stress over long periods (see Table 15.10)

0 No loss of volatiles even at high temperature thus minimiz- ing frequency of gland adjustment

Resistance to a wide range of chemicals

0 High-temperature capability, particularly in non-oxidizing High thermal conductivity

Low friction properties - self-lubricating Exceptionally low chloride content

No adhesion or corrosion problems Fire-safe

Figure 15.51 shows an interesting comparison of performance

on a test gland between half-rings of expanded graphite and a lubricated asbestos yarn packing Not only did the former require fewer gland adjustments during the period of testing but the average leakage rate was much less - to the point of running virtually dry for protracted periods

environments

Seals and sealing

Rather than use the tape form of expanded graphite which is primarily a useful maintenance expedient, moulded rings to a selected and controlled density should be the first choice Although more costly than conventional packing materials, economies of radial width and number of rings used are feasible, quite apart from the performance advantage likely io

be derived from the use of expanded graphite

Hybrid gruphite/PTFE yarn This latest material is a yarn combining PTFE and graphite; the PTFE is not added to the yarn as was the case with many other materials in the past This proprietary blended yarn based on an exclusive process has achieved much success on high-speed rotating pump duties

at speeds in excess of 25 ms-' Many manufactarers are now offering products based on this material, which is particularly insensitive to variations in fitting techniques, etc Such pack- ings bed-in easily and have characteristically lower leakage rates than many conventional variants

Table 15.10

Expanded Expanded Compresr;ibility/recovezy graphite graphite

A S T M F36-66 Procedure H 1 mm thick 2 mm thick

_._ Expanded graphite rings (in halves)

a Lubricated asbestos yearn rings (split)

15/30 Plant engineering

Since these yarns are extremely small in size, extrusion can

occasionally be a problem Hence when using these new

products, extra attention should be given to the extrusion

clearance of the gland and neck bushes

Due to its PTFE content, service temperatures are limited

to 260°C However, this is well below the operational gland

temperature of most rotary shaft applications

One cannot leave materials without special reference to the

vital role played by asbestos - a much-denigrated mineral

fibre without which economic and practical solutions to many

sealing problems would not be feasible Although understand-

able, in an age of correct awareness of health and safety

matters, the over-reaction against asbestos has revealed many

inconsistencies Motor manufacturers may prohibit its use as a

plant-maintenance material but continue to use asbestos in a

brake-lining and clutch-facing role where residual dust is

evident Some users may seek to limit its application in a safe

form as a valve packing or gasket but perpetuate its specifica-

tion for fireproof positions and roofing where the mineral is

cut in a dry form

Hazards exist but adherence to basic advice on handling

asbestos will result in a sensible balance between prohibition

and practicality In this context, users of asbestos-based pump

and valve packings, gaskets or allied components might heed,

to advantage, the statement issued by the Asbestos Informa-

tion Committee to the British Valve Manufacturers’ Associa-

tion 18-22

15.2.1.11 Selection

Selection is the most vexing question as, for many duties, so

many reasonable alternatives exist Much will depend on

personal experience, frequency of maintenance, original cost

level, contamination considerations, size, etc

In the case of a manufacturer producing large quantities of

valves or pumps to standard dimensions there is much to be

said for purchasing sets or rings rather than length-form

packing With the techniques available, packing can be

supplied ready for immediate fitting with substantial reduction

in that overall cost represented by receiving length-form that

must be cut to size by skilled personnel This economy is not

confined to the large manufacturers but it is they who will

enjoy the greater advantage

On the other hand, in many instances, the problem of

stocking rings or sets tailormade for an assortment of valves

varying in origin, type and dimensions can prove intolerable

For these cases, there is a clear need for the versatility of

packing in length-form Comparable with this solution is the

expedient provided by those packings of plastic nature that are

available in loose form but this advantage must be weighed

against the labour cost in the careful fitting required

15.2.1.12 Cross-plait constructionlmixed yarns

Virtually all these new synthetic yardfilament materials are

manufactured into length-form packings using the ‘cross-plait’

construction as shown in Figure 15.43, which results in a far

better and more uniform construction and it does have the

distinct advantage of enabling two different yarns to be used in

the manufacture of a packing This technique has resulted in

the development of a new generation of cross-plait mixed-yarn

products, which can maximize the benefits of the individual

constituents and minimize their shortfalls Typical examples

are packings deploying tough aramid (yellow) yarn at the

corners and the high-speed thermally superior hybrid yarn just

described to most of their rubbing face - these offer excellent

extrusion resistance and a higher-speed capability than a plain

aramid product Also for chemical compatibility PTFE fila- ments can, of course, be combined with others as necessary

so a reduction in the radial sealing force occurs, resulting in an increasing leakage Gland adjustment, i.e compression, then restores the density and the sealing force

The new materials, particularly the graphite types, are self-lubricating and so very little lubricant is added to the packing In consequence, because there is relatively no ma- terial or lubricant to be lost or forced out of the packing, the amount of subsequent gland adjustments are very much reduced, giving a longer life and reduced maintenance re- quirements This aspect is demonstrated perfectly with the graphite foil used for valve sealing The material is virtually pure graphite and is perfectly stable in size and density within the temperature range of -200 to +550”C, and once fitted and correctly adjusted it will usually provide years of trouble- free service

15.2.1.14 Applications

Basically, the fundamental choice is generally that for a dynamic/rotary application the length-form cross-plaited yardfilament product should be used and for valve spindles rings of expanded graphite are the first choice

15.2.1.15 Valves

Expanded graphite is now the most common form of valve stem seal It is the standard for virtually all valves in the power generation and nuclear industries, both in the UK and in Europe Density of the material is very important and usually

a density of 1.6 g/cm3 is used This is the value specified by CEGB in their specification for this material (No 155701) For this reason, it is always preferable to use moulded rings for critical applications rather than the alternative ‘tape’ form of material The rings are manufactured by winding tape coax- ially around a rod until the required o.d is obtained and then compressing these within a mould The material is usually compressed in the order of approximately 60%

Once installed and the optimum gland adjustment is applied, no further adjustments should be necessary The only reason subsequent adjustments are necessary is because the sealing force has reduced - this can only happen if the density has changed, which in turn can only be the result of some loss

of material, i.e extrusion of the top or bottom rings

It is therefore extremely important to have the minimum possible extrusion clearances However, if extrusion proves to

be a problem, then the use of end rings of graphite filament (i.e cross-plaited type) have proved extremely successful in solving this problem Indeed, the use of these ‘combination sets’ is now becoming far more common and in the power- generation industry in France it is virtually a standard approach

Compared with the older materials, these new sets require negligible subsequent gland adjustments and the use of ‘live’ spring-loaded glands has now achieved the ultimate in long trouble-free operations The reason ‘live’ glands can now be used is due to the small degree of compressibility of these graphite materials Once fitted, these sets will probably only compress by about 8-10% at most, whereas the older sets in