Purdue University

Purdue e-Pubs
International Compressor Engineering Conference School of Mechanical Engineering

1988

Valve Losses in Reciprocating Compressors

Friedrich Bauer
Hoerbiger Ventilwerke AG

Follow this and additional works at: http://docs.lib.purdue.edu/icec

Bauer, Friedrich, "Valve Losses in Reciprocating Compressors" (1988). International Compressor Engineering Conference. Paper 631.
http://docs.lib.purdue.edu/icec/631

This document has been made available through Purdue e-Pubs, a service of the Purdue University Libraries. Please contact epubs@purdue.edu for
additional information.
Complete proceedings may be acquired in print and on CD-ROM directly from the Ray W. Herrick Laboratories at https://engineering.purdue.edu/
Herrick/Events/orderlit.html
 VALVE LOSSES IN RECIPROCATING COMPRESSOR
S

Fried rich Bauer

Hoerb iger venti lwerk e AG
Vienn a, Austr ia

ABSTRACT
Over all flow losse s of comp ressor valve
not only by the valve geom etry but also s are influe nced
pocke t losse s. are ampli fied by valve

The magn itude of the pocke t losse s was

typic al pocke t shape s. Key varia ble measu red for four
param eters that influe nce
the losse s were estab lishe d.
The resul t is shown in "Pock et-Fa ctor"
shall perm it optim izatio n of the pocke diagra ms which
stage of cylin der desig n. t shape at the conce ptual

INTRODUCTION
The effor t to reduc e energ y losse s as
all techn ical proce sses requi res much as possi ble in
knowl edge of the reaso ns for thesea losse
funda menta l and detai led
respe ctive share of loss contr ibuto rs. s as well as of the
recip rocat ing comp ressor s are, by now, The basic losse s in
knowl edge. The press ure and subse quent common and well defin ed
gas flow in and out of the cylin der have energ y losse s cause d by
refer red to as "valv e losse s". gener ally been

Becau se of these exist ing valve losse s,

conti nuous effor t throu ghout the indus it has been a
try to desig n more
effic ient autom atic plate valve s for comp
been worki ng in this area for a long resso rs. Hoerb iger has
minim ize the flow resis tance and lossetime and is striv ing to
s of the valve .
Aside from the valve itsel f, the imme
the valve also contr ibute a signi fican diate surro undin gs of
losse s as the gas enter s and exits the t amoun t of the press ure
cylin der.

v
Flg.1 : PV Diagr am at a ~:ompresslan cylin der

263
 e diagra m with the
Figur e 1 shows a typic al press ure volumarge ·line press ures
disch
cylin der press ure (P), the sucti on and as the press ures
(from press ure gauge s Ps and P0 ) as welland PKD) durin g a full
(PK~
measu red in the valve chamb ers
comp ressio n cycle .
the vario us press ure
In the diagra m the areas comp rised by
curve s repre sent the work perfo rmed in a comp ressio n cycle .
losse s, requi res
This work, in conju nctio n with the valve
close r scrut iny:
and PKD - Pp)
The lightl y-sha ded loss areas (P 5 s.- PKs
canno t be attrib uted to the valve
They are cause d by restr ictio ns betwe en the valve chamb er
sucti on line or disch arge line 1 recei ver,
and the P~ have been
respe ctive ly, where the press ures Ps and
and by dynam ic press ure varia tions outsi de the
measu red
cylin der.
sent the losse s
The heavi ly-sh aded areas in Figur e 1 repre valve pocke ts.
cause d by the valve s mount ed in their
t and cagin g are
Besid es the valve itsel f, the valve pocke
other subst antia l loss contr ibuto rs.
itsel f is
The press ure loss of the comp ressor lvalve with the valve
determ ined exper iment ally in a wind tunne
ndicu lar to the
moun ted on a level surfa ce locat ed tsperpe produ ce well- defin ed
direc tion of flow. These measu remen press ure loss throu gh the
coeff icien ts of flow with which the lation exclu des the effec ts
valve can be calcu lated . This calcu der.
of the rest of the flow system in a cylin
ure losse s in the
As furth er tests have shown , the press
valve s mount ed in valve pocke ts can, at times , be great er by a
s determ ined throu gh the
magn itude of two or more than the losse
afore menti oned tests .
PROCEDURE
a given mass flow
lf the ratio of the press ure drop t;, P at ure drop at the same
a pocke t-mou nted valve to the press
throu gh ed as the "Pock et
mass flow throu gh the valve alone is defin of the valve insta lled in
Facto r" (PF), then the press ure loss
area) can be calcu lated
the cylin der pocke t (the heavi ly-sh aded g the press ure loss of
at norma l condi tions by simpl y multir plyin
the valve alone by the pocke t facto PF.
2
A P valve + pocke t ¢valv e alone )
PF ( ¢valv e + pocke t
P valve alone
A
area,
wher e¢ "' A/{:f and repre sents the equiv alent valve
A is the free lift area and 5 is the flow
coeff icien t.

By integ ratio n it follow s that

LOST WORK valve + pocke t
PF ::;,~
LOST WORK valve alone

264
 As the losses at the valve and pocket can be reduced by
"tuning" the shape of the pocket to the valve, measureme nts of
four different commonly used valve pocket configura tions (see
Figure 2) have been made to determine "pocket factors" as a
function of the most important geometry parameter s.

1.)Valve pocket with 2.lVolve pocket with

round passage slot-shape d passage

lxw

3.lValve pocket axis

perpendic ular ~.JValve pocket oxis
to cylinder axis, p or a l l e l to
semicircle passagE cylinder axis,
moonshape d passage

f tat sur face

cylinder dia
cylinder axis

cylinder dla

F I g • 2- A l tern at i v e con t i g u rat i on s o t c y L i n de r pockets

This research shows that the pocket factors for suction

and discharge valves with the same pocket configura tion differ
due to the opposite directions or flow.
Extensive data were accumulate d for each of the pocket
configura tions and the results were compiled into various
diagrams shown in Figures 3 through 6. All plotted values are
based on actual tests.

265
Pocke t Configu~ation 1

PF PF
SUCTION POCKET DISCHARGE POCKET

1+-----~-----+----~~----~
1 2 3 4-

Flg,3-Po~ket fnc:to r for confi gura tion

s (PF) vs. the

Figure 3 shows graph s of the pocke t facto~ uratio n at a
config
flow passag e geome try (FI¢vl for the first ing va~iables were
suctio n and a discha rge pocke t. The followation:
consid ered essen tial in the overa ll evalu

F = the area of pocke t passag e

the ratio of this area to the equiv alent

valve area

c/D the ~atio defini ng the valve distan ce

icient s of
It is evide nt from these graph s that theofcoeff
the fi~st
tions
flow fo~ the va~ious geomet~ic condi valve pocke ts and for
confi gurat ion are differ ent fo~ suctio n s and pocke ts have the
disch arge valve pocke ts even if the valve pocke t in this case is
same dimen sions. The PF for the suctio n rge pocke t.
somew hat highe r than the PF fo~ the discha

266
Pocket Configuratio n 2

E .':1+ (1-J..~ )tan ex

D D 2 D

A
.6

v
v
.4- /
./
.2 / "~·-·

0 •2 .4- .6 .8 1 w/l

PF

Flg.4--Pack et factor for configurat ion 2

A comparison of configuratio ns 1 and 2 shows that at an

equal ratio of F/~ and equal c/D a slot-shaped passage is
better than a round-shaped passage.

The same observation holds for the discharge pocket. The

PF is smaller for the discharge pocket than for the suction
pocket.

267
Pocket Configuratio n 3

PF
1 .5+-----~-------.------~-----,.-~

.02 .06 .1 • 1 4- .18 .2

c/D

Ffg.S-Pock et factor for configurat ion 3

In configuratio n 3 the eenterline of the valve pocket

coincides with the bottom wall of the cylinder end. Half of the
valve is in the cylinder head which prc;>vides a .sloped relief
towards the cylinder. The only variable ia this case is the
angle of the valve pocket re,cess. Subsequently , there is only
one curve shown for the suction pocket and one curve for the
discharge pocket. The pocket factors are smaller in this
configuratio n than in the previous two configuratio ns. In this
case, the PF for the suction pocket is smaller than the PF for
the discharge pocket.

268
Pocket Configura tion 4

ql
.8~~.--.--,---·~-,

1.2 1.4.
0/d

CD+2e-dlt ono:
c=o+
5
PF PF

SUCTION POCKET

Flg.6-Poc ket foetor for configura tion 4-

Configura tion 4 gives a variety of results also. A drawing

of the overlappin g of the cylinder and the valve is provided to
help determine the valve exposure F = D". 'f. The value 'f is
determined by the relationsh ip of ene valve diameter to the
cylinder diameter and by the eccentric ity of the valve to the
cylinder.
Figure 6 shows the previously taken approach plotting PF
dependent of F/¢v and c/D for the sickle-sha ped portion of the
pocket.
Pocket factors for practical valve exposures in Figure 6
are comparable to results found in pocket configura tions 1 and
2.

PF for discharge valves exceed those for suction valves.

269
 CONCLUSIONS
g
In designin g the passage s and valve pocket shapes the followin
observa tions should be noted:
During the short periods when the valve is open, the gas
flows through a number of passage s (i.e., pipes, flanges
on cylinde rs, cylinde r passage s, valve cages, valves,
valve pockets) until it reaches the cylinde r itself. All
these differe nt areas impose changes in velocity and
directio n which result in pressur e losses.
Therefo re, when speaking of "valve losses" , it has to be
kept in mind that the lightly- shaded areas in Figure 1 are
caused by element s outside of the valve chamber and that
the heavily- shaded areas correspo nd to the real valve
losses multipl ied by the pocket factor (PF) which depends
on the shape of the valve pockets .
The valve contain s the only cross-s ection in the whole
system which can be defined as a "dynami c" cross-s ection.
Its cycle goes from being totally closed to totally open
to totally closed during every revolut ion. The valve
motion must be of a nature that no adverse , prematu re
failure occurs. If this function must be perform ed exactly
for a prolong ed period of time, it require s conside rable
design and technol ogical efforts .
The valve area, therefo re, can be regarded as the most
expensiv e area in relation ship to the valve surroun ding
non-var iable areas. For instanc e, one square inch of ing
increase d valve area can be obtaine d only by develop
rather complic ated designs , by opting for a low valve
life, or by utilizin g expensi ve materia ls.
Optimiz ing thE: flow area around the valve is often
attainab le with little added cost and, once properly
designe d, these flow passage s are not affected by dynamic
conditio ns.
Econom ically and technic ally, it would be wrong to size
the " inexpen sive areas" the same or only slightly larger
are
than thE> valve area, conditio ns which, unfortu nately,
too common. The effort must be focused on increasi ng the
"inexpe nsive areas" as much as possibl e in order to
achieve minimum overall losses.
When designin g new cylinde rs, three points have to be
conside red concern ing the valve pockets :
l) Clearan ce volume
2) Total flow resistan ce
3) Economi cal manufac turing
capacity
The influenc e of the clearanc e volume· on the cylinde r permit us
can easily be calcula ted. The diagram s of this study
a
to determin e approxim ately the total flow resistan ce of of
cylinde r at the early design stage. This is certain lyce when
advanta ge as energy losses are of increas ing importan
analyzin g compres sion work.

270