Auld, 1998-2012 : www.aerodynamics4students.

com
Analysis of Propellers
Glauert Blade Element Theory
A relatively simple method o predictin! the perormance o a propeller "as well as ans
or windmills# is the use o $lade %lement &heory. 'n this method the propeller is divided
into a num(er o independent sections alon! the len!th. At each section a orce (alance
is applied involvin! 2) section lit and dra! with the thrust and tor*ue produced (y the
section. At the same time a (alance o a+ial and an!ular momentum is applied. &his
produces a set o non-linear e*uations that can (e solved (y iteration or each (lade
section. &he resultin! values o section thrust and tor*ue can (e summed to predict the
overall perormance o the propeller.
&he theory does not include secondary eects such as ,-) low velocities induced on the
propeller (y the shed tip vorte+ or radial components o low induced (y an!ular
acceleration due to the rotation o the propeller. 'n comparison with real propeller results
this theory will over-predict thrust and under-predict tor*ue with a resultin! increase in
theoretical eiciency o -. to 10. over measured perormance. /ome o the low
assumptions made also (rea0down or e+treme conditions when the low on the (lade
(ecomes stalled or there is a si!niicant proportion o the propeller (lade in windmillin!
coni!uration while other parts are still thrust producin!.
&he theory has (een ound very useul or comparative studies such as optimisin! (lade
pitch settin! or a !iven cruise speed or in determinin! the optimum (lade solidity or a
propeller. 1iven the a(ove limitations it is still the (est tool availa(le or !ettin! !ood
irst order predictions o thrust, tor*ue and eiciency or propellers under a lar!e ran!e
o operatin! conditions.
Blade Element Subdivision
A propeller (lade can (e su(divided as shown into a discrete num(er o sections.
2or each section the low can (e analysed independently i the assumption is made that
or each there are only a+ial and an!ular velocity components and that the induced low
input rom other sections is ne!li!i(le. &hus at section AA "radius 3 r# shown a(ove, the
low on the (lade would consist o the ollowin! components.
V
0
-- a+ial low at propeller dis0, V
2
-- An!ular low velocity vector
V
1
-- section local low velocity vector, summation o vectors V
0
and V
2
/ince the propeller (lade will (e set at a !iven !eometric pitch an!le " 0 # the local
velocity vector will create a low an!le o attac0 on the section. 4it and dra! o the
section can (e calculated usin! standard 2-) aerooil properties. "5ote: propellors use a
chan!ed reerence line : 6ero lit line not scetion chord line#. &he lit and dra!
components normal to and parallel to the propeller dis0 can (e calculated so that the
contri(ution to thrust and tor*ue o the compete propeller rom this sin!le element can
(e ound.
&he dierence in an!le (etween thrust and lit directions is deined as
=0o
&he elemental thrust and tor*ue o this (lade element can thus (e written as
AT=ALcos()ADsin() ,
AQ
r
=ADcos()+ALsin()
/u(stitutin! section data "C

and C
!
or the !iven o # leads to the ollowin! e*uations.
AL=C
L
1
2
jV
1
2
.c.dr , AD=C
D
1
2
jV
1
2
. c.dr per (lade
where j is the air density, " is the (lade chord so that the lit producin! area o the
(lade element is "#dr.
' the num(er o propeller (lades is "B# then,
AT=
1
2
jV
1
2
c(C
L
cos ()C
D
sin()). B.dr
"1#
AQ
r
=
1
2
jV
1
2
c(C
L
sin ()+C
D
cos ()) . B. dr
AQ=
1
2
jV
1
2
c(C
L
sin()+C
D
cos()). B.r.dr
"2#
$nflo% &a"tors
A ma7or comple+ity in applyin! this theory arises when tryin! to determine the
ma!nitude o the two low components V
0
and V
2
. V
0
is rou!hly e*ual to the aircrat8s
orward velocity " V

# (ut is increased (y the propeller8s own induced a+ial low into a

slipstream. V
2
is rou!hly e*ual to the (lade section8s an!ular speed " Dr # (ut is
reduced sli!htly due to the swirlin! nature o the low induced (y the propeller. &o
calculate V
0
and V
2
accurately (oth a+ial and an!ular momentum (alances must (e
applied to predict the induced low eects on a !iven (lade element. As shown in the
ollowin! dia!ram, the induced components can (e deined as actors increasin! or
decreasin! the ma7or low components.
A typical streamtu(e o low passin! throu!h section AA would have velocities
V
0
=bDr , V
X
=V

(1+a)
/o or the velocities V
0
and V
2
as shown in the previous section low dia!ram,
V
o
=V

+a.V

where a is the a+ial inlow actor

V
2
=DrbDr where b is the an!ular inlow actor "swirl actor#
&he local low velocity and the an!le o attac0 or the (lade section is thus
V
1
=
.
V
o
2
+V
2
2
",#
o=0tan
1
(
V
o
V
2
)
"4#
A'ial and An(ular &lo% Conservation of )omentum
&he !overnin! principle o conservation o low momentum can (e applied or (oth a+ial
and circumerential directions.
2or the a+ial direction, the chan!e in low momentum alon! a stream-tu(e startin!
upstream, passin! throu!h
the propeller at section AA
and then movin! o into
the slipstream, must e*ual
the thrust produced (y this
element o the (lade.
&o remove the unsteady
eects due to the
propeller8s rotation, the
stream-tu(e used is one
coverin! the complete area
o the propeller dis0 swept
out (y the (lade element
and all varia(les are
assumed to (e time
avera!ed values.
A & 3 chan!e in momentum low rate
3 mass low rate in tu(e + chan!e in velocity
=j2nr . dr V
o.
(V
s
V

)
$y applyin! $ernoulli8s e*uation and conservation o momentum, or the three separate
components o the tu(e, rom reestream to ace o dis0, rom rear o dis0 to slipstream
ar downstream and (alancin! pressure and area versus thrust, it can (e shown that the
a+ial velocity at the dis0 will (e the avera!e o the reestream and slipstream velocities.
V
o
=(V

+V
s
)/ 2 , which means that V
s
=V

(1+2a)
&hereore
AT=j2nr V

(1+a) . (V

(1+2a)V

) . dr
AT=j2nrV

2
(1+a).2a.dr
AT=j4nrV

2
(1+a). a.dr
"-#
2or an!ular momentum
A 9 3 chan!e in an!ular momentum rate or low + radius
3 mass low rate in tu(e + chan!e in circumerential velocity + radius
AQ=j2nr dr V
o.
(V
0
(slipstream)V
0
( freestream)) .r
$y considerin! conservation o an!ular momentum in con7unction with the a+ial velocity
chan!e, it can (e shown that the an!ular velocity in the slipstream will (e twice the
value at the propeller dis0.
V
0
( slipstream)=2bDr
V
0
( freestream)=0
&hus
AQ=j2nr V

(1+a)(2bDr). r.dr
AQ=j4nr
3
V

(1+a). bD. dr
":#
$ecause these inal orms o the momentum e*uation (alance still contain the varia(les
or element thrust and tor*ue, they cannot (e used directly to solve or inlow actors.
;owever there now e+ists a nonlinear system o e*uations "1#,"2#,",#,"4#,"-# and ":#
containin! the our primary un0nown varia(les A T* A +, a, b. /o an iterative
solution to this system is possi(le.
$terative Solution pro"edure for Blade Element Theory#
&he method o solution or the (lade element low will (e to start with some initial !uess
o inlow actors "a# and "b#. <se these to ind the low an!le on the (lade "e*uations ",#,
"4##, then use (lade section properties to estimate the element thrust and tor*ue
"e*uations "1#,"2##. =ith these appro+imate values o thrust and tor*ue e*uations "-# and
":# can (e used to !ive improved estimates o the inlow actors "a# and "b#. &his process
can (e repeated until values or "a# and "b# have conver!ed to within a speciied
tolerance.
't should (e noted that conver!ence or this nonlinear system o e*uations is not
!uaranteed. 't is usually a simple matter o applyin! some conver!ence enhancin!
techni*ues "ie >ran0-5icholson under-rela+ation# to !et a result when linear aerooil
section properties are used. =hen non-linear properties are used, ie includin! stall
eects, then o(tainin! conver!ence will (e si!niicantly more diicult.
2or the inal values o inlow actor "a# and "b# an accurate prediction o element thrust
and tor*ue will (e o(tained rom e*uations "1# and "2#.
Propeller Thrust and Tor,ue Coeffi"ients and Effi"ien"y#
&he overall propeller thrust and tor*ue will (e o(tained (y summin! the results o all the
radial (lade element values.
T=2AT "or all elements#, and Q=2AQ "or all elements#
&he non-dimensional thrust and tor*ue coeicients can then (e calculated alon! with the
advance ratio at which they have (een calculated.
C
T
=
T
jn
2
D
4
and
C
Q
=
Q
jn
2
D
5
or J=
V

nD
where n is the rotation speed o propeller in revs per second and ! is the propeller
diameter.
&he eiciency o the propeller under these li!ht conditions will then (e
j( propeller)=
J
2n
C
T
C
Q
.
Soft%are $mplementation of Blade Element Theory
&wo pro!rammin! versions o this propeller analysis techni*ue are availa(le. &he irst is
a demonstration pro!ram which can (e used to calculate thrust and tor*ue coeicients
and eiciency or a relatively simple propeller desi!n usin! standard linearised aerooil
section data. &he (lade is assumed to have a constant pitch "p# so that the variation o
0 with radius is calculated rom the standard pitch e*uation.
p=2nr tan(0) .
>lic0 here to download ?ropeller ?ro!ram 1: ?ropel.e+e
&he second is a @A&4A$ script ile or the implementation o this method. &he source
code in this script is (y deault a simple propeller desi!n with linear properties. ;owever
with the inclusion o your own propeller !eometry and section data a more accurate
analysis o the speciic propeller desi!n can (e o(tained.
>lic0 here to download ?ropeller ?ro!ram 2: ?ropel.m
Return to Table of Contents.
All content and images Auld, 1998-2012