M+ = y+
von Karman [7] < 30 M+ = 5 In y+ - 3.05
U+ =
A Single Formula for the Reichardt [15] u+ =
2 . 5 In y +
2 . 5 In ( 1 +
+ 5.5
O A y +)
D. B. SPALDING 1 0 Z y+ < 26 M+ = f
Jo
Summary dy*
Deissler [2|
It is shown that experimental velocity distributions m a y be 1 + nhi + y + (1 - e-"!" + " + )
well fitted, in the laminar sublayer, the transition region, and n = 0.124
the turbulent core, b y the formula: 26 y< u+ = 2 . 7 8 In i/ + + 3.8
to total shear stress agree with the measurements of Laufer [8] 2 0 t y + < 27.5 M+ = 14.54 tan 7i(0.0688i/+)
ltannie [13|
quite closely. 27.5 ? y+ = 2 . 5 In ? / + + 5.5
" See also Hofmann [5], Reichardt [14], Rotta [16], Miles [9],
Nomenclature
Elrod [3], and Frank-Kamenetsky [21].
6 These authors did not, at the dates in question, state the formulas
u = time-mean velocity of fluid in x-direction
attributed to them in the table. However, they did introduce the
u+ = u V p / r
idea of a sharp division between a laminar sublayer and a fully turbu-
x = distance along the wall in the direction of flow lent core; when compared with experimental data, this idea leads
directly to the formulas given.
= .'/ VVp//imolecular
= Mtotal//^moleculnr noted t h a t all t h e a u t h o r s m e n t i o n e d , except Reichardt [15] and
Mmoiocuiar = absolute viscosity of fluid in l a m i n a r motion van Driest [19], have found it necessary to use at least t w o ex-
/Utot»i = ratio of shear stress to gradient of time-mean p r e s s i o n s , v a l i d f o r d i f f e r e n t r a n g e s o f y+, in o r d e r t o d e s c r i b e the
p l e x in f o r m , w h e r e a s v a n D r i e s t ' s i n v o l v e s a q u a d r a t u r e requiring
T = shear stress in fluid, assumed independent of y numerical evaluation. There is n e e d for a simpler, easily evalu-
ated formula.
Introduction
Such a formula would preferably fit the experimental data
P u r p o s e of note. Numerous formulas have been proposed to de- closely, contain sufficient adjustable constants to permit modifi-
scribe the universal turbulent velocity profile, called b y Coles [1] c a t i o n in t h e light of n e w e x p e r i m e n t a l data, a n d h a v e a n analyti-
the " l a w of the wall." T h e present note discloses a n e w formula cal f o r m p e r m i t t i n g easy integration of t h e various functions of the
w h i c h is v a l i d o v e r t h e w h o l e r a n g e o f d i m e n s i o n l e s s d i s t a n c e y+.3 velocity distribution which arise in, for example, the theory of
The new formula has a form which, on the one hand, permits heat transfer through a turbulent boundary layer.
analytical determination of several important boundary-layer
Looked at mathematically, our problem is to establish a
parameters, and, on the other, m a y provide the vantage point
formula which:
for a new look at the theory of the turbulent boundary layer.
The u n i v e r s a l turbulent velocity profile. P r a n d t l ' s [12] p o s t u l a t e , (ii) is t a n g e n t i a l at this point to: u + = y + ;
that the velocity in t h e n e i g h b o r h o o d of a wall should obey the (iii) is a s y m p t o t i c at large y + to:1
relation:
1 Professor of Heat Transfer, Department of Mechanical Engineer- = 2.5 In y + + 5.5 (2)
ing, Imperial College of Science and Technology, London, England.
(iv) fits the experimental points at intermediate y + values.
Manuscript received by A S M E Applied Mechanics Division,
March 8, 1961.
2 Numbers in brackets indicate References at end of Note. 4 Here the most popular constants for the logarithmic velocity pro-
near u + = 0:i/ + = u +
(3)
This satisfies r e q u i r e m e n t s (3) and (4). Does it also fit the ex-
which contains the experimental data of Laufer [8]. Evidently, Fig. 1 Experimental data of Laufer [8] for velocity distribution near the
equation (5) fits the d a t a fairly well, b u t gives values of u + which wall in turbulent pipe flow, compared with various analytical expressions
are approximately 10 p e r cent low when y + lies b e t w e e n 10 and
dy+ ^ TURB
(7)
du + •U T O T A L
O 4
Equation (5) therefore implies the e+(u+) relation:
e+ = 1 -f- 0 . 4 X 0.1108(e°-4"+ - 1)
f
(0.4m + ) 2
1 + 0.04432 + + (8)
less than 4 if there is n o such variation. Equation (8) satisfies Fig. 2 Experimental data of LauFer on turbulent-stress distribution near
neither requirement. 5 However, it is e a s y to see w h a t must be the wall in turbulent pipe flow, compared with various analytical ex-
d o n e t o t h e v e l o c i t y d i s t r i b u t i o n if e i t h e r o f t h e s e r e q u i r e m e n t s is
pressions
to be satisfied: the distribution formula becomes, respectively:
the more precise fit. Whether the (0.4M +)4 term should be in-
5 Nor, incidentally, do the expressions of Reichardt and van Driest (0.4m+) 2 (0.4m + ) 3
(10a)
which appear in Table 1. 2! 31
Comparison of these relations with the experimental curves (iv) When the density varies such that density ratio rj> i s a
shows that the former equation gives the better fit at low y+, known function of u + , is it r e a s o n a b l e to calculate the velocity
while the latter gives the better fit at high y + . However, it is profile from a suitably modified version of (11)? This would
probable that both curves can be regarded as equally satisfactory run:
w h e n e x p e r i m e n t a l s c a t t e r is t a k e n i n t o account.
dhi+
A l s o p l o t t e d in Fig. 2, as a b r o k e n s t e p l i k e c u r v e , is t h e /it„rb/ = 0:4<Kw+)- (13)
du+'
Htotal distribution which corresponds to the assumption of a sharp
b o u n d a r y between a laminar sublayer a n d a fully turbulent outer which can be evaluated by numerical quadrature without dif-
region. Clearly this gives a very poor representation of the ficulty. T h i s t h o u g h t m i g h t lead t o m o r e 6 satisfactory theories of
Further p o s s i b l e i m p r o v e m e n t s . Equation (10) fits the require- equation (13) is n o t as suitable a starting point for analysis as
ment that e + i n c r e a s e s w i t h t h e f o u r t h p o w e r o f u+, and so of y+, that, for example, of v a n D r i e s t [20], w h a t is t h e p h y s i c a l reason
reason w h y the first nonzero term of the expansion should happen I t is n o t i n t e n d e d to suggest answers to these questions here.
to be that which appears in the expansion of 0.1108e°-4uT. In They are put forward solely to provoke thought and criticism.
expansion. Discussion of such further developments will b e de- which represent adequately the experimental data for the uni-
ferred to a later publication. versal turbulent velocity profile w h e n the viscosity and density
of t h e fluid a r e uniform.
Practical u s e of the n e w f o r m u l a . Fig. 1 shows that equations (9) ment of constants in the light of new experimental data, and
or (10) can be used to represent the " l a w of t h e w a l l " within the simple enough in form to permit analytical integration in im-
shows to be necessary. Of course, the constants 0.4 and 0.1108 boundary-layer analysis may be profitably re-examined with
dy+; for this integral can be written as J,u+(dy+/du+)du+, The author expresses his gratitude to Prof. J. O. Hinze of
which can be evaluated in closed form, since dy+/du+ i s e a s i l y Delft, Holland, f o r his helpful comments on an earlier draft of
On Classical Normal Modes of a for each undamped natural frequency coto. [It is interesting to
Damped Linear System note that (7) can be satisfied even in cases when either a or b
out the use of normal co-ordinates and hence without asssuming H e n c e t h e p r i n c i p a l m o d e s h a p e s y{x) will b e t h e s a m e as without
a k n o w l e d g e of the t h e o r y associated w i t h transformations to such any damping, and the value of p in a n y mode will b e such that
co-ordinates. This procedure, in addition to being instructive,
will also lead to explicit results for the damping factor and
cp + p2
= - " t o 1 (9)
natural frequency in any principal mode, and will be seen to
1 + — P
yield some interesting implications. Finally, the method of
simultaneous internal and external damping. where u> k o i s the undamped natural frequency in the H h mode-
eg V
[h] = {H}eP' (3) wk = uke (11)
4 wA,2
where* { H } i s i n d e p e n d e n t o f t h e t u n e t, a n d p i s a c o n s t a n t . Then,
Manuscript received by ASME Applied Mechanics Division, condition that the mode shapes be entirely unaffected by a damping
March 2, 1961. load of the form f(x)i>Y/c)t is that f(x) be proportional to p(.r).