'^//Vc/\'r<^>31^ NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS TECHNICAL MEMORANDUM 1298 ON THE PROBLEM OF GAS FLOW OVER AN INFINITE CASCADE USING CHAPLYGIN'S APPROXIMATION By G. A. Bugaenko Translation 'K Voprosu o Struinom Obtekanii Beskonechnoi Reshetki Gazom v Priblizhennoi Postanovke S. A. Chaplygina." Prikladnaya Matematika i Mekhanika, T. XIIL No. 4, 1949. UNIVERSITY OF FLORIDA DOCUMENTS DEPARTMENT 1 20 MARSTON SCIENCE LIBRAf Y P.O. BOX 11 7011 iAlNESVILLE, FL 32611 -701 U Washington May 1951 SA jf(f f/^^/7 3^7y^33 NATIONAL ADVISOEY COMMITTEE FOR AERONAUTICS TECHNICAL MEtvIORANDUM 1298 ON Tffli: PROBLEM OF GAS FLOW OVER AN INFINITE CASCADE USING CHAPLYGIN'S APPROXIMATION* By G. A. Bugaenko 1. Some well-laiown results of Chaplygin's method (reference l) are first presented. For the adiabatic law of the state of the gas when p = kP , the following relations hold; PO 1 - 2^ J P = PqU - 2\P Y_ 2a,J (1.1) where p is the gas pressure, p is the gas density, V is the modulus of velocity, Pq and Pq are the values of p and p at the critical point of the flow- at which the velocity becomes zero, k is the coefficient of proportionality, y is the ratio of specific heats, and a and P are constants. If the angle 9 between the velocity and the x-axis and the magnitude T equal to V^/2a are considered, then, as was shown by Chaplygin, the equations of gas motion assume the form ar 2T ae b^ (i-T)P 3'-? Si = - 2Th-T)P+l 06 a^ i-(2i3+i)T aij/ where cp is the velocity potential and ^ is the stream function. (1.2) *"K Voprosu Struinom Obtekanii Beskonechnoi Reshetki Gazom v Priblizhennoi Postanovke S. A. Chaplygina. " Prikladnaya Matematika i Mekhanika, T. XIII, No. 4, 1949, pp. 449 - 456. KA.CA TM 1298 Chaplygin reduced these equations to the very simple form So ^ Se b^ s* Sep 1 £0 K Sy (1.3) ty introducing the new variable and the constant K defined by the formulas _ J 2T )P dT To.2 > K = 1-(2P+1)T (1-T)23+1 (1.4) Chaplygin showed that for velocities far removed from the velocity of sound, the magnitude K is approximately unity and equations (1.3) can be integrated by assuming K equal to 1. For K = 1, these equations go over into the conditions of Cauchy-Elemann; hence, oj = + ia will be an analytical function of the complex variable f = cp + i^. The equation for the elementary vector along a streamline in the approximate treatment has, as is known (reference 3), the form dz (ae^'^ + bei'^)dcp 1 + (1-^002^ AT 2atj 1 - (l-'^2)" 2 ATi^g The complex pressure is given by the equation (reference 3) (1.5) iX — 2~J (® " ® >^^ + Po(^ -^o^2' J dz (1.6) 2. The steady potential flow of a gas through an infinite cascade according to the well-known scheme of Kirchho:ff with separations of the jet is next considered. The vanes of the cas- cade will bs assumed to be plane (fig. l). Velocity of the gas in the flow at infinity is denoted by Vo--|_ and its angle with the x-axis by ^3_; velocity of the gas in the jet at infinity is denoted by '\^2 ^^^ ^"^^ angle with the MCA TM 1298 3 X-axis by %,2* '^^^ angle 6 between the velocity vectors and the X-axis lies within the range -2jt<0^O. The velocity field of the flow repeats itself for each dis- placement by the pitch of the cascade, that is, by the vector he"-^ The condition of constancy of the mass flow for steady flow of the gas gives Q = - Pcoi Vo.1 h sin (\+ Qooi) = i::to2 ^2 ^ (2.1) where p^^-^ i^ the density of the gas at infinity in the flow, p^^z is the density of the gas at infinity in the Jet, and n is the width of the gas jet at infinity. By .-naking use cf expressions (l.l) for p, eq^uation (2.1) can be represented in the form - V^-L h [l - -|i-J sin (\+ ^i) = V^2 ^ V^ - -|i-y (2.2) The behavior of the function f = cp + i"i|/' In the z-plane of the gas flow is now considered. For simplicity, the function f is assumed equal to zero at the critical point of the flow (fig. l). From the relation Ss dcp = -^ ds = Vg ds it follows that on moving along the streamline ij/ = 0, the function cp varies monotonically from -«» at the point E (infinity In the flow) to +<» at the point C (infinity in the jet) and passes through the zero value at the critical point where the stream- line branches. The value of the potential f = 9 + iv at the critical point 0' displaced by the period he"'^'^ relative to the point is found and (fig. l) CP(O') = J Vg ds = J Vg ds + J Vg ds + J Vg ds OMM'O' OM MM' M'O' where MM' is a cut parallel to the axis of the cascade. KACA TM 1298 The first and last integrals mutually cancel and therefore as MM' approaches infinity in the flow, the following equation is obtained : ;p (0') = 7„i h cos (\ + e^^) (2.3) Furthermore, from the relation dQ d\j/ = — where dQ is the quantity of gas flowing in unit time between infinitely near streamlines, it follows that on being displaced by the pitch of the cascade the function ^ receives an increment equal to Q/pq* Hence, nO') .^ = - (l - ^^J hV^i sin (\. e^i) (2.4) In this manner, the f-plane with double-sided cuts along t?ie half-straight lines parallel to the axis of reals (fig. 2) corresponds to the region of the gas flow (fig. l). All the cuts, because of the rule by which the cascade was constructed, are obtained from the initial one (the positive cp-axis) by simultaneous displacement along verticals and horizontals at distemces that are multiples of Q/Pq and V -^ h cos (\ + 9 -^) , respectively. • Because 'm - e + ±0 is an analytic function of f = cp + 1^ , the problem ms.y be solved by relating these functions with t?ie aid of a parameter that varies in the upper semicircle of unit radius, as in the Levi-Civita method. By considering the recti linearity of the cuts in the f-plane, the analytic function f(t) is found, which brings about the con- forme,! transformation of the f -plane into the semicircle t. In the t-plane, the flow of an ideal fluid about the boundary of the semi- circle is constructed. For this purpose, sources and vortices of strengths and intensities are located, as shown in figure 3, at the points tec, and t^~~ that are symmetrical with respect to the circle and at the mirror reflection of these points in the diarseter, that is, at the points toe and too" • At the origin of coordinates we place a sink of strength 2q (and a similar sink: at infinity). KACA TM 1298 5 In t.?ie constructed flow in the t-plane, the upper semicircle of unit radius and the diameter of the semicircle are, of course, stream- lines so that the stream function f maintains a constant value at the boundary of the upper semicircle t; in the f -plane, this boundary vrill correspond to straight cuts. The complex potential of the constructed flow will have the form f (t) = ^y Uy + iq.) log (t - ttt) + (-Y + iq) log (t - ^J+ (-■> + iq) log (t - t^) + CY + iq) log ( t - i^ j - 2ig log t + constant or f (t) = ^ [(7 + iq) log ft + i - 2MJ - (■> - iq) log f t + i - 2MJj + constant (2.5) where 7 is the intensity of the vortex and q is the strength of the sources, and The arbitrary constant in equation (2.5) is chosen so that f(t) becomes zero at a certain point t = e^-^, the position of which will be subsequently determined. Because the logarithm has multiple values, the upper semicircle of the t-plane will correspond to an f-plane with an infinite number of straight cuts '^ = kq(k = 0, ±1, ±2,...), where cp changes from k7 to +CC, as easily follows from equation (2.5) by substituting t = ei® (the arc of the semicircle) and t = t^^ where t]_ is the real amount of the interval (-1, +l), the diameter of the semicircle. In this manner, the function (2.5) establishes a conformal mapping of the upper semicircle of the t-plane on the f-plane with the double-sided cuts represented in figure 2. The function (2.5) is used in the work of N. I. Akhiezer (reference 2) where it is obtained by successive conformal mappings: the f-plane on the half plane, the half plane on the unit circle, and finally the circle on the upper semicircle. MCA TM 1298 The point t = is carried by the transformation (2.5) Into f = +«» so that the radii AC and BC go over into the infinite segments (figs. 2 and 3). The conformal property of the transfor- mation breaks down at the points t = -1, t = +1, and t = e-'- ^ (the point e^^ corresponds to the origin of the double-sided cut in the f-olane). The condition df/dt = for t = ei^ gives -^^-^-^ = -2L_:_ia (2.6) cos £-M cos £-M In order to obtain the elements of the motion, an expression for the derivative df/dt is req.uired that is represented in the form df q (t-e^g)(t-e-i^)(t-t-^) dt n (t-t^) (t-i^-^) (t-t^) (t-t^-1) The quantities y and q. are next determined. When a point in the z-plane of the gas flow is displaced by the pitch of the cascade he"-^'^, the point t, corresponding to the point in the z-plane, goes over from one sheet of the Eiemann surface to the next, passing once around the point E (t =1^1,), as a result of which the function f(t) receives an increment y + iq, as follows from equation (2.5). Because the corresponding increments of the functions cp and 'ii are equal to Vcc^^ h cos (\ + Qk^) and Q/Pq^ respectively, the following equations are obtained: 7 = "^1 h cos (\ + da^i) (. q = - I 1 - -^—j V„i h sin (\ + 0001 ) (2.8) From the expression for 7, it is evident, among other things, that 7=0 corresponds to the case where the approaching flow has a velocity at infinity perpendicular to the axis of the cascade. 3. The function ^{t) Is next determined. The function uj = 9 + io is regular within the semicircle t and has the following properties: 1. At the point 0(t = e^^), the real part of the function ^ has a discontinuity, equal to n, because of the branching of the streamline. On the arc AO the angle 6 is equal to -n and on the arc OB it is equal to zero. MCA TM 1298 2. On the real diameter of the semicircle, the function oo(t) = + icJ is real "because its imaginary part a = on the free jets where <»2' 3. At the origin of coordinates t = 0, the f-onction ^{t) is equal to ^g 'because at infinity in the jet 0=0 and e = e, 'o=2' From the preceding discussion, it follows that the function uj(t) admits of analytical continuation in the lower semicircle and may be obtained by the Schwarz formula. Thus, r lt| J e^^ + t «i* - t ^^ = ^ >iP >i
and q from equations (2.8) are used. Thus
ctg(X . e^J = i fi - "^Y ^ ^°^ ^ - (^ ^ ^) (3.10)
In order to compute the length of a cascade blade, equation (1.5)
is used. Replacing dtp by df gives
q ( 1 - te^^ ^ t - e^^ \ (t - e^^)(t - e-^^)(t - t-^^)dt
dz = J a — 7- + b — : ^ — -^-^ — t-
" V e'^^- t te^^ - 1/ (t - t,„)(t - €J(t - tcr~^)(t - V-1)
NACA TM 1298 11
This expression may be put in the form
dz = aa g (t)dt + 21 gp(t)dt (3.11)
n -^ n
where
g (t) = eiqt-e-l£)2(t-t-l)
, (t-to,)(t-too-l)(t-t^)(t-t„-^)
ggCt) = e-iMt-e^^)^t-t-^)
(t-tj(t-t -l)(t-tj(t-t -1)
The expansions of g]_(t) and §2^^^ into the sum of simple
fractions are of the form
gy(t) = + + -r + — 7-3^ + — ^^ - 1^2J
"t — "toe ^"""^OD ^~"^03 '^■*"'^0D
where
H = ^2 = ^ exp(i^i -O^i)
B-, = Dp = ii:!! exp(iea>i + '^*l)
^ ^ 2iq -^ ^
E-|_ = -e'
-ie
^1 = ^2 = ^^ exp(ao.i - i^s
a -H
<
U
(D CjD
cd
o
^ t>5
EjO
CO
(D
rQ r-t
3
CD
^
O P^
PP
CM
tn
•
,H
H
tn
o
<
g
(D
^ OD
•
H
t3
-p a
O
<:
o3
•H
o
b
o
c! ra
>»
g
cd
TO
td
o
O [3
w
s
C\]
(U
^-^
■T3
.
cd
rH
ver an Infinite Case
ion.
•Ci
•H
CQ
(D
i