Question 6 of 6: Drag on a plate standing in a turbulent boundary layer
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Notes on this paper
Paper format. National Exams, May 2015 —
07-Mec-B6 Advanced Fluid Mechanics. Three hours, OPEN BOOK, any
approved Sharp or Casio calculator permitted. Six questions are printed; any five of them
constitute a complete paper and each carries an equal 20 marks, with the item weights shown
in the left margin. No aid sheet is bound into the paper — the open-book rule is the candidate’s table source, so the compressible-flow ratios below are quoted in closed form rather than read from a chart. All six questions are solved here.
Reference texts. Solutions follow the conventions of the texts the
EGBC syllabus recommends for this subject:
F. M. White, Fluid Mechanics, 8th ed. — potential-flow building blocks
(§4.4, §8.2–8.3), turbulent flat-plate layers (§7.4), dimensional
analysis (§5.2–5.4), duct flow with friction (§9.7).
F. M. White, Viscous Fluid Flow, 3rd ed. — exact solutions of the
Navier–Stokes equations and lubrication theory (§3.2, §3.9).
J. D. Anderson, Modern Compressible Flow, 3rd ed. — quasi-one-dimensional
nozzle flow, normal shocks and Fanno flow (Ch. 3 and Ch. 5).
P. K. Kundu, I. M. Cohen & D. R. Dowling, Fluid Mechanics, 6th ed. —
complex potential and plane potential flows (Ch. 6); boundary layers (Ch. 9).
B. R. Munson et al., Fundamentals of Fluid Mechanics, 8th ed. —
Buckingham Pi method and model similarity (Ch. 7).
SI units throughout. Air and mercury properties are those printed in the question; all
pressures are absolute unless a gauge value is stated explicitly, which is standard Canadian
practice for this examination.
Question 6: Drag on a plate standing in a turbulent boundary layer (20 marks)
Given. A thin plate of streamwise length $L$ and height $b$, wetted on
both faces, standing normal to a wall and aligned with the flow. The approaching wall
boundary layer is fully turbulent with thickness $\delta$ and profile
$u(y)=U_\infty (y/\delta)^{1/7}$, and the plate is taken to lie wholly inside it, $b\le\delta$.
Fluid properties $\rho$ and $\nu$ are uniform.
Find. (a) a closed-form drag coefficient for the plate, and (b) the ratio
of that drag to the drag the same plate would feel in a uniform stream $U_\infty$.
Figure 6.1 — The
plate seen edge-on. Each horizontal strip of height $\mathrm{d}y$ meets the flow at the local
speed $u(y)$ and is treated as an independent flat plate of length $L$ wetted on two
faces.
Approach. Treat the plate as a stack of independent strips (the strip, or
blade-element, approximation), assign each strip the standard turbulent flat-plate friction
coefficient evaluated at its own local Reynolds number, integrate over the height, and
normalise.
Part (a) — state the flat-plate friction law used for each strip.
For a fully turbulent plate the local skin-friction coefficient follows the one-seventh-power
correlation $c_{fx}=0.0266\,Re_x^{-1/7}$, whose integral over a plate of length $L$ gives the
plate-average value
$$\begin{aligned}
C_F&=\frac{1}{L}\int_0^{L}c_{fx}\,\mathrm{d}x=\frac{7}{6}(0.0266)\,Re_L^{-1/7}=0.031\,Re_L^{-1/7}\cr
Re_L&=\frac{U L}{\nu}
\end{aligned}$$
This is the standard result quoted in White and is valid for
$5\times10^{5}<Re_L<10^{7}$.
Write the force on one strip. A strip at height $y$, of height
$\mathrm{d}y$, sees the local speed $u(y)$ and is wetted on both faces, so its drag is
$$\mathrm{d}F=2\,C_F\!\left(\frac{u(y)L}{\nu}\right)\tfrac{1}{2}\rho\,u(y)^{2}\,L\,\mathrm{d}y
=0.031\,\rho L\left(\frac{L}{\nu}\right)^{-1/7}u(y)^{2-1/7}\,\mathrm{d}y .$$
The exponent collects to $u^{13/7}$; the strips are treated as independent because the
spanwise pressure gradient along the plate is negligible for a thin aligned plate.
Insert the power-law profile and integrate over the height. With
$u=U_\infty(y/\delta)^{1/7}$ the integrand becomes $y^{13/49}$, and
$$\int_0^{b}u^{13/7}\mathrm{d}y
=U_\infty^{13/7}\,\delta^{-13/49}\!\int_0^{b}\! y^{13/49}\mathrm{d}y
=\frac{49}{62}\,U_\infty^{13/7}\,b\left(\frac{b}{\delta}\right)^{13/49},$$
since $\tfrac{1}{1+13/49}=\tfrac{49}{62}$. Hence the total drag is
$$F=0.031\,\rho\,L^{6/7}\nu^{1/7}U_\infty^{13/7}\left(\frac{49}{62}\right)
b\left(\frac{b}{\delta}\right)^{13/49}.$$
Normalise to a drag coefficient. Referring the force to the total wetted
area $2bL$ and the free-stream speed $U_\infty$,
$$C_D\equiv\frac{F}{\tfrac{1}{2}\rho U_\infty^{2}(2bL)}
=0.031\left(\frac{49}{62}\right)Re_L^{-1/7}\left(\frac{b}{\delta}\right)^{13/49}$$
$$\boxed{\;C_D=0.0245\,Re_L^{-1/7}\left(\frac{b}{\delta}\right)^{13/49}\;}$$
in which the Reynolds number is formed on the free-stream speed,
$Re_L=U_\infty L/\nu$. If the planform area $bL$ is preferred as the reference, the numerical constant doubles to
0.049 and the functional form is unchanged.
Part (b) — form the reference drag in a uniform stream. Immersed
in a uniform stream $U_\infty$ every strip sees the same speed, so the same correlation gives
$$\begin{aligned}
F_0&=0.031\,Re_L^{-1/7}\cdot\tfrac{1}{2}\rho U_\infty^{2}(2bL)\cr
C_{D,0}&=0.031\,Re_L^{-1/7}
\end{aligned}$$
Compare the two. Dividing, everything except the profile factor cancels:
$$\boxed{\;\frac{F}{F_0}=\frac{C_D}{C_{D,0}}
=\frac{49}{62}\left(\frac{b}{\delta}\right)^{13/49}\;}$$
The ratio is always less than unity, so the plate inside the boundary layer always
feels less drag than the same plate in a uniform stream, simply because most of it is
bathed in fluid moving slower than $U_\infty$. The reduction is modest, because the
one-seventh profile is very full: even a plate reaching exactly to the edge of the layer,
$b=\delta$, feels only $49/62=0.790$ of the uniform-stream drag, a 21 per cent saving. A
shorter plate saves more — at $b=0.75\delta$ the ratio is 0.732, and at $b=0.10\delta$
it is 0.427.
Check the algebra on a concrete case. For air at
$U_\infty=30$ m/s, $\nu=1.5\times10^{-5}\ \text{m}^{2}/\text{s}$,
$\rho=1.2\ \text{kg/m}^{3}$, $L=2$ m, $\delta=80$ mm and $b=60$ mm, the closed form gives
$F=0.335$ N against $F_0=0.458$ N, a ratio of 0.732 exactly as predicted, and a
strip-by-strip numerical integration reproduces $F$ to five figures.
Check: the derivation assumes $b\le\delta$, so that
every strip lies inside the power-law profile. If the plate protrudes above the layer the
integral must be split, adding a uniform-stream contribution
$0.031\,Re_L^{-1/7}\tfrac12\rho U_\infty^{2}\,2L(b-\delta)$ for the exposed portion. It also
assumes that each strip develops its own turbulent layer from its leading edge and that the
strips do not interact — the standard strip approximation, accurate for a thin aligned
plate but not for one at incidence.