You are on page 1of 4

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

DEPARTMENT OF MECHANICAL ENGINEERING

CAMBRIDGE, MASSACHUSETTS 02139

2.29 NUMERICAL FLUID MECHANICS FALL 2011

QUIZ 1

Monday, October 24, 2011

The goals of quiz 1 are to: (i) ask some general higher-level questions to ensure you understand broad concepts and
are able to discuss such concepts and issues with others familiar with numerical fluid mechanics; (ii) show that you
can apply methods and schemes that you learned in idealized computational problems; and (iii), evaluate if you can
read numerical codes and recognize what they accomplish. Partial credit will be given to partial answers (e.g.
accurate descriptions in words of computations required but not carried out).

I. Short Concept Questions


Problem I.1 (24 points)
Please provide a brief answer to the following questions (a few words to a few sentences is
enough, with or without a few equations depending on the question).
a) If you utilize the bisection method (to find the root of an equation), can you predict the

maximum number of iterations required for a specific error level ahead in time (i.e. before

the bi-section method is run)? The bisection method converges slowly, so when would one

use it?

b) What types of computations are very sensitive to round-off errors? Provide an example (in

words is fine).

c) Write the differential equation for conservation of mass of a compressible fluid, both in a

conservative and non-conservative form. State one reason why using the conservative form

in computations would be advantageous.

d) For what type of matrices is the Conjugate Gradient method suitable?

e) Provide three properties of hyperbolic equations.

f) You develop a consistent finite-difference algorithm to solve a PDE. You implement the

algorithm and run a numerical test of convergence to verify your code. You find out that

your solution does not converge as expected. Provide 3 reasons why this could happen.

II. Idealized Computational Problems


Problem II.1: Error Estimations (18 points)
The coefficient of lift for a thick cambered and
finite wing of approximately elliptical planform is given by:
2 sin( 2h / c )
CL
1 2 / AR
where is the angle of attack, c is the chord,
h/c is the maximum camber expressed as a
fraction of the chord (h is the camber height), VRXUFHXQNQRZQ$OOULJKWVUHVHUYHG7KLVFRQWHQWLVH[FOXGHGIURPRXU
&UHDWLYH&RPPRQVOLFHQVH)RUPRUHLQIRUPDWLRQ
AR is the finite aspect ratio defined by
2
VHHKWWSRFZPLWHGXIDLUXVH
AR=b /Ap where b is the span and Ap=bc the

so-called planform area. Parameters are illustrated on the figure, except the camber.

- 1 -

a) Use a first-order error analysis based on the standard error formula to determine the
sensitivity of the coefficient of lift to the camber height h and to the chord c.
b) If = 5o, h = 2 m, c = 10 m and b = 30 m, to which of the two parameters (h or c) is the
coefficient of lift the most sensitive to?
c) Due to in-flight forces and vibrations, if your uncertainty (error standard deviation) on your
estimate of h is 10 percent of 2 m and on that of c is 1 percent of 10 m, which of the two
parameters should your onboard flight controller attempt to measure with more precision?
Problem II.2: Conditioning of symmetric positive definite matrices. (18 points)
In scientific and engineering computations, one of the most important classes of matrices is the
symmetric positive definite one. Recall that a matrix A is positive definite if xT Ax 0 x 0 .
Considering such a symmetric positive definite matrix A:
a) Obtain the L2 norm or spectral radius of A, i.e. A 2
b) Using a), express the condition number of A as a function of eigenvalues of A.
c) Based on your result in b), is a positive definite matrix always well-conditioned? Briefly
discuss.
d) What is the minimum condition number of A? To what kind of matrices could this
correspond to? Briefly state why such matrices are a favorite in numerical algebra.
Problem II.3: Finite Differences and Partial Differential Equations (22 points)
Consider the following Partial Differential Equation (PDE):
T 2T
f (t )

t x 2
a) What type of PDE is this equation? Briefly state the types of problems that are normally
solved with this type of PDE.
b) Develop an explicit finite difference scheme to solve this PDE, using centered in space and
forward in time finite differences.
c) If you were using an implicit scheme instead, how would your result from b) change?
Would your method to solve the resulting linear system also change?
d) Given t =0.1 s, x =0.5 m, =1 m2/s, f (k t) 1 K/s, and initial conditions Ti01 = 5K,
0
Ti 0 = 6 K, and Ti1
= 32 K, calculate the value Ti1 (after one iteration only) using the explicit

scheme from part b). Given that the true value of Ti1 is 18K, what is the error?
e) A discretization scheme developed by someone else gave Ti1 = 26 K for the same t = 0.1s
used in f). Then, using t =0.05 s, this discretization scheme gave Ti1 =17 K. What is the
order of convergence of this scheme? Is this scheme better or worse than the one you
developed? Which one gave the smaller error?
III. Numerical code Evaluation Problems
Problem III.1: Linear Systems and matrix decomposition using Matlab (18 points)
Read the following Matlab function for an algorithm applicable to some special matrices:

- 2 -

function R = algorithm(A)

m = size(A,1);

R = A;

for k = 1 : m

for j = k + 1 : m
R(j,j:m) = R(j,j:m) - (R(k,j:m) * R(k,j) / R(k,k));
end
R(k,k:m) = R(k,k:m) / sqrt(R(k,k));
end
R = triu(R);
% triu(R) gives upper triangular part of matrix R

a) What is the name of this algorithm and what does it do in a few words?

b) Give an asymptotic operation count of the algorithm i.e. obtain the values of a and b where

Operation count ~ a. (m)b


Hint: Identify the step in the Matlab code which will require the most number of operations.
Obtain the operation count corresponding to this step only (do not aim to compute the exact
number of operations).
c) This algorithm is applicable only to some special matrices. What are the requirements of
these special matrices?

- 3 -

MIT OpenCourseWare
http://ocw.mit.edu

2.29 Numerical Fluid Mechanics


Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

You might also like