Вы находитесь на странице: 1из 4

Suggested Problems Set #8

ECH 6847, Fall 2016


Completion Date: Wednesday December 14th
The final exam will take place, in our usual classroom, on Wednesday December 14th at
10:00 am. This set of problems represents material, in addition to all previous material, for
which you are responsible for the exam. Example exams for your practice will be posted and
there will be no additional quizzes.
Problem #5 (from problem set #7): Least-squares example
Assume that you have measured a rate constant, k, at several different temperatures T , and
wish to find the activation energy (divided by the gas constant), E/R, and the preexponential, k0 , in the Arrhenius model,
k = k0 exp(E/RT ).
The data are given here:
T(K)
k
a)

300 400 500


1.82 2.12 2.26

Take logarithms of the equation and write a model that is linear in the parameters
ln(k0 ) and E/R. Summarize the data and model with the linear algebra problem
A x = b,
in which x contains the parameters of the least squares problem,


ln(k0 )

x =
E/R

b)
c)

What are A and b for this problem?


Find the least-squares fit to the data. What are your least-squares estimates of ln(k0 )
and E/R?
Is you answer unique? How do you know?

Problem #1: Accuracy and Stability of Euler Methods


Consider equations of the form
dx
= f (x, t) .
dt
These equations can be integrated numerically using either an explicit (forward),
x(t + t) = x(t) + tf (x, t) ,
or implicit (backward),
x(t + t) = x(t) + tf (x, t + t) ,

Euler scheme. Both methods are first order accurate, as both can be developed from firstorder Taylor series approximations. The explicit method estimates the solution for x(t + t)
using known information at the current time t. The implicit method uses the value of f at the
new point, where the value is not known, requiring the solution to a potentially complicated
algebraic equation at each step.
So why might one use the implict method instead of the simpler explicit method? The
implict Euler method is unconditionally stable, whereas the explict method is only conditionally stable. To demonstrate, compute x at t = 14.7 using t = 2.1 using both methods
and compare to the exact solution when f (x, t) = x; plot your results as a function of
time. You will see that the explicit method generates a solution that alternates about the
real solution and that the implicit method does not. The first is unstable, the second is
stable.
A criteria for stability is that |G| < 1 for these methods, where
x(t + t) = Gx(t).
Identify the range of time steps at which the explicit method is stable and unstable. Also
demonstrate that the implicit method is stable, regardless of the time step.
Problem #2: Temperature measurement by analyis of an extended surface
A long rod within an oven having an air temperature of T = 400oC is pressed firmly onto
the surface of a billet. Thermocouples imbedded in the rod at locations 25 and 100 mm from
the billet register temperatures of 325 and 375oC, respectively. What is the temperature of
the billet?

Note that the temperature in the rod is governed by the differential equation,
d2 T (z)
4h

(T (z) T ) = 0
2
dz
kD
where z is the coordinate along the rod, z = 0 at the billet, and the temperature approaches
4h
, can be determined from the provided
400o C as z goes to infinity. The coefficient, kD
information.

Problem #3: Temperature measurement by analyis of an extended surface


Solve the above problem again, but using a finite difference. As shown, use five nodes
separated by equal distances of 25 mm. You should be able to generate three equations from
which you can calculate the temperature of node 1 located at 0 mm, i.e. the temperature of
the billet. Compare with the exact answer from above.

Problem #4: Reaction-diffusion in cylindrical geometry


Solve the reaction-diffusion problem in cylindrical coordinates, with a second-order reaction, using a finite difference method. The governing equation is given by


1 d
dc(r)
r
c(r)2 = 0,
r dr
dr
where c(r = 1) = 1. Generate the solution for four nodes, where the first node is located
at the origin (r = 0), the fourth node is at r = 0.75, and the spacing is, consequently,
r = 0.25.

Problem #5: Diffusion in two dimensions


Solve the two-dimensional Laplace equation,
2T
2T
+
= 0,
x2
y 2
with the boundary conditions as shown. Use a central difference approximation for the second
dervitive in each direction to give a set of four algebraic equations for the temperatures at
the node as indicated. Solve and compare to the analytical solution derived earlier in the
semester.

Вам также может понравиться