

Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Material Type: Assignment; Class: Cmp Mod Fluid Dynm; Subject: Meteorology/Phys Oceanography (MPO) ; University: University of Miami; Term: Unknown 1989;
Typology: Assignments
1 / 2
This page cannot be seen from the preview
Don't miss anything!


MPO 662 { Problem Set 3
unj +1
unj +1 +unj 1 2 t
unj +1 unj 1 2x
Determine the truncation error for this scheme Study the consistency of this FD representation What condition to you have to imp ose on x so that the time errors and spatial errors are of the same order of magnitude? Determine the stability characteristics of this scheme.
unj +1 unj t
unj +1+1 2 unj +1+ unj +1 1 x^2
unj +1 2 unj + unj 1 x^2
Study the stability and convergence of this metho d for p erio dic b oundary conditions. Hint: expand in Taylor series ab out (j x; (n + 12 )t), do the space-expansion rst, and use it in analyzing the time-series.
ht + (hu)x + (hv )y = 0 (2)
u = 2
y
; v = 2
x
h(x; y ; t = 0) = max(0; 1 8 r ); r 2 =
x
y
in the unit square 0 x; y 1. Use a conservative scheme discretized on a C-grid and the following uxes:
a Donor cell ux a 2nd^ order centered ux
Cho ose a time-stepping scheme that is at least as accurate as the space discretization (you can reuse some of the programs you wrote earlier). Contour the numerical solution versus the analytical solution at time t = 1 for a max. Courant numb ers of 0 : 25 and compute the ro ot mean square error and the maximum error in the solution. Monitor the minimum and maximum values of the solution throughout the calculation, as well as the total budget of h; check to see if h is indeed conserved. Use 4 sets of grids using 16 16, 32 32, 64 64, and 128 128 cells and discuss the numerical results. Draw
also the convergence curves on a log-log plot. Note: the exact solution is simply the initial condition as the cone returns to its starting p osition after a full rotation. The are 8 cases to run, and hence you should think how to structure the program so you can do the numerical exp eriments without to o many changes to your co de (and p ost- pro cessing co de). Designing and planning the di erent steps would save you a lot of work later on. It mayb e helpful while debugging to use a unidirectional constant ow to check your solution. Pay particular attention to your co ding near the edge of the domain. Numb ering of the C-grid can b e cumb ersome. It the tracers are dimensioned as Tr(1:m,1:n) then the velo cities have to b e u(1:m+1,n) and v(1:m,1:n+1). The divergence of the ow eld at p oint tracer p oint (i,j) is given by
do j = 1,n do i = 1,m div(i,j) = (u(i+1,j)-u(i,j))/dx + (v(i,j+1)-v(i,j))/dy enddo enddo