Solving a system of PDE using pdepe

7 ビュー (過去 30 日間)
nir livne
nir livne 2022 年 6 月 30 日
コメント済み: Bill Greene 2022 年 7 月 1 日
Hi,
I'm trying to solve system of 2 PDE's. It is a one-dimensional problem (cylinderical coordinates with symmetry):
with the following boundary conditions:
,
(R stands for r=R which is the boundary of the domain).
I'm keep getting index errors but I can't figure out why, I've been stuck for a while now... Specifically, the 'pdefun' is keeping me stuck right now with the following error:
Index exceeds the number of array elements. Index must not exceed 1.
Error in AggSim>pdefun (line 35)
f = [Dn, alpha*(u(1)/u(2)); 0, Dc]* dudx;
Any help would be appreciated! Thanks in advnace :)
%% constants and space/time variables
L = 0.5;
dL = 0.001;
x = 0:dL:L;
t_steps = 100;
t_f = 1;
t = linspace(0, t_f, t_steps);
m = 1;
alpha = 10^-3;
Dn = 4 * 10^-6;
Dc = 9 * 10^-6;
k = 10^-10;
pH0 = 5.5;
beta = 0.1 * 10^-(pH0);
%% solve and plot
sol = pdepe(m,@pdefun,@icfun,@bcfun,x,t);
u1 = sol(:,:,1);
u2 = sol(:,:,2);
surf(x,t,u1)
title('u_1(x,t)')
xlabel('Distance x')
ylabel('Time t')
%% function defs
function u0 = icfun(x)
global pH0
u0 = [1, 10^-pH0];
end
function [c,f,s] = pdefun(x, t, u,dudx)
global alpha Dn Dc k beta
c = [1; 1];
f = [Dn, alpha*(u(1)/u(2)); 0, Dc]* dudx;
s = [0; beta -k*u(1)*u(2)];
end
function [pL,qL,pR,qR] = bcfun(xL,uL,xR,uR,t)
global pH0
pL = [1, 1];
qL = [0; 0];
pR = [1; 0];
qR = [0; uL(2)-10^-pH0];
end
  1 件のコメント
Bill Greene
Bill Greene 2022 年 7 月 1 日
In your equations for the boundary conditions you show some derivatives with respect to t. Are those really supposed to be derivatives with respect to r?

サインインしてコメントする。

採用された回答

Torsten
Torsten 2022 年 6 月 30 日
編集済み: Torsten 2022 年 6 月 30 日
The error is solved, but I think the boundary conditions at the right end can't be set within pdepe.
The condition set at the moment (by me) is not what you want.
I assumed beta = c0 in your code.
%% constants and space/time variables
global alpha Dn Dc k beta
global pH0
L = 0.5;
dL = 0.001;
x = 0:dL:L;
t_steps = 100;
t_f = 10000;
t = linspace(0, t_f, t_steps);
m = 1;
alpha = 10^-3;
Dn = 4 * 10^-6;
Dc = 9 * 10^-6;
k = 10^-10;
pH0 = 5.5;
beta = 0.1 * 10^-(pH0);
%% solve and plot
sol = pdepe(m,@pdefun,@icfun,@bcfun,x,t);
u1 = sol(:,:,1);
u2 = sol(:,:,2);
surf(x,t,u1)
title('u_1(x,t)')
xlabel('Distance x')
ylabel('Time t')
%% function defs
function u0 = icfun(x)
global pH0
u0 = [1; 10^-pH0];
end
function [c,f,s] = pdefun(x, t, u,dudx)
global alpha Dn Dc k beta
c = [1; 1];
f = [Dn, alpha*(u(1)/u(2)); 0, Dc]*dudx;
s = [0; -k*u(1)*(u(2)-beta)];
end
function [pL,qL,pR,qR] = bcfun(xL,uL,xR,uR,t)
global pH0
pL = [0; 0];
qL = [1; 1];
pR = [0; uR(2)-10^(-pH0)];
qR = [1; 0] ;
end
  5 件のコメント
Torsten
Torsten 2022 年 7 月 1 日
  1. You forgot to include the globals in the script part of your code.
  2. s in pdefun has changed. I assumed beta = c0 and thus set s(2) = -k*u(1)*(u(2)-beta).
  3. The boundary condition setting (pL qL, pR, qR) is substantially different from your settings. At the moment, the setting at r=R is c = 10^(-pH0) and D_n*dn/dr - alpha*n/c * dc/dr = 0. I guess that with your definition of f in pdefun, it is impossible to set dn/dr = 0 at r=R in pdepe.
nir livne
nir livne 2022 年 7 月 1 日
Actually, your setting of D_n*dn/dr - alpha * n/c * dc/dr is what I meant to set. Sorry for the confusion I've created! I meant the flux of n to vanish at r=R.
You've been exteremly helpful! Thank you!

サインインしてコメントする。

その他の回答 (0 件)

カテゴリ

Help Center および File ExchangeEigenvalue Problems についてさらに検索

製品


リリース

R2021b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by