pdepe - changing I.C 'midstream'

2 views (last 30 days)
Eric
Eric on 29 Jul 2015
Edited: Hashim on 15 Apr 2021
I'm using pdepe to solve a system of two nonlinear reaction-diffusion equations, Neumann B.C. At some time, t1, I would like to have pdepe use the solution, u(x,t1), a size(t)-by-size(x)-by-2 array, as the new Initial Conditions and have pdepe continue the solution from there. I want to do this because at t1 I want to change the values of some of the parameters in the reaction function specification (in pdefun) and get to the final solution. For the life of me, I cannot figure out how to pass u(x,t1) to icfun so that pdepe will use u(x,t1) as the I.C. I tried the anonymous function approach, but pdepe didn't seem to like that...perhaps I wasn't coding it correctly.

Answers (1)

Torsten
Torsten on 30 Jul 2015
Edited: Torsten on 30 Jul 2015
global counter
counter=0;
...
solold=pdepe(m,@pdefunold,@icfunold,@bcfunold,xmesh,tspanold);
uold(1,:) = solold(end,:,1);
uold(2,:) = solold(end,:,2);
tspannew=...;
solnew=pdepe(m,@pdefunnew,@(x)icfunnew(x,uold),@bcfunnew,xmesh,tspannew);
...
function u0 = icfunnew(x,uold)
global counter
counter=counter+1;
u0(1)=uold(1,counter);
u0(2)=uold(2,counter);
Maybe you will have to adjust the variable "counter" from 0 to 1 in the calling program.
Best wishes
Torsten.
  2 Comments
Eric
Eric on 30 Jul 2015
That solved it! The one additional thing I had to do was 'convert' the 2 u0 values to a column vector. Thank you sooooo much, and kudos for all the help you've given posters in the past!!!
Hashim
Hashim on 14 Apr 2021
Edited: Hashim on 15 Apr 2021
Hi, I think I am trying to do something similar where I have different IC for two different time spans. Would it be possible if I could get some help with my problem? @Torsten
function pdepe_philip_v3a
% 1-b Extend the above calculation to model a potential step to any chosen
% potential assuming that the ratio of the surface concentrations of Os2+
% and Os3+ obeys the Nernst equation both before and after the potential
% step. Now we want to further add a potential step which will make it a
% double potential step.
global D n F R Os_bulk
n = 1; % No. of Electrons Involved
F = 96485.3329; % sA/mol
R = 8.3145; % kgcm^2/s^2.mol.K
D = 5e-06; % cm^2/s
Area = 1; % cm^2
Os_bulk = 1e-05; % mol/cm^3
N = 10;
m = 0; % Cartesian Co-ordinates
% logarithmized xmesh for semi infinite boundary condition
x = logspace(log10(0.000006), log10(0.06), N); x(1) = 0;
% Time Span-1
t1 = linspace(0, 10, N); % s
c0 = Os_bulk;
E = 0.8;
ic_arg = @(x)c0.*ones(size(x));
sol1 = pdepe(m, @pdev3apde, @(x)pdev3aic(x,ic_arg),@(xl, ul, xr, ur, t)...
pdev3abc(xl, ul, xr, ur, t, E, c0), x, t1);
c1 = sol1(:, :, 1);
I_num = ( n*F*Area) * c1(1,:) .* sqrt(D./(t1*pi));
% I Flux For Time Span-1
figure(1);
plot(t1, I_num, 'r--', 'LineWidth',1);
xlabel('t (s)'); ylabel('I (A/cm^2s)');
hold on
% Time Span-2
t2 = linspace(10, 20, N); % s
E = 1.6;
c0 = (c1(1,:))';
ic_arg = @(x)(c0).*ones(size(x));
sol2 = pdepe(m, @pdev3apde, @(x)pdev3aic(x,ic_arg),@(xl, ul, xr, ur, t)...
pdev3abc(xl, ul, xr, ur, t, E, c0), x, t2);
c2 = sol2(:, :, 1);
I_num1 = ( n*F*Area) * c2(1,:) .* sqrt(D./(t2*pi));
% I Flux For Time Span-2
figure(1);
plot(t2, I_num1, 'b--', 'LineWidth',1);
xlabel('t (s)'); ylabel('I (A/cm^2s)');
hold on
end
%% pdepe Function That Calls Children
%%
function [a, f, s] = pdev3apde(x, t, u, DuDx)
global D
a = 1;
f = -D*DuDx;
s = 0;
end
%% Initial Condition
%%
function u0 = pdev3aic(x, ic_arg)
u0 = ic_arg(x);
end
%% Boundry Conditions
%%
function [pl, ql, pr, qr] = pdev3abc(xl, ul, xr, ur, t, E, c0)
global n F R Os_bulk
E0 = 0.2;
alpha = exp((E-E0).*((n*F)/(R*298.15)));
pl = ul - (c0./(1+alpha));
ql = 0;
pr = ur - Os_bulk;
qr = 0;
end
%%
%%

Sign in to comment.

Categories

Find more on Partial Differential Equation Toolbox in Help Center and File Exchange

Community Treasure Hunt

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

Start Hunting!