How can I include a finite integral in a system of non-linear 2nd order differential equation using bvp4c function and solve numerically? Tried a test with Caughey example for large-amplitude whirling of an elastic string
1 view (last 30 days)
Show older comments
In fine, I looking for a numerical solution using bvp4c of a 3rd order differential for a clamped circular membrane where the rhs term depends on a finite integral of the soolution y.
The test example is described in the book "Solving-odes-with-matlab" from Shampine Gladwell Thompson example 1.19 et 3.9, methodology is provide but no correction is given.
Quote ------
Caughy (1970) describes the large-amplitude whirling of an elastic string by a BVP consisting of the ODE:
with boundary conditions
Here α is a physical constant with 0 <α<1. Because the whirling frequency ω is to be determined as part of solving the BVP, there must be another boundary condition. Caughey specifies the amplitude ε of the solution at the origin: µ(0) = ε.
An unusual aspect of this problem is that an important constant H is defined in terms of the solution µ(x) throughout the interval of integration:
Formulate this BVP in standard form. As in the Sturm–Liouville example, you can introduce a new variable y3(x), a first-order ODE, and a boundary condition to deal with the integral term in the definition of H. The trick to dealing with H is to let it be a new variable y4(x). It is a constant, so this new variable satisfies the first-order differential equation y4' =0. It is given the correct constant value by the boundary condition resulting from the definition of H:
Unquote ------
I tried to code several ways, but always get one od the variable y(3) or y(4) out of bound. I miss some methodlogy to get a correct output. Thanks for any help.
The following code include following analysis for differential equation:
and
y4(x)=H=cste and
and boundary conditions
and
function dydx = bvpfcn(x,y,OMEGA)
ALPHA=0.5 ; EPSILON=0.5;
dydx = zeros(4,1);
dydx = [y(2);...
-OMEGA.^2.*(((1-ALPHA.^2)./y(4)).*(1./(1+y(1).^2).^0.5)+ALPHA.^2).*y(1);...
(1./(1+y(1).^2).^0.5);...
0];
end
function res = bcfcn(ya,yb,OMEGA)
ALPHA=0.5 ; EPSILON=0.5;
res = zeros(4,1);
res = [ya(2);...
yb(2);...
yb(3)-(asinh(yb(1))-asinh(ya(1)));...
yb(4)-((1./ALPHA.^2).*(1-(1-ALPHA.^2).*yb(3)))];
end
function g = guess(x)
g = [cos(pi*x);-pi*sin(pi*x)];
end
xmesh = linspace(0,1,10);
OMEGA=pi; % initial condition
solinit = bvpinit(xmesh,@guess,OMEGA);
tol=1E-3;
options = bvpset("RelTol",tol,"AbsTol",tol,"Stats","on",'Nmax',10000);
%
tic;
Testsol4c = bvp4c(@bvpfcn, @bcfcn, solinit);
toc;
2 Comments
Answers (0)
See Also
Categories
Find more on Boundary Value Problems in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!