'Matrix dimensions must agree.' while invoking ode15i

1 次查看（过去 30 天）
Zoltán Csáti
Zoltán Csáti 2012-2-24
I would like to solve a second order ode system consisting of two second order implicit differential equations (this system describes the motion of a point of mass and the centroid of a rigid body (a brick), if the mass is attached to the rigid body via a rope and swings as a pendulum; the motion is taken place in the xy plane and is given by two coordinates x and phi).
Here are the equations of motion:
(m1+m2)*x'' + m2*l*phi''*cos(phi) - m2*l*(phi')^2*sin(phi) = 0 x''*l*cos(phi) + l^2*phi'' = -g*l*sin(phi)
let X = [x x' phi phi'] and Y = X' = [x' x'' phi' phi''], then here is my function:
function Z = implicit(t, X, Y) m1 = 1; % mass of the brick m2 = 1; % mass of the l = 5; % length of the string g = 9.81; % gravitational constant Z = zeros(4,1); Z(1) = (m1+m2)*Y(2)+m2*l*Y(4).*cos(X(3))-m2*l*(Y(3)).^2.*sin(X(3)); Z(2) = Y(2).*cos(X(3))+l*Y(4)+g*sin(X(3)); Z(3) = Y(1)-X(2); Z(4) = Y(3)-X(4);
My first question: have I implemented the system well?
After, I called the solver this way: [t y] = ode15i(@implicit, [0 2], [0 0 pi/4 0], [0 0 0 0]);
Now the initial conditions are: initial coordinate, velocity, acceleration of the brick is 0, initial angle of the pendulum is pi/4, initial angular velocity and angular acceleration is 0.
And I get the following error:
??? Error using ==> max Matrix dimensions must agree.
Error in ==> odenumjac at 103 yscale = max(abs(y),thresh);
Error in ==> ode15ipdupdate at 30 [dfdyp,dfdyp_options.fac,NF] = odenumjac(odefun,{t,y,yp,extras{:}},f,dfdyp_options);
Error in ==> ode15ipdinit at 76 [dfdy,dfdyp,dfdy_options,dfdyp_options,nfcn] = ...
Error in ==> ode15i at 175 [Jac,dfdy,dfdyp,Jconstant,dfdy_options,dfdyp_options,nfcn] = ...
Please help me.
Thank you, András

请先登录，再回答此问题。

采纳的回答

Zoltán Csáti
Zoltán Csáti 2012-4-15
I managed to solve those equations explicitly.
  1 个评论
Jan
Jan 2012-4-15
The explicit solution is not necessarily more efficient for numerical simulations or visualizations. Solving a small linear linear system in each step can be cheaper than a bunch of trigonometric functions.
However, currently the original question is hard to read due to the missing code formatting. Please edit the question - code formatting is trivial. Perhaps you want to post the solution?

请先登录，再进行评论。

更多回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Numerical Integration and Differential Equations 的更多信息

标签

产品

Community Treasure Hunt

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

Start Hunting!

Translated by