ode does not converge

Hi,

I have the following problem with code on the sliding mode control, solving a non linear problem:

running the code, I do not have errors or warning, the integration starts, but does not end, looks like enter in a sort of infinite loop.

Showing some result in the function and running I realized that:

-- with a tollerance about 1e-13 (I always use this), the results are always equal, so the integration carries on always the same result and it will not go on,

-- reducing thetollerance at 1e-5 , I change see results that change, but integration never end, maybe go to converge, but soon starting to converge.

these happen even if I use a small intervall time.

many many thanks for your time and help

best regards

Answer 1

[darova]

I tried

tt = [0 1];
options=odeset('RelTol',1e-3,'AbsTol',1e-3);

[darova]

I suggest you to use structure instead of global variables

S.J = J;
S.rho1_0_costante = rho1_0_costante;
S.sigma1 = sigma1;
S.lambda1 = lambda1;
S.L = L;
[t,X] = ode113(@eq_moto_assetto,[0 1],cond_in,options,S);

and function

function [ derivate ] = eq_moto_assetto(t, X ,S)
J = S.J;
rho1_0_costante = S.rho1_0_costante;
sigma1 = S.sigma1;
lambda1 = S.lambda1;
L = S.L;

[Alessandro Vitiello]

Thank you very much, now run, but does not work, so I suppose something wrong in the equations.

just a little question, why do you suggest to use structure instead of global ?

Again, many many thanks for your time and consideratoins

best regards

[darova]

• now run, but does not work

Don't understand?

• why do you suggest to use structure instead of global ?

Im not an experinced programmer but only heard a lot not to use global variables

Some links: LINK1, LINK2

[Alessandro Vitiello]

thanks for the links.

now run, but does not work : in the sense that the code compile ( no error, no warning) but the results are wrong, their are not what I expect, errors should go to zero, (I integrate for longer time) . I do not know if it occurs because the tollerance is too small ( in gereral I use 1e-13) or because there is a bug error in the model, maybe the second.

By the way, thank you again for your help.

regard