mfile version of Simulink integrator

I'm running a simulation model in Simulink and I want to transform it over to m-code. In my Simulink model I use an integrator and to be sure I was making my m-code correctly I took time-stepped data on both sides of the integral and compared to when I ran those points through my m-code. Sadly they did not match up.

The solver I am using in Simulink is ode3 ... does anyone know the formula for the integrator block in Simulink using an ode3 solver?

BTW, I do not have the toolboxes to do code generation from simulink.

[Ryan G]

Why do you want to convert a model to m-code?

[Scott Hansen]

Because I need to for what I'm working on ... why is that relevant to the question?

[Jorge Gustavo Vazquez Duran]

hi man, i don´t have a solution but i have the same problem like you.

did you solve this or keeping it??

thanks.! regreat!!

Answer 1

[Guy Rouleau]

This should get you started:

http://en.wikipedia.org/wiki/Bogacki%E2%80%93Shampine_method

You might also be interested to look at the code for the ode3.m function in MATLAB.

[Scott Hansen]

I was looking at that exact page yesterday but had a hard time figuring out how to implement the equations. Please let me explain my confusion and hopefully you can point me in the right direction.

My format looks like this:

xdot -----> 1/s -----> x

xdot = input to integrator (acceleration in my system)

1/s = integrator

x = output of integrator (velocity in my system)

So I have time stamped data sets for xdot and x.

When it comes to the wiki page you recommended ... these equations are usually used to determine the next x value from previous values of x (at least that is how I understand it). When looking at the equations on the wiki page I'm not confident which variables are xdot, x and so forth. My understanding is below:

h_n = timestep from x_n to x_n+1

t_n = time at y_n

k1, k2, ... = calculated multipliers

y_n' = f(t_n, y_n) = xdot

y_n = x

Equations I am referring to: http://upload.wikimedia.org/math/6/8/b/68bf5ee4cc6f90e4da54a53046e6a631.png

If this is all correct then my next issue is that I don't have a function to evaluate these at (and neither does the integrator). It doesn't know what the next xdot value will be until it is applied as an input ... so how can I estimate the x value without an understanding of this?

I apologize for the length of this email but I'm really needing some assistance with this. Thanks in advance for your help.

S