I have a problem with simulink because when run the program:

Error in port widths or dimensions. Invalid dimension has been specified for input port 5 of 'Model_MGDL/Subsystem/Sum'

[Olivia Mendoza]

Error in port widths or dimensions. Invalid dimension has been specified for input port 5 of 'Model_MGDL/Subsystem/Sum'.

Error in port widths or dimensions. Output port 1 of 'Model_MGDL/Subsystem/C' is a [8x8] matrix.

Answer 1

[Olivia Mendoza]

Making corrections to the observations that made me, even so I have problems, my inputs are 8x1 and 8x8 and still I do not allow me to perform the correct process, which I still do wrong :(

[Walter Roberson]

If you have an 8 x 1 input and an 8 x 8 input, then what is the desired result? Should each row of the 8 x 1 be replicated into 8 x 8 and that added to the 8 x 8 ? Simulink does not know to do that: if you want inputs to be replicated then you have to do that with a separate block.

Answer 2

[Walter Roberson]

Your Constant block is 1 x 4. Your output of Add1 is 8 x 8. You are trying to add the two together with a Sum block. What output are you expecting?

[Olivia Mendoza]

Yes, but I do not know how to add the matrix signal even though it is of different magnitude

[Olivia Mendoza]

Is the mathematical model of a system of 4 degrees of freedom

[Walter Roberson]

Ignoring the Simulink portion of it for a moment: What is the mathematical formula for your desired outcome for adding a 1 x 4 vector to an 8 x 8 array.

For example, is your desired outcome to replicate the 1 x 4 vector once per row, taking [A B C D] to [A B C D A B C D], and then to replicate that 8 times, so you get

[A B C D A B C D
 A B C D A B C D
 A B C D A B C D
 A B C D A B C D
 A B C D A B C D
 A B C D A B C D
 A B C D A B C D
 A B C D A B C D]

and add that to your 8 x 8 Matrix?

Or should the 1 x 4 be reshaped to 2 x 2, and then that replicated as needed:

   A B A B A B A B
   C D C D C D C D
   A B A B A B A B
   C D C D C D C D
   A B A B A B A B
   C D C D C D C D
   A B A B A B A B
   C D C D C D C D

and that should be added to the 8 x 8?

Or should it be be

   A A A A A A A A
   B B B B B B B B
   C C C C C C C C
   D D D D D D D D
   A A A A A A A A
   B B B B B B B B
   C C C C C C C C
   D D D D D D D D

added to the 8 x 8?

or something else?

Answer 3

[Olivia Mendoza]

I tried with a simpler cycle and still do not do it, I do not know if I'm messing with the integrator or what it is

[Walter Roberson]

Please attach this model.

Answer 4

[Olivia Mendoza]

The simple model and the firts model

[Walter Roberson]

In Subsystem3, input #1 is C, an 8 x 8 matrix. You pass that matrix to several Fcn blocks. Fcn1 for example is coded as -36*u(1) . That is a problem because u is not a scalar or vector, and "General Expression blocks" only permit scalar or vector inputs. You either need to switch to MATLAB Function Block for those, or you need to use a reshape to reshape the 8 x 8 into a vector.

When I look at the subsystem, I see that you only ever use u(1) of the 8x8 blocks. That leads me to wonder why you are sending the entire block in, instead of selecting only the one value and passing that in?

[Olivia Mendoza]

I use fnc blocks because later I need to run the program with different values and that will make it easier to modify it, or block instead of fnc recommend me to replace it or how?

[Walter Roberson]

You may wish to recode as MATLAB Function Blocks

function y = fcn(u)
   y = -36*u(1,1);
end

and the like.

Answer 5

[Olivia Mendoza]

I made the arrays to u (1,1 |) in the matrix blocks and now it appears:

Error while parsing expression: 36*u(1,1) in 'Model_MGDL_P1/Subsystem/C/C/Subsystem/Subsystem3/Fcn1' token "," was found, where as ")" was expected

[Walter Roberson]

You cannot use 2D subscripts in an arithmetic block. That is why I was talking about MATLAB Function Blocks: they can have general MATLAB code.

Answer 6

[Olivia Mendoza]

Hi, i resolved my problem but i have a new problem with the model:

An error occurred while running the simulation and the simulation was terminated

Derivative of state '4' in block 'Model_MGDL/Subsystem/Integrator' at time 0.59593389906600114 is not finite. The simulation will be stopped. There may be a singularity in the solution. If not, try reducing the step size (either by reducing the fixed step size or by tightening the error tolerances

[Walter Roberson]

Your Matrix Inversion is probably failing. Matrix inversion is fragile. At the very least you should consider using https://www.mathworks.com/help/dsp/ref/pseudoinverse.html

Probably reformulating as a matrix division instead of multiplying by inverse would be less fragile than what you have now, but pinv would be less fragile than that. You should also consider using a MATLAB Function Block in order to be able to use the \ operator if you are using discrete time instead of continuous time.

Answer 7

[Olivia Mendoza]

I already considered other methods of inverse and sends me the same error

[Walter Roberson]

Well, create a new Question for this and post the model there.

Note: to reply to what someone else has posted, click on "Comment on this Answer" instead of creating a new Answer.