How can I solve this error xdot = A*x + B*u; y = C*x + D*u; Incorrect dimensions for matrix multiplication A*x

[Dina]

@Melda Ulusoy @Walter Roberson

[Dina]

This is my code:

[Dina]

this is the mtlab function block in detail .. i need help ??

[Torsten]

It seems that x is 4x1 and A is 6x6. So how should A*x work in your opinion ?

[Sam Chak]

@Dina, How many initial values have you entered in the Integrator block? 4 or 6?

[Dina]

how can i know that? the input to the integrator block is xdot which is afunction of matrix A,B,U,X, A[6x6] , B[6x1],u =[1] , x default as deduced from equation and i try to change its size but i cant ??

Answer 1

[Walter Roberson]

Your A is 6x6. Your x is 4x1. You cannot matrix multiply the two.

Your A should have (number of outputs) rows and (number of inputs) columns

[Dina]

yes ,I know that but the matrix x how to change its size as it deduced from the equation x dot =AX+BU and A ,B ,U KNOWN but x matrix size deduced automatically as the previous equation and ican't change it ?how can i change its size? thanks in advance

[Walter Roberson]

please attach your model

[Sam Chak]

Hi @Dina

This issue is, "What caused when the state vector is in the feedback loop?"

Can you double-click the Integrator block and take a snapshot?

Also, I suspect that the culprit is the ydot external signal is a vector . Yet multiple ydot's appear in the matrix elements (when scalars are expected). How do you interpret that mathematically?

[Walter Roberson]

I am wondering whether that "X" block that shows [4x1] in and [2x1] out has been configured as expecting an input of length 4, and that information is then flowing backwards ?

It would help to have the model to examine.

[Sam Chak]

That's a selector switch. That could also be a reason. Hope @Dina can attach "a portion of the model", if reluctant to share the full model. Or double-click the Selector block and take a snapshot.

[Dina]

I tried to change the size from internal size from model explorer and finally the size of x changed correctely as follows but i still solving some other problems thanks for help

[Dina]

[Walter Roberson]

Question: why are your Vx and x the same signal ?

You have the output from the 1/s going along the bottom and forking into the oval for "1" with the other fork going to input x . But the oval for "1" goes into Vx so both x and Vx are the same signal.

Notice that the Vx is specifically dimensioned as 1 even though Vx port is size -1 (inherit) so Simulink must somehow figure that the oval with "1" is a scalar. But the fork that went into x is derived size ? rather than 1, so it must be something about the oval . Is there a different goto "1" somewhere?