Can someone please tell me where is my fault?

[Milan Grahovac]

[Milan Grahovac]

Answer 1

[Walter Roberson]

At line 11 insert

AKM = zeros(4,1);

or

AKM = zeros(1,4);

as appropriate. That way Simulink will not have to "grow" AKM as you assign the values to AKM(1), AKM(2) and so on.

[Milan Grahovac]

Thank you very much, this helped me. But i have one more question, Ramp block (M_ter) should load values from excel table and i cant load with xlsread. Can you tell me how i can load the values for Ramp block from excel? I get error if i do that with xlsread.

[Walter Roberson]

Does your data happen to have some NaN (not a number) ?

What is the error you get when you use xlsread() ?

[Milan Grahovac]

[Milan Grahovac]

[Milan Grahovac]

These errors i get.

[Walter Roberson]

1) Before your call to xlsread() you need to insert coder.extrinsic()

coder.extrinsic('xlsread');

The code that is generated will be usable for lower levels of acceleration but not for deployment or Real Time Target.

2) Simulink believes that the output port for the function is configured to be 1 x 3, but you are attempting to store as 1 x 4. Either reconfigure the output port or do not store the 4th output in AKM.

[Milan Grahovac]

Thank you very much sir, i will reconfigure model and then i hope it will work. Thank you one more time. Best regards!

[Milan Grahovac]

I have reconfigured my model and i have made it with 4 fcn blocks(1 block for every equation) . Please can u answer me this?

[Milan Grahovac]

[Milan Grahovac]

[Milan Grahovac]

[Milan Grahovac]

[Milan Grahovac]

[Milan Grahovac]

[Milan Grahovac]

This is 1st block

function nnovo =fcn(n, M_ter, I_phod)

coder.extrinsic('xlsread'); filename='Moment_terecenja' ; T=xlsread(filename) ; M_ter=T(:,1) ; I_phod=4.17 ; n_phod=1498 ;

nnovo=-0.0284*(M_ter*M_ter)-1.3569*M_ter+n_phod ;

nnovo=zeros(1,1) ;

2nd block

function I =fcn(n, M_ter, I_phod)

coder.extrinsic('xlsread'); filename='Moment_terecenja' ; T=xlsread(filename) ; M_ter=T(:,1) ;

I=0.0033*(M_ter*M_ter) + 0.1019*M_ter + I_phod ;

I=zeros(1,1) ;

3rd block

function eta =fcn(n, M_ter, I_phod)

coder.extrinsic('xlsread'); filename='Moment_terecenja' ; T=xlsread(filename) ; M_ter=T(:,1) ;

eta=-0.001*(M_ter*M_ter)+0.0588*M_ter ;

eta=zeros(1,1) ;

4th block

function cos_phi =fcn(n, M_ter, I_phod)

coder.extrinsic('xlsread'); filename='Moment_terecenja' ; T=xlsread(filename) ; M_ter=T(:,1) ;

cos_phi=-0.0008*(M_ter*M_ter)+0.0542*M_ter-0.3941 ;

cos_phi=zeros(1,1) ;

[Milan Grahovac]

I hope u can see reworked model.

[Milan Grahovac]

This is error report

Subscripting into an mxArray is not supported.

Function 'f(n)' (#31.131.137), line 6, column 7: "T(:,1)" Launch diagnostic report.

Undefined function or variable 'M_ter'. The first assignment to a local variable determines its class.

Function 'f(n)' (#31.187.192), line 13, column 16: "M_ter"

Errors occurred during parsing of MATLAB function 'f(n)'

Error in port widths or dimensions. Output port 1 of 'Modes_sa_4_bloka/f(n)/M_ter' is a one dimensional vector with 1 elements.

[Walter Roberson]

I do not understand why you have M_ter and I_phod listed as parameters but you then read M_ter from a file and you use a constant for I_phod . I think you have constructed your flow incorrectly.

I have written the below code following your logic.

One of the parts that I was not certain of was whether your Moment_terecenja file is intended to have multiple rows. The sizes of your variables were not consistent if it could have multiple rows, so in the code I assumed that there should be only one row. I assumed that there were a maximum of 64 values on that one row.

Is your Moment_terecenja a text file or a .xls file? Please describe the content for it, the number of fields and the number of rows. An example might help.

First I introduce a new function to fetch the M_ter. I think you should not have to do this, as I think you should be passing it as an input to your blocks. However, I assume you had a reason for reading it from the file instead of using the parameter passed. By putting the code in a function of its own, we only have to get it right once.

None of this code has been tested. I am working from the documentation, as I have never had access to the Simulink software.

function M_ter = get_Mter()
%this is a new function. It must be stored in get_Mter.m

    coder.extrinsic('xlsread');
    filename = 'Moment_terecenja' ;

    coder.varsize('T', [1 1]);
    T = zeros(1,64);
    T = xlsread(filename) ;

    M_ter = zeros(1,1);
    M_ter = T(1,1);

First block

function nnovo = fcn(n, M_ter, I_phod)

    I_phod = 4.17 ;
    n_phod = 1498 ;

    M_ter = zeros(1,1);
    M_ter = get_Mter();

    nnovo = zeros(1,1) ;
    nnovo = -0.0284 * M_ter.^2 - 1.3569 * M_ter + n_phod ;

Second block

function I = fcn(n, M_ter, I_phod)

    I_phod = 4.17 ;

    M_ter = zeros(1,1);
    M_ter = get_Mter();

    I = zeros(1,1) ;
    I = 0.0033 * M_ter.^2 + 0.1019 * M_ter + I_phod ;

Third block

function eta = fcn(n, M_ter, I_phod)

    M_ter = zeros(1,1);
    M_ter = get_Mter(); 

    eta = zeros(1,1) ;
    eta = -0.001 * M_ter.^2 + 0.0588 * M_ter ;

Fourth block

function cos_phi = fcn(n, M_ter, I_phod)

    M_ter = zeros(1,1);
    M_ter = get_Mter(); 

    cos_phi=zeros(1,1) ;  
    cos_phi = -0.0008 * M_ter.^2 + 0.0542 * M_ter - 0.3941 ;

[Milan Grahovac]

Moment_terecenja is xls file (Moment_terecenja.xls) and that file consist of 7 values in one column and that values i need for Ramp. What i should write between this parenthesis get_Mter(HERE)?

[Milan Grahovac]

I_phod are constant and n_phod are constant.

[Walter Roberson]

You do not need anything in the parenthesis, just get_Mter()

If Moment_terecenja is a column vector of 7 values, then your previous code would have failed. Your had

M_ter * M_ter

which would have been a (7 x 1) matrix-multiplied by (7 x 1). For matrix multiplication, the "inner dimensions" must agree -- the number of columns of the left array must match the number of rows of the right array. You can use expressions such as

M_ter' * M_ter

which would result in a (1 x 7) by (7 x 1) which would give a 1 x 1 result, or you could have

M_ter * M_ter'

which would result in a (7 x 1) by (1 x 7) which would give a 7 x 7 result.

Your expressions then proceed to add something multiplied by M_ter . If the M_ter * M_ter is intended to give a 1 x 1 result then when it would be acceptable in MATLAB to add the 7 x 1 vector that is M_ter, giving a 7 x 1 result. That in itself would not be a problem. However, your existing code appears to expect each block to return a 1 x 1 scalar result. It is possible to select the first value out of the 7 x 1 vector and ignore the rest, but if you are doing that then you can skip some of the calculations for efficiency.

If the M_ter * M_ter is intended to give a 7 x 7 result, then adding the 7 x 1 vector is a problem . You could replicate the vector to be 7 x 7 with all columns the same, and add that. But the result would be 7 x 7 and again you seem to be expecting a 1 x 1 result.

I will wait for you to explain what you want the M_ter * M_ter to mean before I offer code changes.

[Milan Grahovac]

I write (M_ter*M_ter) cus it's square of M_ter from same row. Excel table with values of M_ter have 1 column and 7 rows and i need results for every M_ter from excel table. Result of every block should be 7 x 1 vector and then i need to plot every block to see function graphicly. X axis should be M_ter and y axis in first block should be nnovo (for example in first block).

[Milan Grahovac]

In simulink ramp should use my values from excel for M_ter, cus M_ter is Ramp block.

[Milan Grahovac]

Only value that change during simulation is M_ter. Every block should give me 7 results for every change of M_ter.

[Walter Roberson]

I need to do some work on my house now, but I will return to this later.

I do not understand what you are saying about Simulink and ramp. Are you saying that the values for M_ter are outputs of a Ramp block that you then write to an excel file and then read back? If so then why not have the ramp block as direct input to these sets of blocks without using the excel file?

[Walter Roberson]

I still do not understand about the Ramp or why M_ter is being read from a file instead of being passed as a signal, or why you are passing I_phod as a signal but ignoring the value. Still here is the version with 7 values expected for M_ter

function M_ter = get_Mter()
%this is a new function. It must be stored in get_Mter.m

    coder.extrinsic('xlsread');
    filename = 'Moment_terecenja' ;

    coder.varsize('T', [1 0]);
    T = zeros(64,1);          %we only expect 7 values but allow up to 64 before we bomb
    T = xlsread(filename) ;

    M_ter = zeros(7,1);
    M_ter = T(1:7,1);

First block

function nnovo = fcn(n, M_ter, I_phod)

    I_phod = 4.17 ;
    n_phod = 1498 ;

    M_ter = zeros(7,1);
    M_ter = get_Mter();

    nnovo = zeros(7,1) ;
    nnovo = -0.0284 * M_ter.^2 - 1.3569 * M_ter + n_phod ;

Second block

function I = fcn(n, M_ter, I_phod)

    I_phod = 4.17 ;

    M_ter = zeros(7,1);
    M_ter = get_Mter();

    I = zeros(7,1) ;
    I = 0.0033 * M_ter.^2 + 0.1019 * M_ter + I_phod ;

Third block

function eta = fcn(n, M_ter, I_phod)

    M_ter = zeros(7,1);
    M_ter = get_Mter(); 

    eta = zeros(7,1) ;
    eta = -0.001 * M_ter.^2 + 0.0588 * M_ter ;

Fourth block

function cos_phi = fcn(n, M_ter, I_phod)

    M_ter = zeros(7,1);
    M_ter = get_Mter(); 

    cos_phi=zeros(7,1) ;  
    cos_phi = -0.0008 * M_ter.^2 + 0.0542 * M_ter - 0.3941 ;

[Milan Grahovac]

I have finished my work succesfully, thank you for all your help. Greetings!