CAT arguments dimensions not consistent.

Hello,

I am getting the following messages while running the program for multiple values of 'l', (e.g: the program works well by assigning single value to l, like l=5 or l=10, l=15, l=20 etc, but doesnot work for l=5:5:20 or for l=linspace(5,20)).

error: CAT arguments dimensions not consistent.

clear all
syms logP s
P = exp(logP);
N=-P;
% l=10 % program works for individual values of l
l= 10:5:50; % problem encounter unsing a l as a vector  
%   l=linspace(10,50);
b=0.2; 
E=200;n=2;
h0=3; h1=h0;
I0=b.*h0.^3./12; I1=b.*h1.^3./12; 
L0=sqrt(P./(E.*I0-P.*n));
L1=sqrt(P./(E.*I1-P.*n));
x=(E.*I1+n.*N)./(E.*I0+n.*N);
   a11=sin(L0*s);a12=s;a13=0;a14=0;
   a21=0;a22=0;a23=-sin(L1*(l-s));a24=(s-l).*cos(L1*l);
   a31=-L0.*cos(L0*s).*cos(L1*l);a32=-cos(L1*l);a33=L1.*cos(L1*(l-s));a34=cos(L1*l);
   a41=L0.^2.*sin(L0*s).*cos(L1*l);a42=0;a43=x.*L1.^2.*sin(L1*(l-s));a44=0;
   A=[a11 a12 a13 a14;a21 a22 a23 a24;a31 a32 a33 a34;a41 a42 a43 a44];
   deltaA=det(A);
         s=1;
         delta=subs(deltaA,s);
F=matlabFunction(delta);
logP1=[];
   for j = log(0.01:10)
   logPP = fzero(F, j);
   logP1 = vertcat(logP1, logPP); 
   end
 logP1 = [logP1; logPP];
P1 = exp(logP1);
P_cr = min(P1(P1 > 0))

[Rik]

How would you do this with pen and paper if you have a vector l? Should A be a 3D matrix? Should it be wider? Taller?

It looks like you need a loop. It doesn't look to me like your computations are properly defined for a non-scalar l.

NB: you should avoid lowercase l as a variable name, because it looks a lot like a 1 in many mono-spaced fonts.

[HINA]

Thank you Rik for your valuable suggestions. I have already tried with the for loop as well, and that's also not working for me. I am sharing the code and the error

clear all
syms logP s
P = exp(logP);
N=-P;
 for l=linspace(10,50,5);
b=02; 
E=200;n=2;
h0=3; h1=h0;
I0=b*h0^3/12; I1=b*h1^3/12; 
L0=sqrt(P/(E*I0-P*n));
L1=sqrt(P/(E*I1-P*n));
x=(E*I1+n*N)/(E*I0+n*N); 
   a11=sin(L0*s);a12=s;a13=0;a14=0;
   a21=0;a22=0;a23=-sin(L1*(l-s));a24=(s-l)*cos(L1*l);
   a31=-L0*cos(L0*s)*cos(L1*l);a32=-cos(L1*l);a33=L1*cos(L1*(l-s));a34=cos(L1*l);
   a41=L0^2*sin(L0*s)*cos(L1*l);a42=0;a43=x*L1^2*sin(L1*(l-s));a44=0;
   A=[a11 a12 a13 a14;a21 a22 a23 a24;a31 a32 a33 a34;a41 a42 a43 a44]
   deltaA=det(A);
         s=1;
         delta=subs(deltaA,s);
F=matlabFunction(delta);
logP1=[];
for j = log(001:10)
   logPP = fzero(F, j); 
   logP1 = vertcat(logP1, logPP); 
end
P1 = exp(logP1);
P_cr = min(P1(P1 > 0))
 end

it returns the following error

Too many input arguments.

Answer 1

[Rik]

There are several issues with your code. After applying the correct indentation, here are my edits.

The source of your error is probably the fact that you overwrite s with 1.

l_list=linspace(10,50,5);
P_cr=zeros(size(l_list));
for n=1:numel(l_list)
    P_cr(n)=calculate_P_cr(l_list(n));
end
disp(P_cr)

  142.7270   43.8971   20.1769   11.4483    7.3452

function P_cr=calculate_P_cr(l)
%Write documentation for this function here.
%describe inputs and expected output

You should also write comments for every idea. An idea can span a single line or a few lines. Make sure other people reading your code can understand what you're doing.

%clear all

clear all is mostly equivalent to restarting Matlab every time you run your code. It will do what you need, but at an unnecessary performance cost. clear or clearvars is enough in scripts. Outside of debugging you should use functions to keep your workspace clean.

%initialize symbolic variables
syms logP s
P = exp(logP);
N=-P;

%initialize constants
b=02;
E=200;
n=2;
h0=3;
h1=h0;

I0=b*h0^3/12;
I1=b*h1^3/12;
L0=sqrt(P/(E*I0-P*n));
L1=sqrt(P/(E*I1-P*n));
x=(E*I1+n*N)/(E*I0+n*N);

A little bit of whitespace might improve how readable your code is

%create the A matrix
a11=sin(L0*s);               a12=s;         a13=0;                   a14=0;
a21=0;                       a22=0;         a23=-sin(L1*(l-s));      a24=(s-l)*cos(L1*l);
a31=-L0*cos(L0*s)*cos(L1*l); a32=-cos(L1*l);a33=L1*cos(L1*(l-s));    a34=cos(L1*l);
a41=L0^2*sin(L0*s)*cos(L1*l);a42=0;         a43=x*L1^2*sin(L1*(l-s));a44=0;
A=[a11 a12 a13 a14;a21 a22 a23 a24;a31 a32 a33 a34;a41 a42 a43 a44];

deltaA=det(A);
s=1;%this overwrites the symbolic variable with an explicit value
delta=subs(deltaA,s);
F=matlabFunction(delta);

exp_list = log(001:10);

Are you sure you don't mean log10(0.01):log10(10)?

logP1=zeros(size(exp_list));
for n=1:numel(exp_list)
    logPP = fzero(F, exp_list(n));
    logP1(n) = logPP;

You had a dynamically expanding array here. You shouldn't do that if you can determine the size beforehand.

end
P1 = exp(logP1);
P_cr = min(P1(P1 > 0));

Because exp(x)==0 will only be true for x=-inf, you could consider skipping the indexing.

end

[HINA]

A big thanks for such a detailed description. That’s a great help.