I am writing an FEM code in matlab. I want to make the sizes of repmat(N_dash, 3, 1) and repelem(ey.', 3, 1) the exact same. Original size of repmat(N_dash, 3, 1) is 3x3 and that of repelem(ey.', 3, 1) is 9x1. Any help will be much appreciated.

for epsilon3=-1:1:1
N_dash(1)=-0.5+epsilon3;
N_dash(2)=-2*epsilon3;
N_dash(3)=0.5+epsilon3;
end

ly=(x(5)-x(2)); (x,y,z are constants like 0.9876,0.8876..etc)
my=(y(5)-y(2));
ny=(z(5)-z(2));

lve=sqrt(((ly)^2)+((my)^2)+((ny)^2));       %%vertical length of element

ey = [ly; my; ny] / lve;

A=repmat(N_dash, 3, 1) .* repelem(ey.', 3, 1)

[Walter Roberson]

What size does the output need to be?

Answer 1

[madhan ravi]

a=repelem(ey.', 3, 1)

a= reshape(a,3,3) %and then multiply a with repmat 

A=repmat(N_dash, 3, 1) .* a

[madhan ravi]

upload the datas required to test

[Virajan Verma]

Still getting the same error, though i have used size and numel command and sizes and no. of elements are same. I dont know what is the problem. The LHS of eqaution is B_matrix(1,1:8) which is defined as:

for i=1:8 
for j=1:26 
B_matrix(i,j)=0; 
end 
end

[madhan ravi]

hows is this related to the question??

[Virajan Verma]

%%Strain for the (Axial) Extensional Flexural Effects
for epsilon3=-1:1:1
N_dash(1)=-0.5+epsilon3;
N_dash(2)=-2*epsilon3;
N_dash(3)=0.5+epsilon3;
end

(TAKE ANY CONSTANT VALUE OF x,y,z like 0.8798,0.3456)
%%Tangent vector (ey) about y dirsction
ly=(x(5)-x(2));
my=(y(5)-y(2));
ny=(z(5)-z(2));

lve=sqrt(((ly)^2)+((my)^2)+((ny)^2));       %%vertical length of element

ey = [ly; my; ny] / lve;

for i=1:8
      for j=1:26
         B_matrix(i,j)=0;

A=reshape(repmat(ey, 3, 1),3,3)
  for i=1:8
B_matrix(1,1:8)=1/s*repmat(N_dash, 3, 1) .* A
end
end
end

[Virajan Verma]

Here is the code required and u will know how this is related.

[madhan ravi]

>> COMMUNITY
Undefined function or variable 'y'.

Error in COMMUNITY (line 9)
my=(y(5)-y(2));

>>

[madhan ravi]

whats x,y ,s and z ? provide the values

[Virajan Verma]

I have a separate input file for that. U can take any decimal values like 0.9821,0.8736 etc

[madhan ravi]

value of s?

[madhan ravi]

for epsilon3=-1:1:1
N_dash(1)=-0.5+epsilon3;
N_dash(2)=-2*epsilon3;
N_dash(3)=0.5+epsilon3;
end
% (TAKE ANY CONSTANT VALUE OF x,y,z like 0.8798,0.3456)
%%Tangent vector (ey) about y dirsction
ly=2-5;
my=7+18;
ny=4;
lve=sqrt(((ly)^2)+((my)^2)+((ny)^2));       %%vertical length of element
ey = [ly; my; ny] / lve;
for i=1:8
      for j=1:26
         B_matrix(i,j)=0;
A=reshape(repmat(ey, 3, 1),3,3)

B_matrix=1/4*repmat(N_dash, 3, 1) .* A

end
end

[madhan ravi]

It works I made slight modification also I assumed value of s as 4(if it matters)

[Virajan Verma]

x1=0  
y1=0  
z1=0
x2=0.5  
y2=0  
z2=0.049406182
x3=1  
y3=0  
z3=0.097122055
x5=0.5  
y5=0.51  
z5=0.049406182

    for i=-1:1:1
        epsilon3=(i)
    lx=(epsilon3-0.5)*x(1)-(2*epsilon3)*x(2)+(epsilon3+0.5)*x(3);
    mx=(epsilon3-0.5)*y(1)-(2*epsilon3)*y(2)+(epsilon3+0.5)*y(3);
    nx=(epsilon3-0.5)*z(1)-(2*epsilon3)*z(2)+(epsilon3+0.5)*z(3);

    %%Jacobian Factor
    s=sqrt(((lx)^2)+(((mx)^2)+(((nx)^2))));

[madhan ravi]

x(1)=0  
y(1)=0  
z(1)=0 
x(2)=0.5  
y(2)=0  
z(2)=0.049406182 
x(3)=1  
y(3)=0  
z(3)=0.097122055 
x(5)=0.5  
y(5)=0.51  
z(5)=0.049406182
epsilon3=-1:1:1
for i=1:numel(epsilon3)

lx(i)=(epsilon3(i)-0.5)*x(1)-(2*epsilon3(i))*x(2)+(epsilon3(i)+0.5)*x(3);
mx(i)=(epsilon3(i)-0.5)*y(1)-(2*epsilon3(i))*y(2)+(epsilon3(i)+0.5)*y(3);
nx(i)=(epsilon3(i)-0.5)*z(1)-(2*epsilon3(i))*z(2)+(epsilon3(i)+0.5)*z(3);
end
%%Jacobian Factor
s=0.5025;
% (TAKE ANY CONSTANT VALUE OF x,y,z like 0.8798,0.3456)
%%Tangent vector (ey) about y dirsction

lve=sqrt(((ly)^2)+((my)^2)+((ny)^2));       %%vertical length of element
ey = [ly; my; ny] / lve;
for i=1:8
      for j=1:26
         B_matrix(i,j)=0;
A=reshape(repmat(ey, 3, 1),3,3)

B_matrix=(1/s)*repmat(N_dash, 3, 1) .* A

end
end

[madhan ravi]

problem solved!

[Virajan Verma]

s=0.5025

[madhan ravi]

see the comment above

[Virajan Verma]

But i am getting this error: Subscripted assignment dimension mismatch.

Error in practisee2 (line 225) B_matrix(1,1:9)=1/s*repmat(N_dash, 3, 1) .* A

[madhan ravi]

B_matrix=(1/s)*repmat(N_dash, 3, 1) .* A

[Virajan Verma]

I am working on 2015 version. Is that a problem??

[madhan ravi]

replace

B_matrix(1,1:9)=(1/s)*repmat(N_dash, 3, 1) .* A

with

B_matrix=(1/s)*repmat(N_dash, 3, 1) .* A

[madhan ravi]

if you got the answer for your question accept the answer so people know the question is solved

[Virajan Verma]

I have not got the desired result yet, otherwise i would have done that.

[madhan ravi]

whats your desired result and its size ?

[Virajan Verma]

@madhan ravi. Thanks, Got the answer.

[madhan ravi]

Really really glad that I was able to help because you were struggling for days with this problem.

[Virajan Verma]

Yes i was really struggling. Its people like u who support others. Really thanks for ur valuable suggestions and time.