What does Assignment has more non-singleton rhs dimensions than non-singleton subscripts mean?

Hey in a project I am trying to build a 3d matrix based off another 3d matrix to find a final temperature profile. The code up to the error is:

X = 0:0.1:5;
Y = 0:0.1:1;
Z = 0:0.1:20;

layer = zeros(length(X),length(Y));

layer1 = layer;
layers = layer;
layer1(:,:) = 52;
layers(:,:) = 22;

temperature_profile(:,:,1) = layer1;

for x = 2:length(Z)
    temperature_profile(:,:,x) = layers;
end

layers2 = layer;
layers2(:,:) = 1;

for x = 1:length(Z)
    temperature_profile2(:,:,x) = layers2;
end

H = zeros(length(X),length(Y),length(Z));
H(:,:,:) = 0.0000001;
G = ones(length(X),length(Y),length(Z));

V = temperature_profile2-temperature_profile;
t = 0;
Cp = 1;
V = 1;
A = 1;
rho = 1;
h = 1;
T0 = 22;
T1 = 52;
k = 1;

T = V < H;
while T ~= G
    for z = 2:length(Z)
        for y = 1:length(Y)
            for x = 1:length(X)
                if z == 1
                    temperature_profile2(x,y,z) = 52; 
                elseif z == 2
                    if y == 1
                        if x == 1
                            temperature_profile2(x,y,z) = ((0.00001)./(V.*rho.*Cp)).*(V.*k.*((((temperature_profile(x+2,y,z)-temperature_profile(x,y,z))./(0.2^2)))+(((temperature_profile(x,y+2,z)-temperature_profile(x,y,z))./(0.2^2)))+(((temperature_profile(x,y,z+2)-3.*temperature_profile(x,y,z)+2.*temperature_profile(x,y,z-1))./(4.*(0.2^2))))+A.*h.*(T0-T1)))+temperature_profile(x,y,z);

when it is run the Assignment has more non-singleton rhs dimensions than non-singleton subscripts mean error pops up and i'm not really sure what it means.

Answer 1

[Rory Greer]

Figured out the problem V was later replaced with a 3d matrix later in the code on accident, the problem is fixed.

Answer 2

[the cyclist]

Your code (after adding 7 "end" statements) has been running for over 10 minutes for me, neither finishing nor crashing. Is it possible to post an example that errors out more quickly?

[Walter Roberson]

The posted code does not change T or G, so if it made it into the while loop at all, it is going to infinite loop because of the while

Answer 3

[Image Analyst]

When I get rid of the while, it runs with no error. Please post all the error - that means ALL the red text, not just part of it like you did.

This runs with no error:

X = 0:0.1:5;
Y = 0:0.1:1;
Z = 0:0.1:20;

layer = zeros(length(X),length(Y));

layer1 = layer;
layers = layer;
layer1(:,:) = 52;
layers(:,:) = 22;

temperature_profile(:,:,1) = layer1;

for x = 2:length(Z)
  temperature_profile(:,:,x) = layers;
end

layers2 = layer;
layers2(:,:) = 1;

for x = 1:length(Z)
  temperature_profile2(:,:,x) = layers2;
end

H = zeros(length(X),length(Y),length(Z));
H(:,:,:) = 0.0000001;
G = ones(length(X),length(Y),length(Z));

V = temperature_profile2-temperature_profile;
t = 0;
Cp = 1;
V = 1;
A = 1;
rho = 1;
h = 1;
T0 = 22;
T1 = 52;
k = 1;

T = V < H;
%   while T ~= G
for z = 2:length(Z)
  for y = 1:length(Y)
    for x = 1:length(X)
      if z == 1
        temperature_profile2(x,y,z) = 52;
      elseif z == 2
        if y == 1
          if x == 1
            temperature_profile2(x,y,z) = ((0.00001)./(V.*rho.*Cp)).*(V.*k.*((((temperature_profile(x+2,y,z)-temperature_profile(x,y,z))./(0.2^2)))+(((temperature_profile(x,y+2,z)-temperature_profile(x,y,z))./(0.2^2)))+(((temperature_profile(x,y,z+2)-3.*temperature_profile(x,y,z)+2.*temperature_profile(x,y,z-1))./(4.*(0.2^2))))+A.*h.*(T0-T1)))+temperature_profile(x,y,z);
          end
        end
      end
    end
  end
end
msgbox('Done');
%   end

Answer 4

[Walter Roberson]

You define temperature_profile and temperature_profile2 as three dimensional objects.

You define V = temperature_profile2-temperature_profile, so V is a three dimensional object.

Your code has

                              temperature_profile2(x,y,z) = ((0.00001)./(V.*rho.*Cp)).*(V.*k.*((((temperature_profile(x+2,y,z)-temperature_profile(x,y,z))./(0.2^2)))+(((temperature_profile(x,y+2,z)-temperature_profile(x,y,z))./(0.2^2)))+(((temperature_profile(x,y,z+2)-3.*temperature_profile(x,y,z)+2.*temperature_profile(x,y,z-1))./(4.*(0.2^2))))+A.*h.*(T0-T1)))+temperature_profile(x,y,z);

On the left side we can see assignment to a single location.

On the right side, every reference to temperature_profile and temperature_profile2 are subscripted so those reference are all scalars as well. k and rho and cp and T0 and T1 and A are all scalars as well, so those are fine.

But the code refers to V in a couple of places, and V is a three dimensional object, so the result of the right hand side is going to be a three dimensional object. And you cannot store that three dimensional object into a scalar.

I suspect you want the references to be to V(x,y,z)

It looks to me that you could probably vectorize the code. For example,

temperature_profile(x+2,y,z)-temperature_profile(x,y,z)

could potentially be

temperature_profile(3:end,:,:)-temperature_profile(1:end-2,:,:)

The boundary conditions like filling in the 52 can be handled in vectorized form as well by expanding the matrix by one in that direction and writing in 52 to that slice at each step.