Overwriting within an if/else

Hey!

I've this code but I don't quite understand why at the first iteration j = 2 and when i = 60 the for loop at line overwrites the last input of t(:,1) - I already defined t(:,1) before initializing the if loop.

Any help would be appreciated.

clear clc

r=zeros(60,20); t=zeros(60,20); epsilontheta=zeros(60,20); epsilonn=zeros(60,20); epsilonr=zeros(60,20); den=zeros(60,20); der=zeros(60,20); det=zeros(60,20); epsilons=zeros(60,20); sigmas=zeros(60,20); sigman=(-1)*ones(60,20); sigmar=zeros(60,20); k=zeros(60,20); l=zeros(60,20); A=zeros(60,20); F=zeros(60,20);

for i =1:60 r0=110; %mm t0=1; %mm R=1.15; K=530; %N/mm^2 = MPa n=0.268; my=0.2;

    for j=1:20
        r(1,j)=r0-0.5*(j-1);
        r(i,1)=r0-1*(i-1);
    end

    %j=1
    t(1,1)=t0;
    t(i,1)=t(1,1);

    epsilonn(i,1)=0;
    epsilontheta(i,1)=0;
    epsilonr(i,1)=0;

    det(i,1)=0;
    den(i,1)=0;
    der(i,1)=0;

    epsilons(i,1)=0;
    sigmas(i,1)=0;

    k(i,1)=0;
    l(i,1)=0;

    sigmar(i,1)=0;
    sigman(i,1)=0;
    F(i,1)=0;
end

j=2; i=1; processtop = 0; dt=0.0001; zonestop = 0; processkift = 0;

while processkift == 0 %næste procestrin while processtop == 0 %kraftligevægtskriterie while zonestop == 0 %antal ringelementer if j ~= 1 t(i,j)=t0+dt; %startantagelse af tykkelse, dt ændrer sig t(:,j)=t(i,j);

                if i ~=60                                                                              
                    r(i+1,j)=sqrt((2*(r(i,j)^2*((t(i,j)+t(i+1,j))*(1/2))-(r(i,1)^2-r(i+1,1)^2)*t(i,1)))/(t(i,j)+t(i+1,j)));
                end

                epsilonn(i,j)=log(t(i,j)/t0);
                epsilontheta(i,j)=log(r(i,j)/r(i,1));
                epsilonr(i,j)=-(epsilonn(i,j)+ epsilontheta(i,j));

                den(i,j)=epsilonn(i,j)-epsilonn(i,j-1);
                det(i,j)=epsilontheta(i,j)-epsilontheta(i,j-1);
                der(i,j)=-(den(i,j)+det(i,j));

                epsilons(i,j) = sqrt(R+1)*sqrt(R*(der(i,j)-det(i,j))^2+(der(i,j)-R*den(i,j))^2+(det(i,j)-R*den(i,j))^2)/(2*R+1);
                sigmas(i,j)=K*(epsilons(i,j))^n;

                k(i,j)=(sigmas(i,j)/epsilons(i,j))*((1+R)/(1+2*R))*(2*det(i,j)+den(i,j));
                %k(i,j) = sigmatheta(i,j) - sigmar(i,j)
                l(i,j)=(sigmas(i,j)/epsilons(i,j))*((1+R)/(1+2*R))*(den(i,j)*(1+R)+det(i,j));
                %l(i,j) = sigman(i,j) - sigmar(i,j)

                sigmar(1,j)=0;
                if i ~=60
                    if i ~=1
                        sigmar(i+1,j) = (sigmar(i, j)+(1/2*(k(i+1, j)/r(i+1, j)+k(i, j)/r(i, j))+my*(l(i+1, j)+sigman(i, j))/t(i, j))*(r(i+1, j)-r(i, j)))/(1-my*(r(i+1, j)-r(i, j))/t(i, j));
                    end
                end
            end

            if j==1 %et loop inden i et loop - kunne nok gøres smartere
                sigman(i,1)=0;
            else
                if i == 1
                    sigman(1,j)=l(1,j); %ved i=1 er sigmar = 0
                else
                    sigman(i,j)=l(i,j)+sigmar(i,j);
                end
            end
            if j ~=1
                if sigman(i,j)>=0
                    zonestop = 1
                else
                    i = i+1
                    zonestop = 0;
                end
            end
        end

        if zonestop >= 1
            Fanvendt = 3,9*10^4 %N
            for i=1:i-1 %alle ringelementer frem til radius r(i,j)
                A(i,j)=((r(i,j))^2-(r(i+1,j))^2)*pi;
                F(i,j)=sqrt((sigman(i,j))^2)*A(i,j); %i stedet for numerisk tegn
            end
            if sum(F(:,j))-Fanvendt > 0.01*Fanvendt
                dt = dt + 0.01;
                zonestop = 0; %geninitialiserer zonestop
                processtop = 0;
                i = 1; %starter forfra ved i = 1
            else
                processtop = 1;
            end
        end
    end

    if processtop >=1
        j=j+1;
        %restarter alt undt. j
        zonestop = 0;
        processtop = 0;
        i = 1;
        dt = 0;
    end

    if j>20
        processkift = 1 %terminerer alt
    end
end

[Walter Roberson]

http://www.mathworks.com/matlabcentral/answers/13205-tutorial-how-to-format-your-question-with-markup

[Jan]

The command "clear clc" removes the function of variable called "clc" from the memory. Are you sure you want to do this?!

Answer 1

[Walter Roberson]

There is no such thing as an "if loop". http://www.mathworks.com/help/techdoc/ref/if.html

Answer 2

[Jan]

I suggest to use the debugger to step through your code line by line.

This will most likely not produce the expected result:

Fanvendt = 3,9*10^4 %N

Now Fanvendt is 3 and "9e4" appears in the command window. You need a dot instead of the comma. Calculating 10^4 and the multiplication is much more expensive than using "3.9e4" directly.

Sahdowing the function "det" by a variable with the same name might have unexpected effects.

[Linda]

Arhh of course! But I don't understand what's wrong with the det-function?

[Jan]

"det" is a builtin function. You *can* shadow it by redefining it as a local variable. But this leads to problems frequently, as you can check in this forum. Example:

max = 1:5;

... Some other code, until you forget, that "max" was used before...

disp(max(rand(1, 100))); % ERROR!