Array indices must be positive integers or logical values.

I don't know what I'm doing wrong! I'd appreciate any help! I'm using MATLAB 2019b. It keeps giving me this error

Array indices must be positive integers or logical values.

Here is my code:

  n = 32;  	    % choose a number which is a exponent of 2
    Cp = 0.24; % cal/g C
    K = 0.007; % cal/(cm s C)
    rho = 2.7; % g/cm3
    kapp = K/(Cp*rho); % cm2/s
    d = 5*100; % cm
    start_temp = 1000;
    t = 1*24*60*60;  % (days)

    T=zeros(1:numel(n));
    
    for j = 1:n/2
       T(j) = start_temp;
    end
    
    for j = n/2+1:n
       T(j) = -start_temp;
    end
    
    figure(1);
    
    x = linspace(0,d,length(T));
    plot(x,T,'^')
    hold on
    % plot(T,n)
    xlabel('distance (units)')
    ylabel('Temperture (units)')
    % Title('Heated Dike, Lab #3, #...days')
    
    %%%%%% TAKE FAST FOURIER TRANSFORM (FFT) of your Temp function %%%%
    % (You are now working in frequency (or in this case - wave number) space
    % (This will output a coefficient in alternating sine,cos for each point) 
    
    %%%%%% here is my function
    
%             function f = dfft(y)
%             % First use standard Matlab routine to find Fourier transform of y.
%             % z contains the complex coefficeints of the Fourier exponential series.
%              z=fft(y); 
%              n = length(y);
%              half=n/2;
%             % This section takes the exponential series coefficients and gives the     
%             % coefficients of the Fourier Sine and Cosine series.
%             % f(1) = constant term = average value of the function over period sampled
%             % f(2) = 
%              for i = 1:2:n
%               j=(i+1)/2;
%               f(i)=real(z(j))/half;
%               f(i+1)=-imag(z(j))/half;
%              end  
%              j=n/2 +1;
%              f(2)=real(z(j))/half; 

   
    y=dfft(T);
    
    %%%%% GET FOURIER COEFFICIENTS (Bo, An, Bn) %%%%%%
    
    COEFF(1) = y(1)/2;
    COEFF(2) = y(2)/2;
    for j = 3:2:n		% increment from 3 to n by 2's
       COEFF(j) = y(j);
       COEFF(j+1) = y(j+1);
    end
    
    %%%%%%  SOLVE FOR TEMPERATURE CHANGE using solution to heatflow eqn %%%%%
    %   Y = exp(1i*pi)
    % 
    for j = 3:2:n
       NEWT(j) = y(j).*exp(-kapp*t(2*pi*j/(2*d))^2)*(sin((pi*j)));
       NEWT(j+1) = y(j+1).*exp(-kapp*t(2*pi*j/(2*d))^2)*(sin((pi*j)));
    end
    
    NEWT = T(2)*exp(-kapp*t(2*pi*j/(2*d))^2)*(sin((pi*j)));

[Chad Greene]

This'll be pretty easy to diagnose, if you can tell us where (which line) the error is occurring. Is the error message giving you that info?

[cruxsign]

here is the line it gives the error

NEWT(j) = y(j).*exp(-kapp*t(2*pi*j/(2*d))^2)*(sin((pi*j)));

Answer 1

[Chad Greene]

I think this is the error

t(2*pi*j/(2*d))

The first time through the loop, j=3, which means

2*pi*j/(2*d) = 0.02

It's trying to index the 0.02'th entry in t, but t is a scalar. Perhaps, did you mean to multiply t by the parenethesis? Like this:

t*(2*pi*j/(2*d))

[cruxsign]

Yes! Thank you! It is supposed to be multiplied by the parenthesis. It's fixed!

[Chad Greene]

Great, glad to help. Can you mark this question as answered?

[cruxsign]

Thanks again! Sure!