Problems with using a loop structure within an fzero command.

I have a file which uses the fzero command and works just fine as it gives me values of x when Re is chosen. See below:

F = @(f,Re) 2*log10(Re*sqrt(f))-0.8-(1/sqrt(f));  % parameterized function
Re = 1e7;                      % parameter
fun = @(f) F(f,Re);                               % function of x alone
x = fzero(fun,.05)

I am trying to use several scalar values for Re via a looping structure, but I am struggling to understand how to use it properly. My attempt is below:

   F = @(f,Re) 2*log10(Re*sqrt(f))-0.8-(1/sqrt(f));  % parameterized function
  Re = linspace(1e4,1e7,31);                      % parameter
  fCW = zeros(size(Re));
  for k = 1:length(Re)
      fCW(k) = fzero(@(f) F(Re(k), f), 0.05);
  end

[Yianni]

I am trying to mimic a script like this:

f = @(x,y) y.^3 + y - x;
% 50 points equally spaced from -5 to 2, inclusive
xvals = linspace(-5, 2, 50);
% Preallocate storage space for the solutions
yvals = zeros(size(xvals));
% Loop through all the values
for j = 1:length(xvals)
 yvals(j) = fzero(@(y) f(xvals(j), y), 0);
end

[per isakson]

Have you tried to plot F(f) for a few values of Re? See&nbsp ezplot

[Yianni]

Individually it works just fine, but the looping doesn't yield anything on a plot.

Answer 1

[per isakson]

First compare the order of the input arguments in

    F = @(f,Re) 2*log10(Re*sqrt(f))-0.8-(1/sqrt(f));

and in

    fCW(k) = fzero( @(f) F( Re(k), f ), 0.05 );

then replace

    fCW(k) = fzero( @(f) F( Re(k), f ), 0.05 );

by

    fCW(k) = fzero( @(f) F( f, Re(k) ), 0.05 );

and the loop will return a vector of numbers

    >> fCW = cssm()
    fCW =
        0.0309    0.0104    0.0093    0.0087    0.0084    0.0081

where

    function    fCW = cssm()
        F = @(f,Re) 2*log10(Re*sqrt(f))-0.8-(1/sqrt(f));  
        Re = linspace(1e4,1e7,6);                         
        fCW = zeros(size(Re));
        for k = 1:length(Re)
            fCW(k) = fzero( @(f) F(f,Re(k)), 0.05 );
        end
    end

&nbsp

In response to comment

    [fCW,fB,fSJ,Re] = cssm()
    figure, semilogy( fCW, Re, 'd' )

where

    function    [fCW,fB,fSJ,Re] = cssm()
        F = @(f,Re) 2*log10(Re*sqrt(f))-0.8-(1/sqrt(f));  
        Re = logspace( 4,7,31 );                         
        fCW = zeros(size(Re));
        fB  = zeros(size(Re));
        fSJ = zeros(size(Re));
        for k = 1:length(Re)
            fCW(k) = fzero( @(f) F(f,Re(k)), 0.05);
            fB(k)  = 0.3164/Re(k)^0.25;
            fSJ(k) = 0.25/(log(5.74/Re(k)^0.9))^2;
        end
    end

[Yianni]

Excellent this works great!

Now is there a way to incorporate two more equations in here?

fB(:,k)=0.3164/Re(k)^0.25;
fSJ(:,k)=0.25/(log(5.74/Re(k)^0.9))^2;

These do not use fzero and I want to be able to plot these three equations as Re vs. (fCW,fB, fSJ)

[per isakson]

Did you try? Why

    fB(:,k)

rather than

    fB(k)

Right hand side

    0.3164/Re(k)^0.25

returns a scalar