Two variable function computation

2 views (last 30 days)
friet
friet on 31 Jan 2017
Answered: Walter Roberson on 31 Jan 2017
Hello,
I am trying to reproduce the Matlab plot I found in a research paper. The function in P(t,z) (it is time and space dependent). P(t,z) have two parameters (alpha and beta) which are frequency dependent.
I have two loops for computing P(t,z). My code is below, however, I can't find a similar solution which what is mentioned. This is my code
clc
clear all
close all
f=linspace(0,20*10^6,100);
t=1./f;
x=linspace(0,0.15*10^-2,100);
alpha=(0.45*10^-11)*f.^2;
beta=(0.15*10^-4)*f;
figure(1)
subplot(2,1,1)
plot(f/10^6,alpha/100)
ylabel('alpha')
subplot(2,1,2)
plot(f/10^6,beta/100)
ylabel('beta')
p = zeros(length(x),length(t));
for it= 1:length(t)
for ix=1:length(x)
p(ix,it)= exp(-alpha(it).*x(ix)).*exp(-i.*1.*(f(it).*t(it)-beta(it).*t(it)));
end
end
figure(2)
plot3(x,t,p)
Any help will be appriciated.
Thanks
  2 Comments
Walter Roberson
Walter Roberson on 31 Jan 2017
Note that your p is complex, so you should abs() before plotting the magnitude.
You are changing alpha and omega and beta over time, which is not part of the formula. That p formula is for some particular alpha and beta and omega.
Your definition of t is non-linear, but the plot on the right appears to use linear time.
In the right hand plot, the Distance does not appear to have a scaling factor, but you have used 10^-2 scaling factor in your definition of x.
friet
friet on 31 Jan 2017
Edited: friet on 31 Jan 2017
Hello,
Thanks for your comment. I have made some modification based on your kind comment. I have selected omega as 10MHz and select the corresponding values of alpha and beta. The factor 10^-2 is change from cm to m. All the units are in for x are in m and time in sec.
clc
clear all
close all
f=linspace(0,20*10^6,100);
t=linspace(0,1*10^-6,100);
x=linspace(0,0.15*10^-2,100);
alpha=(0.45*10^-11)*f.^2;
beta=(0.15*10^-4)*f;
omega=10*10^6;
a=5.5*100;
b=1.8*100;
figure(1)
subplot(2,1,1)
plot(f/10^6,alpha/100)
ylabel('alpha')
subplot(2,1,2)
plot(f/10^6,beta/100)
ylabel('beta')
p = zeros(length(x),length(t));
for it= 1:length(t)
for ix=1:length(x)
p(ix,it)= exp(-a.*x(ix)).*exp(-i.*1.*(omega.*t(it) - b.*t(it)));
end
end
tt= repmat(t,100,1);
xx= repmat(x,100,1);
figure(2)
waterfall(xx,tt,abs(p))
But still, can't find the same solution. By the way, how can I change the origins of the 3d plot, the distance and time to be start at 0 and the magnitude origin separately as it was plotted in the original plot.
Thanks

Sign in to comment.

Answers (2)

Honglei Chen
Honglei Chen on 31 Jan 2017
You may want to check out waterfall
HTH

Walter Roberson
Walter Roberson on 31 Jan 2017
Your expression just does not act like that, at least not over those parameter ranges.
See attached that explores a number of possibilities. (Note the vectorized implementations of the equations.) I put up several isosurface plots to explore the way the surface grows along the 4 different parameters. The last couple of plots explore the possibility that your x is too small or too large. The last one does suggest that possibly the x range should be much smaller; for example if you were to use x/200 then it would have room for only one peak.
On the isosurface plots notice that I plot the real component of p rather than the absolute magnitude. The absolute magnitude are completely boring; plotting the real component at least allows you to track through a couple of phase cycles.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!