# Fitting 4 data sets to non-linear least squares

1 view (last 30 days)
jgd0008 on 1 Dec 2016
Edited: jgd0008 on 2 Dec 2016
Hello there,
Im trying to fit 4 data sets to an analytical model. Im looking for K1, where K1(u,v,r,theta). It gives me a "Exiting due to infeasibility: 1 lower bound exceeds the corresponding upper bound" error I´m pretty sure is because the 4 parameters as lsqcurvefit seems it works only with 2 sets of data, and I have 4. For data, you can use the attached txt file. Thanks in advance JGD0008
This is my code: *************
clear all; close all; clc; format long
E=200000; % MPa
nu=0.3; % poisson
global G k u v;
G=E/(2*(1+nu)); % shear mod
k=(3-nu)/(1+nu); %kolosov
A = xlsread('img55_DCT.xlsx'); %excel file with experimental data
x=A(:,1); y=A(:,2);
u=A(:,3);
v=A(:,4);
[m,n]=size(A);
r=zeros(m,1);
theta=zeros(m,1);
for i = 1:1:m %angle rotation according to Irving´s Ref System (X to left, Y down)
r(i,1)=sqrt(x(i,1)^2+y(i,1)^2);
if x(i,1) >= 0 && y(i,1)>=0 % I Q
theta(i,1)=acos(x(i,1)/r(i,1)); %IQ
end
if x(i,1) < 0 && y(i,1)>=0 % II Q
theta(i,1)=acos(x(i,1)/r(i,1)); %II Q
end
if x(i,1) < 0 && y(i,1)<0 % III Q
theta(i,1)=-acos(x(i,1)/r(i,1)); %-I Q
end
if x(i,1) > 0 && y(i,1)<0 % IV Q
theta(i,1)=-acos(x(i,1)/r(i,1)); %-II Q
end
end
% x0 = [1 25]; % initial guess K1 & K2
x0 = 30; % initial guess K1
% [K1, K2] = lsqcurvefit(@series,x0,r,theta); % two parameters
[K1, resnorm]= lsqcurvefit(@series,x0,r,theta, u, v);
****************
%function
function fseries = series(K1,r,theta)
global G k u v;
fseries=2*G*sqrt(2*pi./r)*(u./(sin(theta/2)*(k+1+2*(cos(theta/2))^2))-v./(cos(theta/2)*(k-1-2*(sin(theta/2))^2)))-K1(1);
***************
##### 2 CommentsShowHide 1 older comment
jgd0008 on 2 Dec 2016
thanks