How to use fminsearch for my code

Hello everyone, I am new in MATLAB. Please help me find out my problem.

I prepared a code that aimed at finding parameters to fit the data. (data is recorded temperature vs. time)

it is work by changing that three parameters manually but I would like to use fminsearch but always faild!

Thanks in advance.

here is my code (the objective is RMSE):

close all, clear all,  clc
load GHE1_6kW_P12_V1.txt
[min,n]=size(GHE1_6kW_P12_V1);
tau = GHE1_6kW_P12_V1(240:end,1);
Tfm_mis = GHE1_6kW_P12_V1(240:end,2);

rho_s = 2533;                       %[kg/mc]
c_s = 840;                         %[J/kg.K]
eps = 0.4;
T_0 = 15.96;
rho_w = 1020;                       %water density [kg/mc] 
c_w = 4186;
r_bh= 0.076;                        %borehole ID/OD=151/165
q = 110.22;                            %[W/m]
a_s = 0.25;
par=[10,1.7E-6,0.09];
C_0 = rho_s*c_s*(1-eps)+rho_w*c_w*eps; 

est = r_bh^2*C_0/4*tau*par(1);
f=(1./est).*exp((-1./est)-((par(2)^2)*c_w^2*r_bh^2*rho_w^2)/16*par(1));
T=cumtrapz(est,f);

dteta= 45*pi/180;
teta =(dteta : dteta : 2*pi);
for j=1:numel(teta)
    z(1,j)= q/(4*pi*par(1))*exp((par(2)*rho_w*c_w*r_bh*cos(teta(j)))/2*par(1));
end

g=T*z;
A=mean(g');
B = T_0+A;
Tfm_MLS= B + (q * par(3));

plot(Tfm_MLS,'r')
hold on
plot (Tfm_mis,'g')

RMSE = (sum((Tfm_MLS-Tfm_mis).^2)/length(tau)).^0.5;

[Walter Roberson]

Which variables need to be optimized over? Is it the par vector?

Is RMSE the equation whose value is to be minimized? \

Is there a reason you are using a loop for

for j=1:numel(teta)
    z(1,j)= q/(4*pi*par(1))*exp((par(2)*rho_w*c_w*r_bh*cos(teta(j)))/2*par(1));
end

instead of

z = q./(4*pi*par(1)).*exp((par(2).*rho_w.*c_w.*r_bh.*cos(teta))./2.*par(1));

which is vectorized?

Answer 1

[Saeid Bina]

Yes Par need to be optimaized and while RMSE is minimized. RMSE is the diffrence between calculated and measured data.

I have used loop beacuase that fuction need to be calculated for 8 teta. It is kind a 3D matrix. The data are calculating for timly basis and at each min for 8 teta (degree) and then using Mean for finding average temperature for each min.

by the way if you think is not needed, your idea is respected.

[Walter Roberson]

Vectorizing that loop that calculates z would be more efficient.

Make sure, though, that you use ./ and .* instead of / and * like I show in my code.

[Saeid Bina]

Yes, I have used that one, it results in same. Please help me to optimaize par.