Estaimation of kinetic papremeters using optimization

1 view (last 30 days)
Hi,
I am trying to solve the pde system(stiff) for my kinetics experiments. We have around 100 experimental data points.
unknowns are: {k1,E1, k2,E2} are positive values
r=(k1*exp(-E1/T)*c2+k2*exp(-E2/T)*c5)c1*c2
equations are :
(1) dc1/dt=D*d^2(c1)/dx^2-r;
(2) dc2/dt=D*d^2(c2)/dx^2-2*r;
(3) dc3/dt=D*d^2(c3)/dx^2+r;
(4) dc4/dt=D*d^2(c4)/dx^2+r;
(5) dc5/dt=D*d^2(c5)/dx^2;
and i am using ode15s and collocation method for solving the equations. After getting the solution i need find the flux and compare with experimental flux values.
I need to minimize the error between experimental and modelled flux.
I tried to fminsearch but most of the times my objective function is not zero.
objective function=(flux_exp_flux_model)^2/(flux_exp*flux_model)
I need to mention that for one complete run of all data points it takes around 300 seconds.
So pelase suggest me how can i get the optimized values.
  1 Comment
Matt J
Matt J on 25 Mar 2014
I tried to fminsearch but most of the times my objective function is not zero.
One doesn't expect a numerical minimization routine to reach zero exactly. Is there other evidence you see of a poor fit?

Sign in to comment.

Answers (0)

Products

Community Treasure Hunt

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

Start Hunting!