Info

This question is closed. Reopen it to edit or answer.

idnlgrey model - solver computing other values than ode45

1 view (last 30 days)
Hi all,
I'm trying to do an parameter estimation of a model with three ODEs and six parameters with System Identification Toolbox. I'm using a nonlinear grey-box-model. I'm getting a fit of about 77%, but my state variables take complete unrealistic values. When solving the system with ode45 and my initial values computed per hand, the solution is complete different (but about how I expect). So the model should work.
When trying to find initial states with findstates(), MATLAB computes totally wrong values.
Any suggestions what I could do?
The equations are:
function [dx, y] = muehle2(t, x, u, k_c, k_pf, k_pc, k_ppf, k_mil, k_ppa, varargin) % Output equations.
%Outputs
y = x(3); %Druckverlust
%Stoffwerte etc.
%CM=0.04; %Feuchte der Kohle
%T_amb=30; %Umgebungstemperatur °C
%C_pc=1.3; %spez. Wärmekapazität Kohle kJ/kg*K
%C_pa=1.005; %spez. Wärmekapazität Luft kJ/kg*K
%H_evap=2257; %Verdampfungsenthalpie Wasser bei 100°C kJ/kg
%T_cin=20; %Temperatur Rohkohle °C
%Dichteberechnung
%rho=u(6)/(287.058*u(2));
%rho=1.5;
%Zustandsgleichungen
dx =[ u(1) * 1000/3600 - k_c * x(1); %Massenbilanz Rohkohle in Mühle
k_c * x(1) - k_pf * u(3) * x(2); %Massenbilanz gemahlene Kohle in Mühle
k_pc * x(1) + k_ppf * x(2) - k_mil * x(3) + k_ppa * u(3)]; %Druckverlust über Mühle
Thanks.

Answers (0)

Community Treasure Hunt

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

Start Hunting!