pelda4.m

Incorporating Prior Knowledge in Cubic Spline Approximation - Application to the Identification of R

clear all
global kin te ya yb yc

%t1 = 89.5;  %az elso bemeres idotartama (eddig semmi sem tortenik, csak bomlik az anyag)
%t2 = 149.7; %a masodik bememeres idotartam (ekkor indul be igazan a reakcio)

load rc1_403.txt
t = rc1_403(:,1);  %ido (perc)
M = rc1_403(:,2);  %osszes tomeg
Mv = rc1_403(:,3); %vinileszter tomeg
Mk = rc1_403(:,4); %kloroform tomeg
Mp = rc1_403(:,5); %proxibrat tomeg
M3 = Mv+Mk+Mp; %a harom anyag tomege egyutt
tt = t(2:end)-t(2);
ca = Mv(2:end);
cb = Mk(2:end);
cc = Mp(2:end);
ca0 = Mv(2);
cb0 = Mk(2);
cc0 = Mp(2);
te = [0:1:250]'; %rajzolasi pontok
tk = [0:1:250]'; %soft-korlatok pontjai az iterativban es az fk3-ban
%a csompontokat a beadagolas jeloli ki, es min. 3 db pont legyen egy tartomanyban
knots = [87.7 145 168 250+87.7]-87.7;

figure(1)
hold on
plot(tt,ca,'k.');
plot(tt,cb,'k.');
plot(tt,cc,'k.');

%-------------------------------
%Spline illesztes

spline_kezd
ya = drawspline(spa1,te,'k--');
yb = drawspline(spb1,te,'k--');
yc = drawspline(spc1,te,'k--');

spline_kemeny
y = [ya;yb;yc];
yold = zeros(size(y));
while mean(abs(y-yold))/mean(abs(y))>1E-3,
 pause(0.1);
 mean(abs(y-yold))/mean(abs(y))
 yold = y;
 %k1,k2,k3 becslese a gorbekbol
 kin = fk3(spa1,spb1,spc1,tk,ca0+cb0+cc0);
 %Iterativ becsles
 lambda1 = 1;
 spline_iter
 %uj y
 ya = drawspline(spa1,te);
 yb = drawspline(spb1,te);
 yc = drawspline(spc1,te);
 y = [ya;yb;yc];
end