mpc_predmat.m

Model-based predictive control: a practical approach此程序为模型预测控制的例程

%%%%  Generate prediction equations (using recursions) from 
%%%%  MFD model   A y(k+1) = B u(k)
%%%%
%%%%  yfut = H *Dufut + P*Dupast + Q*ypast
%%%%
%%%%  [H,P,Q] = mpc_predmat(A,B,ny)
%%%%
%%%%  ny is the output horizon
%%  
%% Author: J.A. Rossiter  (email: [e-mail])

function [H,P,Q] = mpc_predmat(A,B,ny);


%%%%% Add integral to model
sizey = size(A,1);
D = [eye(sizey),-eye(sizey)];
AD = convmat(A,D);
nA = size(AD,2);
nB = size(B,2);

%%%%  Initialise recursion data
%%%% nominal model    y =  Bo ut + B2 Dupast  + A2 ypast
A2 = -AD(1:sizey,sizey+1:nA);
B2 = B(1:sizey,sizey+1:nB);
Bo = B(1:sizey,1:sizey);
nB2 = nB-sizey;
nA2 = nA-sizey;

P1=Bo;
P2=B2;
P3=A2;
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

%%%%% Loop updating models using recursion
for i=2:ny;
   
   vecold = (i-2)*sizey+1:(i-1)*sizey;
   vecnew = (i-1)*sizey+1:i*sizey;
   Phi = P3(vecold,1:sizey); 

   vecufut = 1:sizey*i;
   vecufut2 = 1:sizey*(i-1);
   P1(vecnew,vecufut) = [(P2(vecold,1:sizey)+Phi*Bo),P1(vecold,vecufut2)];
   
   vecupast = sizey+1:nB2;
   vecypast = sizey+1:nA2;
   
   temp = [P2(vecold,vecupast),zeros(sizey,sizey)] + Phi*B2;
   P2(vecnew,1:nB2) = temp;
   P3(vecnew,1:nA2) = [P3(vecold,vecypast),zeros(sizey,sizey)] + Phi*A2;
end

H=P1; P=P2; Q=P3;


%    function [C]=convmat(A,B)
%    to convolve two matrix polynomials stored as [a(0),a(1),a(2)...]
%   routine works by forming a convolution matrix [a(0),0   ,0   ,...
%                                                a(1),a(0),0   ,...
%                                                a(2),a(1),a(0),...]
function [C]=convmat(A,B)
[na,ma]=size(A);[nb,mb]=size(B);orda=ma/nb;ordb=mb/na;
mat = [];
for i=1:orda;
    s = zeros(nb,((orda-1)*na+mb));
    s(:,(i-1)*na+1:(i-1)*na+mb) = B;
    mat = [mat;s];
end;
C=A*mat;