Errors occur during grey-box modeling

Currently I am trying to build up a thermal model of a lecture theater using grey-box modeling in system identification toolbox. First I specified the model structure in a function file whose content is shown below:

function [A, B, C, D, K, x0] = thermalmdl_ode(R1, R2, R3, R4, C1, C2, Ts) % This function parameterizes the thermal model of the lecture theatre

% R1 is the convective heat transfer resistance of the outside surface of

% the wall and roof

% R2 is half of the conductive heat transfer resistance of the wall and

% roof

% R3 is the conductive heat transfer resistance of the window

% R4 is the convective heat transfer resistance of the inside surface of

% the wall and roof

% C1 is the heat capacity of the wall and roof

% C2 is the heat capacity of the lecture theatre

% Define some useful variables a=1/R1+1/R2+1/R3; b=1/R2+1/R3+1/R4; c=R3^2*a*b-1;

% A matrix

A=zeros(2,2);

A(1, 1)=(1-b*R4)/(C2*b*R4^2);

A(1, 2)=(1+R3*a)*R3/(C2*R2*R4*c);

A(2, 1)=(c+b)/(b*c*R4*C1);

A(2, 2)=(2*R3+R3^2*(a+b)-2*c)/(R2*C1*c);

% B matrix

B=zeros(2,2);

B(1, 1)=R3/(R1*R4*c);

B(1, 2)=R3/(R4*C2*c);

B(2, 1)=R3*(b*R3-1)/(R1*R2*C1*c);

B(2, 2)=R3*(1+b*R3)/(C1*c);

% C matrix

C=[1 0];

% D matrix

D=zeros(2, 2);

% K matrix

K=[-1/C2; 0];

% Initial state

x0=[298.15; 298.15];

end

The input and state of my model are both a 2*1 matrices, the output is a scalar.

Then I tried to create a grey-box model object with idgrey command, but errors occurred:

thermalmdl=idgrey('thermalmdl_ode', {0.000133; 0.000114; 0.0000212;0.000133; 638645.04; 4456.38}, 'cd', {}, 60)

Error using pmodel.LinearODE/checkOutputArgConsistency (line 214)

The sizes of the matrices returned by the ODE function must be consistent with the input/output size and order of the model. Create a new model if you need to change the sizes.

Error in pmodel.LinearODE/checkDataType (line 195)

M = checkOutputArgConsistency(M, Ts, Nargout);

Error in idgrey (line 376)

Structure = checkDataType(pmodel.LinearODE(Orders, Fcn, FcnType, Par, Extras), Ts);

Anyone has any ideas what is going on here?

Answer 1

[Rajiv Singh]

D matrix is 2-by-2 suggesting that the number of inputs (nu) = number of outputs (ny) = 2. However, C is 1-by-2 implying ny = 1. Also K is 2-by-1 implying ny=1. It seems likely that your C and K equations need to be fixed.

[Haoquan]

Yeah, I have already spotted that mistake. But thanks for your response, Rajiv!