Models "Opportunities and difficulties with 5x5 distillation control"

From: Petter Lundstrøm (petter.lundstrom@energos.no)
Date: Mon Jan 29 2001 - 09:31:02 CET


Dear Professor Kalid,

I am happy to hear that you want to use the article "Opportunities and
difficulties with 5x5 distillation control" in your undergraduate and
graduate courses of chemical engineering at the Universidade Federal da
Bahia.

The appended file "linemcmodel.mat" contains two matrices "g" and "g15"
on Matlab format.
"g" contains the linearised state space model (A,B,C,D matrices) before
model reduction, i.e. the model has 123 states.
"g15" contains the linearised state space model after model reduction to
15 states, using "ohkapp" in Robust Toolbox (se section "Controllability
analysis" in the article).

 <<linemcmodel.mat>>

Both "g" and "g15" are stored on "mu-tools format", i.e.

 A(1,1) A(1,2) ... A(1,nx) B(1,1) ... B(1,nu) nx
 A(2,1) A(2,2) ... A(2,nx) B(2,1) ... B(2,nu) 0
   : : : : : :
 A(nx,1) A(nx,2) ... A(nx,nx) B(nx,1) ... B(nx,nu) 0
 C(1,1) C(1,2) ... C(1,nx) D(1,1) ... D(1,nu) 0
 C(2,1) C(2,2) ... C(2,nx) D(2,1) ... D(2,nu) 0
   : : : : : :
 C(ny,1) C(ny,2) ... C(ny,nx) D(ny,1) ... D(ny,nu) 0
   0 0 ... 0 0 ... 0 -Inf

where nx = number of states
      nu = number of inputs
      ny = number of outputs

In both cases the inputs and outputs are:
        Inputs: L_T, Q_R, Q_C, D, B, F, z_F, P_F i.e. 8
inputs
        Outputs: x_D(1), x_B(1), P_D, M_D, M_B i.e. 5
outputs

Scaling of inputs and outputs, as given in Table 3 of the article, are:
% Scaling (Table 3)
inscale = diag([2.7 130 130 0.5 0.5 0.15 0.1 0.025]);
outscale = diag([1/0.01 1/0.01 1/0.05 1/30 1/10]);
gsc = mmult(outscale,g,inscale);

If you have any questions, please let me know.

Best regards,
Petter Lundstrøm

Dr. Ing. Petter Lundstrøm Tlf: +47 72 89 61 00/30
Energos asa Faks: +47 72 89 61 01
Vestre Rosten 81 Mobil: +47 926 22 232
N-7075 Tiller E-post: petter.lundstrom@energos.no
Norge

        -----Opprinnelig melding-----
        Fra: skoge@chembio.ntnu.no [SMTP:skoge@chembio.ntnu.no]
        Sendt: Friday, December 22, 2000 8:40 AM
        Til: Petter Lundstrøm
        Emne: Fra Brazil

>From kalid@ufba.br Thu Dec 21 17:53:19 2000

        Dear Professor Sigurd Skogestad=20

        I am assistant professor of the Universidade Federal da Bahia
(Brazil) =
        and would like to use the article "Opportunities and
difficulties with =
        5x5 distillation control" in my undergraduate and graduate
courses of =
        chemical engineering. But I need the transfer functions of this
process. =
        Can you help me to find or to give me this model?=20

        Thank you for your attention

        Ricardo de Ara=FAjo Kalid
        __________________________
        kalid@ufba.br



This archive was generated by hypermail 2.1.3 : Fri Sep 05 2003 - 16:42:15 CEST