Purpose
To obtain the state-space model (A,B,C,D) for the feedback inter-connection of two systems, each given in state-space form.Specification
SUBROUTINE AB05ND( OVER, N1, M1, P1, N2, ALPHA, A1, LDA1, B1,
$ LDB1, C1, LDC1, D1, LDD1, A2, LDA2, B2, LDB2,
$ C2, LDC2, D2, LDD2, N, A, LDA, B, LDB, C, LDC,
$ D, LDD, IWORK, DWORK, LDWORK, INFO )
C .. Scalar Arguments ..
CHARACTER OVER
INTEGER INFO, LDA, LDA1, LDA2, LDB, LDB1, LDB2, LDC,
$ LDC1, LDC2, LDD, LDD1, LDD2, LDWORK, M1, N, N1,
$ N2, P1
DOUBLE PRECISION ALPHA
C .. Array Arguments ..
INTEGER IWORK(*)
DOUBLE PRECISION A(LDA,*), A1(LDA1,*), A2(LDA2,*), B(LDB,*),
$ B1(LDB1,*), B2(LDB2,*), C(LDC,*), C1(LDC1,*),
$ C2(LDC2,*), D(LDD,*), D1(LDD1,*), D2(LDD2,*),
$ DWORK(*)
Arguments
Mode Parameters
OVER CHARACTER*1
Indicates whether the user wishes to overlap pairs of
arrays, as follows:
= 'N': Do not overlap;
= 'O': Overlap pairs of arrays: A1 and A, B1 and B,
C1 and C, and D1 and D, i.e. the same name is
effectively used for each pair (for all pairs)
in the routine call. In this case, setting
LDA1 = LDA, LDB1 = LDB, LDC1 = LDC, and LDD1 = LDD
will give maximum efficiency.
Input/Output Parameters
N1 (input) INTEGER
The number of state variables in the first system, i.e.
the order of the matrix A1. N1 >= 0.
M1 (input) INTEGER
The number of input variables for the first system and the
number of output variables from the second system.
M1 >= 0.
P1 (input) INTEGER
The number of output variables from the first system and
the number of input variables for the second system.
P1 >= 0.
N2 (input) INTEGER
The number of state variables in the second system, i.e.
the order of the matrix A2. N2 >= 0.
ALPHA (input) DOUBLE PRECISION
A coefficient multiplying the transfer-function matrix
(or the output equation) of the second system.
ALPHA = +1 corresponds to positive feedback, and
ALPHA = -1 corresponds to negative feedback.
A1 (input) DOUBLE PRECISION array, dimension (LDA1,N1)
The leading N1-by-N1 part of this array must contain the
state transition matrix A1 for the first system.
LDA1 INTEGER
The leading dimension of array A1. LDA1 >= MAX(1,N1).
B1 (input) DOUBLE PRECISION array, dimension (LDB1,M1)
The leading N1-by-M1 part of this array must contain the
input/state matrix B1 for the first system.
LDB1 INTEGER
The leading dimension of array B1. LDB1 >= MAX(1,N1).
C1 (input) DOUBLE PRECISION array, dimension (LDC1,N1)
The leading P1-by-N1 part of this array must contain the
state/output matrix C1 for the first system.
LDC1 INTEGER
The leading dimension of array C1.
LDC1 >= MAX(1,P1) if N1 > 0.
LDC1 >= 1 if N1 = 0.
D1 (input) DOUBLE PRECISION array, dimension (LDD1,M1)
The leading P1-by-M1 part of this array must contain the
input/output matrix D1 for the first system.
LDD1 INTEGER
The leading dimension of array D1. LDD1 >= MAX(1,P1).
A2 (input) DOUBLE PRECISION array, dimension (LDA2,N2)
The leading N2-by-N2 part of this array must contain the
state transition matrix A2 for the second system.
LDA2 INTEGER
The leading dimension of array A2. LDA2 >= MAX(1,N2).
B2 (input) DOUBLE PRECISION array, dimension (LDB2,P1)
The leading N2-by-P1 part of this array must contain the
input/state matrix B2 for the second system.
LDB2 INTEGER
The leading dimension of array B2. LDB2 >= MAX(1,N2).
C2 (input) DOUBLE PRECISION array, dimension (LDC2,N2)
The leading M1-by-N2 part of this array must contain the
state/output matrix C2 for the second system.
LDC2 INTEGER
The leading dimension of array C2.
LDC2 >= MAX(1,M1) if N2 > 0.
LDC2 >= 1 if N2 = 0.
D2 (input) DOUBLE PRECISION array, dimension (LDD2,P1)
The leading M1-by-P1 part of this array must contain the
input/output matrix D2 for the second system.
LDD2 INTEGER
The leading dimension of array D2. LDD2 >= MAX(1,M1).
N (output) INTEGER
The number of state variables (N1 + N2) in the connected
system, i.e. the order of the matrix A, the number of rows
of B and the number of columns of C.
A (output) DOUBLE PRECISION array, dimension (LDA,N1+N2)
The leading N-by-N part of this array contains the state
transition matrix A for the connected system.
The array A can overlap A1 if OVER = 'O'.
LDA INTEGER
The leading dimension of array A. LDA >= MAX(1,N1+N2).
B (output) DOUBLE PRECISION array, dimension (LDB,M1)
The leading N-by-M1 part of this array contains the
input/state matrix B for the connected system.
The array B can overlap B1 if OVER = 'O'.
LDB INTEGER
The leading dimension of array B. LDB >= MAX(1,N1+N2).
C (output) DOUBLE PRECISION array, dimension (LDC,N1+N2)
The leading P1-by-N part of this array contains the
state/output matrix C for the connected system.
The array C can overlap C1 if OVER = 'O'.
LDC INTEGER
The leading dimension of array C.
LDC >= MAX(1,P1) if N1+N2 > 0.
LDC >= 1 if N1+N2 = 0.
D (output) DOUBLE PRECISION array, dimension (LDD,M1)
The leading P1-by-M1 part of this array contains the
input/output matrix D for the connected system.
The array D can overlap D1 if OVER = 'O'.
LDD INTEGER
The leading dimension of array D. LDD >= MAX(1,P1).
Workspace
IWORK INTEGER array, dimension (P1)
DWORK DOUBLE PRECISION array, dimension (LDWORK)
LDWORK INTEGER
The length of the array DWORK. If OVER = 'N',
LDWORK >= MAX(1, P1*P1, M1*M1, N1*P1), and if OVER = 'O',
LDWORK >= MAX(1, N1*P1 + MAX( P1*P1, M1*M1, N1*P1) ),
if M1 <= N*N2;
LDWORK >= MAX(1, N1*P1 + MAX( P1*P1, M1*(M1+1), N1*P1) ),
if M1 > N*N2.
Error Indicator
INFO INTEGER
= 0: successful exit;
< 0: if INFO = -i, the i-th argument had an illegal
value.
> 0: if INFO = i, 1 <= i <= P1, the system is not
completely controllable. That is, the matrix
(I + ALPHA*D1*D2) is exactly singular (the element
U(i,i) of the upper triangular factor of LU
factorization is exactly zero), possibly due to
rounding errors.
Method
After feedback inter-connection of the two systems,
X1' = A1*X1 + B1*U1
Y1 = C1*X1 + D1*U1
X2' = A2*X2 + B2*U2
Y2 = C2*X2 + D2*U2
(where ' denotes differentiation with respect to time)
the following state-space model will be obtained:
X' = A*X + B*U
Y = C*X + D*U
where U = U1 + alpha*Y2, X = ( X1 ),
Y = Y1 = U2, ( X2 )
matrix A has the form
( A1 - alpha*B1*E12*D2*C1 - alpha*B1*E12*C2 ),
( B2*E21*C1 A2 - alpha*B2*E21*D1*C2 )
matrix B has the form
( B1*E12 ),
( B2*E21*D1 )
matrix C has the form
( E21*C1 - alpha*E21*D1*C2 ),
matrix D has the form
( E21*D1 ),
E21 = ( I + alpha*D1*D2 )-INVERSE and
E12 = ( I + alpha*D2*D1 )-INVERSE = I - alpha*D2*E21*D1.
Taking N1 = 0 and/or N2 = 0 on the routine call will solve the
constant plant and/or constant feedback cases.
References
NoneNumerical Aspects
NoneFurther Comments
NoneExample
Program Text
* AB05ND EXAMPLE PROGRAM TEXT
* Copyright (c) 2002-2010 NICONET e.V.
*
* .. Parameters ..
INTEGER NIN, NOUT
PARAMETER ( NIN = 5, NOUT = 6 )
INTEGER N1MAX, N2MAX, NMAX, M1MAX, P1MAX
PARAMETER ( N1MAX = 20, N2MAX = 20, NMAX = N1MAX+N2MAX,
$ M1MAX = 20, P1MAX = 20 )
INTEGER LDA, LDA1, LDA2, LDB, LDB1, LDB2, LDC, LDC1,
$ LDC2, LDD, LDD1, LDD2
PARAMETER ( LDA = NMAX, LDA1 = N1MAX, LDA2 = N2MAX,
$ LDB = NMAX, LDB1 = N1MAX, LDB2 = N2MAX,
$ LDC = P1MAX, LDC1 = P1MAX, LDC2 = M1MAX,
$ LDD = P1MAX, LDD1 = P1MAX, LDD2 = M1MAX )
INTEGER LDWORK
PARAMETER ( LDWORK = P1MAX*P1MAX )
DOUBLE PRECISION ONE
PARAMETER ( ONE=1.0D0 )
* .. Local Scalars ..
CHARACTER*1 OVER
INTEGER I, INFO, J, M1, N, N1, N2, P1
DOUBLE PRECISION ALPHA
* .. Local Arrays ..
INTEGER IWORK(P1MAX)
DOUBLE PRECISION A(LDA,NMAX), A1(LDA1,N1MAX), A2(LDA2,N2MAX),
$ B(LDB,M1MAX), B1(LDB1,M1MAX), B2(LDB2,P1MAX),
$ C(LDC,NMAX), C1(LDC1,N1MAX), C2(LDC2,N2MAX),
$ D(LDD,M1MAX), D1(LDD1,M1MAX), D2(LDD2,P1MAX),
$ DWORK(LDWORK)
* .. External Subroutines ..
EXTERNAL AB05ND
* .. Executable Statements ..
*
OVER = 'N'
ALPHA = ONE
WRITE ( NOUT, FMT = 99999 )
* Skip the heading in the data file and read the data.
READ ( NIN, FMT = '()' )
READ ( NIN, FMT = * ) N1, M1, P1, N2
IF ( N1.LE.0 .OR. N1.GT.N1MAX ) THEN
WRITE ( NOUT, FMT = 99992 ) N1
ELSE
READ ( NIN, FMT = * ) ( ( A1(I,J), J = 1,N1 ), I = 1,N1 )
IF ( M1.LE.0 .OR. M1.GT.M1MAX ) THEN
WRITE ( NOUT, FMT = 99991 ) M1
ELSE
READ ( NIN, FMT = * ) ( ( B1(I,J), I = 1,N1 ), J = 1,M1 )
IF ( P1.LE.0 .OR. P1.GT.P1MAX ) THEN
WRITE ( NOUT, FMT = 99990 ) P1
ELSE
READ ( NIN, FMT = * ) ( ( C1(I,J), J = 1,N1 ), I = 1,P1 )
READ ( NIN, FMT = * ) ( ( D1(I,J), J = 1,M1 ), I = 1,P1 )
IF ( N2.LE.0 .OR. N2.GT.N2MAX ) THEN
WRITE ( NOUT, FMT = 99989 ) N2
ELSE
READ ( NIN, FMT = * )
$ ( ( A2(I,J), J = 1,N2 ), I = 1,N2 )
READ ( NIN, FMT = * )
$ ( ( B2(I,J), I = 1,N2 ), J = 1,P1 )
READ ( NIN, FMT = * )
$ ( ( C2(I,J), J = 1,N2 ), I = 1,M1 )
READ ( NIN, FMT = * )
$ ( ( D2(I,J), J = 1,P1 ), I = 1,M1 )
* Find the state-space model (A,B,C,D).
CALL AB05ND( OVER, N1, M1, P1, N2, ALPHA, A1, LDA1,
$ B1, LDB1, C1, LDC1, D1, LDD1, A2, LDA2,
$ B2, LDB2, C2, LDC2, D2, LDD2, N, A, LDA,
$ B, LDB, C, LDC, D, LDD, IWORK, DWORK,
$ LDWORK, INFO )
*
IF ( INFO.NE.0 ) THEN
WRITE ( NOUT, FMT = 99998 ) INFO
ELSE
WRITE ( NOUT, FMT = 99997 )
DO 20 I = 1, N
WRITE ( NOUT, FMT = 99996 ) ( A(I,J), J = 1,N )
20 CONTINUE
WRITE ( NOUT, FMT = 99995 )
DO 40 I = 1, N
WRITE ( NOUT, FMT = 99996 ) ( B(I,J), J = 1,M1 )
40 CONTINUE
WRITE ( NOUT, FMT = 99994 )
DO 60 I = 1, P1
WRITE ( NOUT, FMT = 99996 ) ( C(I,J), J = 1,N )
60 CONTINUE
WRITE ( NOUT, FMT = 99993 )
DO 80 I = 1, P1
WRITE ( NOUT, FMT = 99996 ) ( D(I,J), J = 1,M1 )
80 CONTINUE
END IF
END IF
END IF
END IF
END IF
STOP
*
99999 FORMAT (' AB05ND EXAMPLE PROGRAM RESULTS',/1X)
99998 FORMAT (' INFO on exit from AB05ND = ',I2)
99997 FORMAT (' The state transition matrix of the connected system is')
99996 FORMAT (20(1X,F8.4))
99995 FORMAT (/' The input/state matrix of the connected system is ')
99994 FORMAT (/' The state/output matrix of the connected system is ')
99993 FORMAT (/' The input/output matrix of the connected system is ')
99992 FORMAT (/' N1 is out of range.',/' N1 = ',I5)
99991 FORMAT (/' M1 is out of range.',/' M1 = ',I5)
99990 FORMAT (/' P1 is out of range.',/' P1 = ',I5)
99989 FORMAT (/' N2 is out of range.',/' N2 = ',I5)
END
Program Data
AB05ND EXAMPLE PROGRAM DATA 3 2 2 3 1.0 0.0 -1.0 0.0 -1.0 1.0 1.0 1.0 2.0 1.0 1.0 0.0 2.0 0.0 1.0 3.0 -2.0 1.0 0.0 1.0 0.0 1.0 0.0 0.0 1.0 -3.0 0.0 0.0 1.0 0.0 1.0 0.0 -1.0 2.0 0.0 -1.0 0.0 1.0 0.0 2.0 1.0 1.0 0.0 1.0 1.0 -1.0 1.0 1.0 0.0 1.0Program Results
AB05ND EXAMPLE PROGRAM RESULTS The state transition matrix of the connected system is -0.5000 -0.2500 -1.5000 -1.2500 -1.2500 0.7500 -1.5000 -0.2500 0.5000 -0.2500 -0.2500 -0.2500 1.0000 0.5000 2.0000 -0.5000 -0.5000 0.5000 0.0000 0.5000 0.0000 -3.5000 -0.5000 0.5000 -1.5000 1.2500 -0.5000 1.2500 0.2500 1.2500 0.0000 1.0000 0.0000 -1.0000 -2.0000 3.0000 The input/state matrix of the connected system is 0.5000 0.7500 0.5000 -0.2500 0.0000 0.5000 0.0000 0.5000 -0.5000 0.2500 0.0000 1.0000 The state/output matrix of the connected system is 1.5000 -1.2500 0.5000 -0.2500 -0.2500 -0.2500 0.0000 0.5000 0.0000 -0.5000 -0.5000 0.5000 The input/output matrix of the connected system is 0.5000 -0.2500 0.0000 0.5000
Click here to get a compressed (gzip) tar file containing the source code of the routine, the example program, data, documentation, and related files.
Return to index