benubah / control Goto Github PK

I got this error when trying to convert the model back to TF with these statements:

g1 <- tf(10,c(1,6,5,0))
s1 <- tf2ss(g1)
ss2tf(s1)

y1:Error in poly2str(argnum, svar = "s", smul = " ") :
Argument 'p' must be a numeric vector.

Using the step function to plot the step response of a transfer function gives this error message.

Error

Error in balance(baP$z, "S") :
BLAS/LAPACK routine 'DGEBAL' gave error code -3

Im not an expert but is it

> TF("exp(s)")

 y1:
     2.718282 s^1 + 1 
   - - - - - - - - - - -
                 1 

result y1 right ? Shouldn't be exp(s) ?

The current care() function requires package QT which is now unavailable.

However, commenting out line 69 as follows:
# tmp <- QZ::qz.zgees(val) # schur decomposition from QZ

and un-commenting line 68 as follows:
tmp <- Matrix::Schur(val) # coerces imaginary parts

solves the problem and gives the same results as Octave function lqr(), but not Octave care(), on a sample problem.

Let me know if anything unclear.

Regards,

Graham W Griffiths
[email protected]

Both TF("-1/s") and TF("1/(-s)") give error (R 3.6.1)

Error in args[[2]] : subscript out of bounds

The R transfer functions generated from a ss model using the control package, differ from the Octave transfer functions. The Octave results seem to be correct.
jetEngine_tf_contr_2.zip

Hi

Thanks for your work in Control package. It seems the results of place() example cannot be reproduced by Matlab place function.

Here is the code of place example:

F <- rbind(c(0,1),c(0,0))
G <- rbind(0,1)
H <- cbind(1,0);
J <- 0
t <- 1
sys  <-  ss(F,G, H,J)
A  <-  c2d(sys,t);
j <- sqrt(as.complex(-1));
pc  <- rbind(0.78+0.18*j, 0.78-0.18*j)
K  <-  place(A$A, A$B, pc)

And the results just like the follows:

> K
        [,1]
[1,] -0.0808
[2,] -0.3188

But in Matlab, the results are different:

F = [0,1;0,0];
G = [0;1];
H = [1,0];
J = [0];
t = 1;
sys = ss(F,G,H,J);
sysd = c2d(sys,t);
pc = [0.78+0.18*sqrt(-1);0.78-0.18*sqrt(-1)];
place(sysd.A,sysd.B,pc)'

ans =

    0.0808
    0.3996

I really want to know where the problem lies since I am a green hand in control filed . I would greatly appreciate any suggestions.

Regards

JY

On running an Octave example from:

https://wiki.octave.org/Control_package#Examples

the R version generated some errors. I attach a zipped set of R/Octave code that reproduces the errors, together with a pdf of the R code with the error lines marked in yellow.

Let me know if anything is unclear.

Regards,
Graham W griffiths
[email protected]
OctaveEx_markup.zip