Normalizing a transfer function

Specific Domains Modelling & Simulations
, ,
1

Is there already a function to normalize a transfer function?
I have:

TransferFunction{Continuous, ControlSystemsBase.SisoRational{Float64}}
                                 -951.729934283765s^2
---------------------------------------------------------------------------------------
3.1777394243568e7s^3 + 7.907227012152197e6s^2 + 540639.1480777722s + 5918.8473721761775

and I would like all coefficients to be divided by -951.729934283765.

2

OK, wrote a function that does this myself:

function normalize(tf_::TransferFunction{Continuous, ControlSystemsBase.SisoRational{Float64}})
    expr = tf_.matrix[1]
    factor = expr.num.coeffs[end]
    if factor == 0.0
        return tf_
    end
    tf(reverse(expr.num.coeffs./factor), reverse(expr.den.coeffs./factor))
end
3

You are reaching deep into the internals of ControlSystems here. See the functions numvec and denvec which would make your implementation simpler and avoid using internals.

No, because there are many possible ways one could possibly normalize a transfer function, and different people prefer different conventions.

4

Thank you!