Specifing input variables for Linearization in Physics/Mechanics

[AlessandroZavoli]

When using the linearization routines in sympy.physics.mechanics
it is not clear how to define the system inputs $r$

In particular, the form
$M \delta x' = A \delta x + B \delta r$
returns B empty even in the case of applied non-conservative forces that depend on a parameter (i.e., the system input).

What is the correct way to specify that a symbolic variable is a system input?

[moorepants]

I don't think you have to do anything. If you have symbols that are functions of time in your equations of motion that are not state variables, then the linearize method treats them as inputs.

[AlessandroZavoli]

Do you mean that r is needs to be a Function of t or a dynamicvariable?

[moorepants]

Yes.

[AlessandroZavoli]

Using a dynamicsimbol, the matrix B is no longer empty.
Yet, I think something is not working properly: before the linearization i would expect to be able to set the operating-point value (so that the EoMs are respected).
Alos, I should be able to retrieve the list of reference inputs in some way after the linearization

[moorepants]

You can set the operating point for any dynamicsymbol in that kwarg. The method returns A, B, and r, no? or M, A, B, r. So the list of inputs is there.

[AlessandroZavoli]

Using the dynamicsymbols the matrix B is now non-empty. Yet, the method does not return r as 4th argument
M, A, B, r = linearizer.linearize(A_and_B=False, op_point=op_point) returns ValueError: not enough values to unpack (expected 4, got 3).
Nevertheless, by the reference variables are correctly stored in linearizer.r on the creation of the linearizer object.

I will try to write a minimum example to reproduce the error (or point me to one if you have it)
Best,
Alessandro

[moorepants]

It doesn't look like the Linearizer returns r:

https://github.com/sympy/sympy/blob/master/sympy/physics/mechanics/linearize.py#L259

[moorepants]

The linearize() methods of LagrangesMethod and KanesMethod return r: https://github.com/sympy/sympy/blob/master/sympy/physics/mechanics/kane.py#L632