Hysteresis and Non-Linearity
Hysteresis is
a nonlinear phenomenon that can be found in diverse disciplines such as smart
materials [54], ferromagnetism and superconductivity [55]. The hysteresis refers
to a static memory effect rate independent hysteresis is characterized by (output
vs. input) hysteresis loops [56]. Practical hysteresis loops often demonstrate rate
dependent behaviour due to intimate coupling between hysteresis and traditional
dynamics [53].
One of the major
obstacles in the application of smart actuators is their hysteretic behaviour
and some degree of hysteresis. It has been observed by [53], the hysteresis
modelling and compensation methods used for smart actuator devices and
compensation is required to provide precision positioning systems.
A typical loop
response for a piezoceramic actuator driven by a voltage source is shown in
[25]. It was shown the position is multivalued with respect to the input
voltage and also it is the response is non-linear.
It is a
significant difference between the charging and discharging parts. The
repeatability of the loop is good to within a few percent. This fact can be
used to achieve highly repeatable open loop operation where the actuation is
synchronised to operate on one part of the loop.
The major obstacles
in the application of smart actuators are their hysteretic behaviour. If the hysteresis
effect is not incorporated into the control system it will act as an unmodelled
phase lag whose presence will cause instability in a closed-loop system if
sufficient phase margin is not provided in the control design. Integration of
actuators with these dynamics into mechatronic systems requires some degree of
hysteresis compensation, and also in some cases creep compensation, to provide
precision positioning [54].
The degree and
severity of hysteresis effects in smart actuators can often be reduced by
constraining input levels, or incorporating local control feedback. To improve the
control of the hysteresis and the linearity can do by controlling the
polarisation (charge) rather than the voltage [26]. The system controls the
charge on the actuator to achieve highly linear open loop control.
An alternative
strategy for controlling hysteresis is to apply a decaying oscillation to a
target DC voltage level. The displacement of the actuator converges to a point removing
hysteresis from the system. The hysteresis loop converges to the anhysteric
line when applied over a range of dc voltages. The non-linearity can be removed
by software compensation [25]. In present scanning tunnelling microscopic (STM)
designs, PID or robust control laws are used to reduce hysteresis effects [57,
58].
Reference:
No comments:
Post a Comment