| Summary: | Thesis (MScEng (Mechanical and Mechatronic Engineering))--University of Stellenbosch, 2006. === This thesis investigates practical methods of modelling and control of the vertical
motions of a hydrofoil assisted catamaran, the HYSUCAT. The aim of the
control application is to reduce the motions, and consequently the motion
sickness of the passengers.
First, a potential flowcommercial program, POWERSEA,was used to model
the system. This uses 2-D strip methods to model the planing hull-form of
the vessel, and the Peter du Cane hydrofoil theory for modelling of the foils.
These simulations are compared to experimental towing tank results, with fair
agreement at lower speeds, but limited applicability at high speeds. Thus for
the control design the agreement was insufficient.
As an alternative, a simple coupled 2 degree-of-freedom spring - mass -
damper model is proposed, for which the equations of motion are derived.
This has 9 unknown parameters; three of these aremeasured directly, two are
modelled, and the remaining four were identified using an experimental parameter
estimation technique. Representative parameter values were calculated
frommultiple experiments for application in the control design.
The design of a control system was based on the above model. First, an
output-weighted Linear Quadratic Regulator (LQR) was designed to obtain
the full state feedback gains. A non-linear ’bang-bang’ control design was
then implemented to try and speed up the response of the system. These
control strategies, as well as no control, were applied in the towing tank in
regular waves, with good results at low and medium frequencies. At the design
point, 32% and 65% reductions in rms motions were achieved for pitch
and heave, respectively. At high frequencies, though, not much improvement
was achieved due to the bandwidth limitation of the control system. The LQR
results were better overall (reduced motions) across the frequency range than
the bang-bang controller, as well as having a lower added resistance in waves.
The control design of the output-weighted LQR was then revised to be
based on alternative outputs, as a possible improvement. However, a further
two controller designs did not yield any noticeable improvement and were
not developed further.
|
|---|
