Adaptive rotor systems
with piezoelectric actuation
P.J. Sloetjes, A. de Boer
Institute of Mechanics, Processes and Control - Twente Chair of Structural Dynamics and Acoustics, University of Twente
P.O. Box 217, 7500 AE Enschede, The Netherlands phone +31-(0)53-4893405, email p.j.sloetjes@ctw.utwente.nl
Introduction
High speed rotating machines can be supplied with
unbalance compensation, stabilization & excitation
functions by adding piezoelectric fiber actuators or
ultrasonic motors with accompanying control &
identification algorithms. The resultingadaptive rotor
systemsare complex but can be modelled by taking advantage of e.g. energy conservation principles,
spatiotemporal periodicity and weak coupling
between subsystems & physical subdomains.
Experimental work is required to investigate the full system dynamics without tedious simulation.
Objective
The aim is to gain insight in model-based monitoring & control of weakly actuated rotor dynamic systems.
Methods
‘Functional complexity’ of engineering systems is related to the uncertainties which complicate the achievement of functional requirements. Sources of
uncertainty in rotating machinery are damping,
unbalance, misalignment, bearing properties and
component degradation. ‘Artificial intelligence’ methods are able to succeed in spite of uncertainty. Such methods can be passive, e.g., methods which combine algorithms for discernment, classification &
memorization of measurement data features, or
active, e.g. methods which use planned diagnostic
teststo improve the convergence speed of model & fault identification (see Fig.1a). For example, an active balancing device may be used to apply a large oriented excitation for identification purposes.
Results
Arotor dynamics code for general asymmetric rotor systems with sensors, actuators and controllers was
developed in Matlab (Fig.1b). Anexperimental setup
with piezoelectric actuators was built (Fig.1c).
Submodels for high voltage generation and power
harvesting were simulated and tested. Currently, submodels and components are being integrated and cosimulation of nonlinear electromechanical models is being considered. Further research will focus mainly on extending the current models by nonlinear models of hysteresis, supports, bearings and failure.
Fig.1a) Advanced models and fault identification
fgf
Fig.1b) Basic models and control algorithms
Fig.1c) Rotor dynamics setup with actuators
0RGHOVIRUV\VWHPGHVLJQDQDO\VLVDQGFRQWUROOHUGHVLJQ $GYDQFHGPRGHOVDQGILOWHUV &RPSXWHUV\VWHPIRUFRQWURODQGLGHQWLILFDWLRQ /LQHDUURWRUDQGGLVWXUEDQFHPRGHOV )LEHUDFWXDWRU KYDPSOLILHU 8OWUDVRQLF PRWRUGULYHU $FWXDWRUGHYHORSPHQW $FFHOHURPHWHU FRQGLWLRQHU (GG\FXUUHQW FRQGLWLRQHU 0RWRUIP DPSOLILHU $QJXODU FRQGLWLRQHU s p s q t i s rg s rg s rg i s sg s sg s sg s mg s mg s mg e g e Mx Cx Kx m e f f x K x C x M x K x C x M + + Ω + = + + + + + + + + + Ω 2 2[ ] ] [ ] [ θ 9HORFLW\IHHGEDFN IRUVWDELOL]DWLRQ $GDSWLYHXQEDODQFH FRPSHQVDWLRQ 5RWRUDQGDFWXDWRU SDUDPHWHUHVWLPDWLRQ 'LVWXUEDQFHHVWLPDWLRQ 6LJQDOSURFHVVLQJ 'HVLJQRIIDLOXUHVFHQDULRWHVWV )RXULHUZDYHOHW WLPHIUHTXHQF\ 'HYHORSPHQWRIQRQOLQHDUPRGHOVRIVXSSRUWV K\VWHUHVLVEHDULQJVFUDFNVFRQWDFW 'HYHORSPHQWRIVLJQDOSURFHVVLQJPHWKRGVIRU LGHQWLI\LQJFUDFNVEHDULQJIDLOXUHORRVHFRPSRQHQWV 7HVWVFHQDULRGHVLJQ &RQWUROOHGXQEDODQFH $UELWUDU\H[FLWDWLRQ 9LUWXDOUHVSRQVHPRGHO 0HPRU\ 1RUPDOUHVSRQVH &DVHKLVWRU\ 1RQOLQHDUUHVSRQVHSUHGLFWLRQ )DXOWPRGHOFRVLPXODWLRQ )DXOWLGHQWLILFDLWRQ )X]]\ORJLF GHFLVLRQVFKHPH $FWXDWRUPRGHOV 0 1− = + = Ω − Tr q q q q t i q q s q Klve Rq Cq Klx f ) , ) , ( , (2 1 ) ( 2 12 n ns nc ns nc n n t i p p p s p dm e Jφ M v v v v σ φ φ ∠ + = Ω = Ω++ l f
