1 September 2016 in Journal of Sound and Vibration
This paper presents a novel topology for enhanced vibration sensing in which wireless MEMS accelerometers embedded within a hollow rotor measure vibration in a synchronously rotating frame of reference. Theoretical relations between rotor-embedded accelerometer signals and the vibration of the rotor in an inertial reference frame are derived. It is thereby shown that functionality as a virtual stator-mounted displacement transducer can be achieved through appropriate signal processing. Experimental tests on a prototype rotor conrm that both magnitude and phase information of synchronous vibration can be measured directly without additional stator-mounted key-phasor sensors. Displacement amplitudes calculated from accelerometer signals will become erroneous at low rotational speeds due to accelerometer zero-g osets, hence a corrective procedure is introduced. Impact tests are also undertaken to examine the ability of the internal accelerometers to measure transient vibration. A further capability is demonstrated, whereby the accelerometer signals are used to measure rotational speed of the rotor by analysing the signal component due to gravity. The study highlights the extended functionality aorded by internal accelerometers and demonstrates the feasibility of internal sensor topologies, which can provide improved observability of rotor vibration at externally inaccessible rotor locations
1 November 2019 in Fusion Engineering and Design
24 April 2017 in IEEE/ASME Transactions on Mechatronics
An active rotor topology for vibration control in rotating machinery is proposed. Microelectromechanical system accelerometers and miniature mass-balancer actuators are placed within a hollow rotor to produce a self-sensing self-actuating design. When compared with conventional stator-mounted active systems, this topology enables an increased number of sensors and actuators, with greater freedom to select their location along the rotor. The construction and evaluation of a prototype is reported. An algorithmic direct search controller is adopted to provide non a priori vibration control. The searching algorithm is specified to solve a discrete problem, accounting for the limits in resolution of sensors and actuators. In addition, features have been introduced to prevent premature convergence at saddle points. The controller has been applied both in simulation and experimentally, achieving a substantial reduction in rotor vibration.