AMOS: Analytical Methods for Optimal Vibration Reduction on General Rotors
Engineering Design, Machine Elements, Product Development
Final Report Abstract
Rotating machines are omnipresent in the technical world and are always subject to unbalance. This excites the rotor and thus the surrounding structure during operation. An active piezoelectric bearing for vibration reduction has been investigated in this project. The common control objectives of reducing the bearing forces and/or the rotor displacements are in conflict with each other. The use of the bending energy of the rotor was subject of the investigations of the present project. The objective was to reduce the bearing forces while also reducing the rotor deflection within the bending resonances of the unsupported rotor. A new relative displacement coordinate was introduced, which equals to the difference between the rotor displacements and the free actuator displacements, to represent the bending energy of the rotor system. The minimization of this relative coordinate in combination with the bearing forces shows advantages in the analytical investigations compared to the use of the absolute deflections. However, these advantages vanish on the real system. The experiments showed that using the relative deflections in combination with the bearing forces led to a reduction of the bearing forces and the required peak voltages which was the objective of using the the bending energy. However, if a sufficiently large weighting of the actuator voltages is implemented, the difference between the use of the relative and absolute deflections vanishes; both in theory and in the experiments. In practice, this weighting is required to increase the stability, which is why the use of the relative deflections no longer provides an advantage. The rotor deflections should only be considered in resonances which are poorly observable with the bearing forces, in order to achieve better control results. Outside of resonances, the rotor deflections can only be influenced in the order of magnitude of the actuator deflections. A control approach was developed for the active bearing which does not require any model of the rotor and is a combination of Integral Force Feedback (IFF) and the Least-Mean-Squares-Algorithm (LMS). Only the delay time of the system is required for implementation, but requires a collocation of actuator and sensor. The control can be used for a safe excitation of the rotor system during operation to estimate the gyroscopic transfer functions and in this way to implement a model-based control comprising IFF and LMS. In theory, a model-free control without the use of the delay time can be realised with rotating actuators. For this purpose, a prototype of a bearing with rotating piezo actuators was built and investigated. The experiments show that such an active bearing is capable of reducing the vibrations caused by the unbalance while reducing the required power for the actuators by 99 %. However, there are also undesirable side effects, such as strong vibrations of the third rotor order.
Publications
-
Active vibration control of a gyroscopic rotor using experimental modal analysis. Bulletin of the Polish Academy of Sciences Technical Sciences, 138090-138090.
Jungblut, Jens; Fischer, Christian & Rinderknecht, Stephan
-
A new active balancing device utilizing rotating piezo actuators. Mechanical Systems and Signal Processing, 181, 109521.
Jungblut, Jens; Haas, Julia & Rinderknecht, Stephan
-
Active vibration control of an elastic rotor by using its deformation as controlled variable. Mechanical Systems and Signal Processing, 165, 108371.
Jungblut, Jens; Haas, Julia & Rinderknecht, Stephan
-
System identification of a gyroscopic rotor throughout rotor-model-free control using the frequency domain LMS. IFAC-PapersOnLine, 55(25), 217-222.
Jungblut, Jens; Fischer, Christian & Rinderknecht, Stephan
-
„Exploiting gyroscopic effects for resonance elimination of an elastic rotor utilizing only one piezo actuator“. In: Proceeding of 15th SIRM – European Conference on Rotordynamics. Darmstadt, 2023.
J. Jungblut
