Abstract.A concept for an optomechanical derotator is presented. The derotator allows optical measurements of rotating objects during operation. To guarantee a stationary optical image, the optical axis of the derotator needs to be coaxially aligned to the rotational axis of the measured object. The correlation between the movement of the optical image and a miscalibration is explained by a mathematical model. The movement of the optical image is tracked with a high-speed camera. Approximating the tracked path as a Limacon of Pascal, a parameter corresponding to the magnitude of the miscalibration is identified. This parameter is minimized by modifying the position and the orientation of the derotator with a hexapod. The limitations of this procedure are analyzed and further approaches are discussed.

In this paper an optomechatronical image derotator is used for vibration measurements on rotating objects. First of all, the concept of the derotator is explained and it is shown that the phase position and the rotational velocity of the derotator and the measurement object have to be aligned. Therefore, a highly dynamic tracking-control is needed. Considering the nonlinear friction of the synchronous motor, a model of the system which considers this non-linearity is evolved. This is accomplished by using neural networks for the approximation of the friction term. In this case General Regression Neural Networks (GRNN) are used for the learning algorithm. Moreover, the system's parameters, eg. the friction term and the inertia, are identified based on the nonlinear model. Then a feedback control is designed by using the controllable canonical form through feedback linearization. Finally, the results of vibration measurements on a rotating blisk using the nonlinear control concept for the derotator are shown.

No abstract