Theoretical and experimental applications of a rotor balancing technique without using trial weights based on augmented Kalman filter

Abstract

Unbalance is the most common fault of rotating machines and efforts devoted to developing balancing techniques are the goal of several researchers around the world. The most common balancing approach is the well-known influence coefficient method, which requires the use of trial weights to determine the relationship between unbalance forces and vibration responses of the rotor; consequently, it is time-consuming regarding industrial applications. The present contribution investigates the augmented Kalman filter (AKF) technique to balance a flexible rotor without using trial weights. There are few studies in this area and the majority of the previous research works are dedicated to theoretical studies using simple or reduced rotor models. This paper considers a finite element (FE) model of a realistic rotor system and investigates the effectiveness of the AKF as a balancing technique by considering several unbalance conditions. Therefore, the novelty of the present contribution is characterized by a complete numerical and experimental AKF based balancing procedure that is performed by using a realistic rotor system. As mentioned above, limited experimental contributions are found in the literature involving real-world applications. Additionally, a sensitivity analysis on the force covariance matrix is proposed to investigate the most favorable conditions to apply the AKF technique in the context of rotor balancing. The methodology proved to be effective to deal with a variety of rotor balancing applications.