A novel balancing method for rotor based on unsupervised deep learning is proposed in this paper. The architecture of the proposed deep network is described. In the proposed network, compared to the supervised deep network, additional convolution layers are applied not only for the learning of the inverse mapping but also for identifying the unbalanced force without labeled data. The equivalent value and position of imbalances in two correction planes are obtained. A case study of a rotor with two discs supported by sliding bearings is conducted. Preset imbalances are balanced well by the proposed method. And, using the state values at different time intervals, no extra weight trails are needed. The results show that the proposed balancing method gives consideration to both cost and accuracy.

The modal balancing method (MBM) is an effective method for reducing the vibration caused by the unbalancing of a rotor. A rotor’s modal parameters, especially its modal shapes, need to be accurately calculated in this method. This paper proposes an optimized modal balancing approach for flexible rotors. The vibration modes of the rotor are first obtained with experimental modal analysis based on the rotor’s response signals while the rotor is speeding up. The rotor balancing strategy is subsequently optimized by the sensitivity analysis of the mode shapes. Orthogonal trial masses are obtained based on the orthogonality of each vibration mode, and the correction masses are finally calculated by using the influence coefficients of the trial masses. An experimental result is shown to demonstrate and validate that the proposed approach is able to achieve superior accuracy compared to the conventional MBM.