Balancing multiple speeds flexible rotors without trial weights using multi-objective optimization

This paper proposes a multi-objective optimization method for the optimal balancing of flexible rotors without using trial weights to overcome the drawbacks of repeatedly starting and stopping and applying trial weight during the balancing process of multidisc flexible rotors. Finite elements are first used to develop a dynamic model of the flexible rotor. An optimization method and an inverse identification approach are then used to minimize the objective function by considering constraints and vibration levels of the flexible rotor during balancing. Finally, an optimal balancing scheme is solved by using a genetic algorithm. The optimization process here described consists in a global multi-objective optimization framework transformed into single-objective (SO) and dual-objective (DO) optimizations. Simulations and experiments are carried out on a three-disc flexible rotor balancing rig without trial weights, to benchmark the multi-objective optimization technique proposed here. The results show that the proposed methods can effectively suppress the unbalance vibration of the rotor. The balancing effect of the DO is better than the one offered by the SO during speedup of the rotor, and the effect of three-plane balancing is better than that of the double-plane one. The proposed method is efficient to avoiding multiple starts and stops during the balancing process, provides an enhanced balancing effect and also has strong scalability.