Effects of Unbalance Identification Locations on Transient Dynamic Balancing Without Trial Weights Performance of Power Turbine Rotor

Highlights: The main findings of this paper are summarized as follows: First, a transient dynamic balancing method without trial weights was developed for a specific type of power turbine rotor. Based on modal balancing theory, this method identifies the rotor unbalance by calculating the unbalance excitation force. Second, the applicability of unbalance identification at different axial correction mass positions was systematically analyzed for the investigated rotor model. The implications of the main findings are generalized as follows: First, the proposed transient dynamic balancing method requires only a single rotor startup and identifies the rotor’s unbalance without adding any trial weights, which significantly improves balancing efficiency. Second, the research on the applicability of unbalance identification across various axial correction mass positions on an actual rotor model significantly improves the efficiency of on-site dynamic balancing operations. This study proposes a dynamic balancing method without trial weights for power turbine rotors and investigates how the axial location chosen for unbalance identification affects the balancing performance. A finite element model of the power turbine rotor system was established to compute transient vibration responses and principal modes. Both continuous and isolated unbalances are employed to identify unbalanced excitation forces, enabling the determination of unbalance parameters. Furthermore, variations in identification accuracy across four designated axial positions on the rotor were analyzed. Simulations and experiments conducted on boss 2 and boss 3 confirmed the method’s efficacy: the maximum vibration amplitudes were reduced by 70.48% and 45.81% for boss 2, and by 64.48% and 61.00% for boss 3, respectively. These results verify the effectiveness of the proposed method. The unbalance parameters identified from simulations exhibited errors within ±6°, ±0.12 g, and ±0.15 × 10⁻⁴ m, while experimental errors remained within ±5°, ±0.11 g, and ±0.10 × 10⁻⁴ m, demonstrating high accuracy and reliability. Notably, this method improves balancing efficiency by requiring only a single startup and facilitates vibration data acquisition in confined spaces.