Abstract
The accurate prediction of dynamic characteristics is critical for mitigating resonance risks in marine propulsion shafting. Conventional single-point support models, while computationally simple, are fundamentally flawed for large-aspect-ratio bearings as they neglect the lateral rotational constraint. This omission can lead to significant miscalculations of natural frequencies, posing a hidden threat to vessel safety. To address this critical limitation, a novel double-point equivalent support model is proposed for water-lubricated stern bearings. The model parameters are derived from a comprehensive mixed-lubrication analysis, which accounts for the fluid and contact pressure distributions across a range of operational speeds and lateral deflection angles. Compared to the benchmark mixed-lubrication model, the presented method achieves equivalence in lateral rotational constraint, whereas even distributed support models exhibit obvious deviations. A subsequent dynamic analysis demonstrates that the proposed double-point model, through its adaptive support stiffness and positions, faithfully replicates the lateral rotational constraint and exhibits high sensitivity to operational variations. Conversely, the conventional single-point model fails to capture these effects, leading to substantial deviations in predicted natural frequencies and modal shapes, thereby increasing resonance risk. These findings highlight the necessity of incorporating lateral rotational constraints in large-aspect-ratio bearing models to enhance the reliability of shafting system design.