Stochastic propeller force and moment reconstruction at a shaft end based on an improved Kalman filter

Abstract

Measuring the unsteady hydrodynamic excitation of a marine propeller at a shaft end accurately from multiple directions is crucial for minimizing the unwanted vibration and noise of ships. Due to the synchronous loading of high static thrust and torque, an innovative piezoelectric multicomponent balance embedded into a hollow shaft end is designed. However, complex coupling between the balance and shaft structure is formed. This leads to crosstalk disturbance and a modal dynamic amplification effect on all measuring channels. To overcome this, an improved Kalman filter algorithm based on the subspace modal identification (SMI) method is established, thus allowing us to reconstruct and locate equivalent unsteady excitations accurately in multiple directions with response-only networks. The effectiveness of proposed framework is test on a simulated nonrotating shaft end system with twelve force and three acceleration response channels. The selection of the smoothing delay parameter and whether crosstalk is considered are two main factors that influence the reconstruction accuracy.