Inverse modeling approach for transformation of propeller shaft angular deformation and velocity to propeller torque load
In full-scale trials or model-scale tests, the propeller torque load is typically measured indirectly on the propeller shaft. Due to structural flexibility and mass of the propeller and propeller shaft, the measured propeller shaft torque response includes a time delay and amplitude difference compare to the propeller load. These differences are usually small during propeller operation, but may be significant during transient states (i. e. starting of the engine, rough sea, or operation in ice-covered waters). In this paper, an inverse model is used to calculate propeller load based on the propeller shaft response. The results indicate that the inverse model can be less complex than the complete propulsion machinery model and consider only the propeller and propeller shaft. Several modeling approaches are tested based on four proposed measuring setups and three forms of system discretization. Out of five inverse modeling approaches, two approaches showed the ability to transform measured propeller shaft angular deformation and velocity responses, colored with noise, into the corresponding rule-based propeller torque load with acceptable accuracy. Furthermore, the robustness of the inverse approaches to the random ice-propeller interaction, sampling frequency, measuring location along the propeller shaft and starting time (before and during ice-propeller interaction), is discussed.