On the inverse design of marine ducted propulsor blading

A three-dimensional computational design method is presented for multi-component marine ducted propulsors operating in axisymmetric shear flow. An inverse design approach is adopted whereby the blade shapes at the specified design point are determined from a given blade circulation distribution and a condition of zero blade incidence. The corresponding three-dimensional propulsor flow field is also obtained, enabling the design-point performance of the synthesised design to be assessed. Although based on an existing inviscid turbomachinery design technique, the method incorporates numerous modifications, firstly to enable propulsor mass flow to be determined, and secondly to model the shear flows that are encountered by propulsors operating within the boundary layers of ships. Employing an assumption of inviscid flow and using the Clebsch representation of vorticity, both the propulsor through flow and bypass flow are described by simultaneous partial differential equations which are solved using finite difference and Fourier techniques. The development of circumferential variations in both velocity and vorticity within the blade passages are included, as are the effects of spanwise variations of blade circulation. Moreover, by assuming that the duct wake remains axisymmetric slipstream contraction is accounted for. In addition to the above formulation, a simplified three-dimensional formulation which neglects the development of circumferential variations in stagnation pressure is described, as is an "actuator duct" approach which neglects the circumferential variations of all flow quantities. Computational results for practical ducted propulsors show the development of circumferential variations in flow quantities to have relatively little effect on either blades shape or overall propulsor performance predictions. Finally, the experimental verification of the design method using a low speed wind tunnel is outlined. Velocity measurements conducted with a three-hole pitot probe show reasonable agreement with the predictions of the design method, whilst measurements of shaft power and mass flow show much closer agreement