| Summary: | <p>This thesis adopts a blade element actuator disc method, embedded in a Reynolds Average Navier-Stokes (RANS) solver, to design multi-rotor tidal arrays. The constructive interference effect is taken into account in the design process to maximize the hydrodynamic efficiency. The design process considers rotor arrays in a fence arrangement, investigating the relative merits of in-situ and uniform rotor designs. The designed rotor is fabricated and tested experimentally. The experimental results show that the rotor achieved superior power efficiency and substantial performance benefits are possible through the fence arrangement.</p>
<p>High-fidelity blade-resolved computations are conducted to evaluate the design process and to reveal detailed flow physics. The blade loading distributions in the mid-span region from the design model shows good agreement with the those from the blade-resolved model. Discrepancies are observed in the blade root and tip regions, primarily due to the three- dimensional flow effects that are not fully resolved in the RANS-BE model.</p>
<p>To increase the modelling accuracy of the reduced order design model, a theoretically derived blade loading correction model is developed to account for three-dimensional spanwise flow effects along the rotating blades. The correction formulation, based on inviscid flow assumptions and the radial component of the governing equations, relates the spanwise pressure gradient on the blade surface to the shedding of bound circulation to form the blade's trailing vortex sheet. The new model corrects for spanwise flow induced pressure effects, which cannot be simply accounted for by manipulation of the inflow to two-dimensional aerofoils modelling through sectional lift and drag data. The analytic model, closed with appropriately calibrated empirical coefficients, is used to anisotropically modify the lift and drag forces on blade sections within the framework of the Blade Element Momentum model. With the new correction model applied, in addition to the Glauert tip correction to account for finite blade number, the new model is found to significantly improve the prediction of the thrust and torque on outboard sections of the blade where spanwise flow effects are significant. The empirical coefficients in the model only show a slight dependence on the operating tip-speed-ratio and the flow blockage. However, the sensitivity of the empirical coefficients on different blade geometries are not examined in this thesis.</p>
|
|---|