Abstract
Low-center-of-gravity wind turbines (LCGWTs) characterized by tapered blades whose chord length c increases nonlinearly from the top (where c = 0.11 m) to the bottom (where c = 0.17 m) of each blade. Further, turbines featuring these blades do not need any arms, or even a center pole in the rotor. Two experimental LCGWTs (diameter: 0.4 m; height: 0.25 m) with symmetrical blades (NACA 0018) and cambered blades were built. A dead band, which is a band of tip speed ratio (TSR) where the rotor has negative torque at TSR lower than that where the maximum power-coefficient condition is achieved, was observed when symmetrical blades were subjected to low wind speed. In contrast, no dead band was observed for the cambered blades. Under high wind speeds and over a wide range of TSR values, performance of the LCGWTs was better with cambered blades than with symmetrical blades. Computational fluid dynamics (CFD) analysis of 2-dimensional rotors whose blade sections corresponded to the blade sections at the equatorial planes of both types of LCGWTs showed the same tendency. Performance predictions by the blade element momentum (BEM) method using aerodynamic data on the NACA 0018 blades showed some agreement with the CFD analysis. For the cambered blade rotor, Wilson and Walker's empirical correction of the thrust coefficient, a correction that is typically used in simulations of horizontal axis wind turbines, brought the BEM prediction closer to the CFD prediction than Glauert's correction did. However, the agreement between the BEM prediction with Wilson and Walker's correction and the CFD prediction of the cambered blade rotor was thought to be just a coincidence due to large difference on the torque variations between BEM and CFD. At least, the Wilson and Walker's correction predicts larger torque than the Glauert's correction at high TSR region.
