Small-scale Vertical Axis Wind Turbine Design for Testing and Optimization in the Princeton High Reynolds Number Test Facility

Abstract

Vertical axis wind turbines (VAWTs) have recently gained recognition for their suitability for niche applications, such as in urban and offshore environments. Further testing of VAWTs is required to provide an improved understanding of VAWTs, especially for self-starting characteristics and optimizing VAWT configurations for performance. Experiments with scaled down models offer a cheap and efficient way to study wind turbine behavior over resource intensive large-scale experiments and computationally expensive numerical simulations. A major challenge in wind turbine testing has been the inability to match both the non-dimensional Reynolds number and tip speed ratio that are required for full dynamic similarity. The Princeton High Reynolds Number Test Facility (HRTF) allows for wind tunnel tests at up to 238 atm, allowing Reynolds numbers to be increased independently of the tip speed ratio, and so offers the unique prospect of simultaneously matching both Reynolds number and tip speed ratio in a small-scale experiment. The aim of this thesis project is to design and construct a 1 scale model of the Hopewell Wind Power Mark II VAWT to withstand the high forces in the HRTF while featuring a modularity which allows for efficient optimization and iterative testing. The model allows for Reynolds numbers up to 2.5×107 to be matched at an operating tip speed ratio of 3. This corresponds to wind speeds of 9.8 m s−1 for the large scale prototype, well above typical average wind speeds. The estimated coefficient of power is 0.26. Preliminary force analyses offer some estimation of turbine performance and predict that the turbine is able to self-start. The experimental trials confirm the ability of the VAWT model to self-start to low tip speed ratios. Experimental results also provide some validation of the simulation models and offer valuable insight into further optimization and interesting aspects of VAWT performance that can be investigated with further testing of the small-scale model.