Experimental and Numerical Investigation of Blade-Tip Rotors on Vertical-Axis Wind Turbines

Abstract

A study is conducted on the effects of tip-mounted rotors on a conventional Darrieus H-VAWT; this study is complementary to the X-rotor concept proposal. This proposal features a novel offshore wind turbine concept that utilises blade-mounted tip rotors to extract energy from the flow. Following a literature study phase, an experimental campaign is devised that utilises various porous actuator mesh disks as an approximation of blade tip-mounted HAWT geometries. Additionally, the implementation of a blade element vortex model is also investigated.

Experimentation using an existing VAWT is conducted at the Open Jet Facility at TU Delft. The test is conducted at a freestream velocity of V=4.0 m/s at an average Reynolds number of Re=7.5E4. Key outputs from the test are the normal forces on the primary (VAWT) rotor blades, the torque measured at the shaft of the primary rotor and the thrust force of the primary rotor. Through post-processing, it is also possible to determine the thrust of the tip-mounted actuator mesh disk and verify the outcome with empirical models. Moreover, large-scale volumetric flow field measurements are also conducted using a PTV system which comprises a coaxial volumetric velocimetry probe mounted to a robotic arm. The PTV system uses Helium Filled Soap Bubbles as tracers to measure the flow field.

Key results from the mesh disk study are consistent with initial assumptions; the thrust of the actuator disk is inversely proportional to the porosity. However, there are discrepancies observed between the different means of measuring the actuator disk thrust, this suggests that experimental inaccuracies are present. Effects of the actuator mesh disk on the blade performance showed a reduction of blade normal forces, an increase of the primary rotor thrust force as well as changes to the turbine thrust vector direction.

The measurements conducted using PTV highlight changes to the blade wakefield which has consequences for the BVI. Notable differences between the cases with and without actuator disks include the lack of the upwind shed vortex in the investigated region when the actuator mesh disk is present. In the case without the actuator mesh disk, the interaction with the downwind and upwind shed vortices is destructive, thus the vorticity in the wake dissipates more rapidly. Tests with the actuator mesh disk conducted at TSR=4.0 showed that the wake of the actuator disk is able to persist longer and thus interact with the second blade leading to BVI. The lack of the upwind shed vortex destructing the downwind vortex likely means that the wake dissipates more slowly. Moreover, the presence of reverse flow behind the actuator disk is identified as another contributor to the large velocity deficit observed in the wake.

The numerical vortex model that is proposed is an extension of the pre-existing CACTUS vortex model developed by Sandia National Laboratories. The changes to the CACTUS numerical model enable the implementation of tip rotor geometries to VAWTs. Initial testing shows behaviour that is within expectations, however, the results are limited by a coarse spatial grid; this highlights the necessity for optimisation of the numerical model in the future. The numerical vortex model demonstrates that it is possible to use this model for the investigation and optimisation of blade tip-mounted rotors.