CFD surface effects on flow conditions and tidal stream turbine performance.
Catherine Lloyd*, Matthew Allmark, Robert Ellis, Stephanie Ordonez, Allan Mason Jones, CameronJohnstone, Tim O’Doherty, Gregory Germain and Benoit Gaurier*Cardiff Marine Energy Research Group (CMERG), School of Engineering, Cardiff University, Cardiff, Wales, UK
This work was undertaken as part of the DyLoTTA project – Dynamic Loading of Tidal Turbine Arrays.This work was undertaken as part of the DyLoTTA project – Dynamic Loading of Tidal Turbine Arrays.The project seeks to help understand the loading of horizontal axis tidal turbines subjected to avariety of operating scenarios. Within this work, as a pre-curser to the modelling of tidal arrayssubjected to wave climates, the effect of modelling a single HATT subject to a regular wave, usingtwo differing boundary setups was studied. A ‘free surface’ and a ‘free slip’ numerical model werecreated, to assess the differences between their surface boundary conditions and how this causeschanges in the generated flow conditions and turbine performance under uniform current flowconditions. The ‘free slip’ model was a single phase, incompressible flow model with the ‘top’boundary at the still water level using the ‘free slip’ boundary condition. The ‘free surface’ modelwas a homogenous, multiphase model with a distinct free surface interface between the water andair phases. The ‘top’ boundary was in the air region of the model and specified as an ‘opening’allowing bidirectional flow across the boundary. Development of the model was split into 3 mainssections: Geometry, Mesh & Physics setup. The 2 different numerical models were compared toexperimental data obtained at IFREMER, Boulogne Sur Mer, France . The average performancecharacteristics show little difference between the numerical models. Therefore, the transientdynamic loadings were investigated (Figure 3). There was little difference in average individual bladetorque and thrust between the models. The ‘free surface’ model estimated the average thrust onthe hub to be over double that of the ‘free slip’ model, giving 20.1 N and 8.2 N respectively. This isdue to the addition of hydrostatic forces in the ‘free surface’ model which affect the hubsignificantly. The fluctuation of the individual blade thrust results show good agreement betweenthe 2 model types.