Comparison between CFD simulations and experiments for predicting the far wake of horizontal axis tidal turbines

Abstract

Actuator discs may be used as a simple method for simulating horizontal axis tidal turbines, both in experiments and CFD models. They produce a similar far wake to a real turbine, but eliminate some of the scaling issues which occur in experiments, and reduce the mesh density required in CFD simulations. This paper examines methods for applying a simple actuator disc in a commercial CFD code, Ansys CFX, and compares the wake produced with experimental results for similar values of disk thrust coefficient (C T). The results show that the CFD model gives reasonable agreement with the experimental results. The main factors affecting the wake structure are the initial C T value, the ambient turbulence levels, and potentially the disc induced turbulence. The main differences between the models and experiments were in terms of the turbulence levels throughout the model. With further development, it is considered that the CFD actuator disc could be an accurate and validated method for numerically modelling tidal turbines.

Accuracy of the actuator disc-RANS approach for predicting the performance and wake of tidal turbines

“…Since then it has been applied to several other applications such as the wake of wind turbines in a farm [3]. Recently, it has also been used in a number of cases to study unducted tidal stream turbine wakes [4][5][6][7][8][9][10][11][12][13][14]. Although this technique has computational advantages for modelling tidal stream turbines, the assumptions can cause inaccuracies in the representation of the near wake of a turbine.…”

“…Ambient turbulence data are not available for either of these cases, so it is difficult to specify why the variation between the models occurs. Centreline normalized velocity profiles behind a range of actuator disc models by different authors, the thrust coefficient of the turbine being 0.9, other than the case of [6] (solid line, Roc et al [11]; dashed line, Sun et al [6]; dashdotted line, Gant & Stallard [7]; line with crosses, MacLeod et al [4]; squares, Turnock et al [14]; line with circles, Harrison et al [12]). …”

Modelling of the flow field surrounding tidal turbine arrays for varying positions in a channel

“…Device-generated turbulence and the initial velocity reduction in the near wake of tidal turbines are not equivalent between actuator discs/fences and mechanical rotors. However, a crucial factor mentioned in Harrison et al [8] which justifies this method is that the downstream spacing between tidal turbines in a farm is likely to be seven turbine diameters or more, as with many wind turbine farms to date.…”

“…The computational domain used was almost identical to the basic setup of a conventional water flume, with sidewalls and an inflow, outlet and seabed. As with the work presented in Harrison et al [8] and Lartiga & Crawford [9], horizontal axis turbines were modelled using a momentum sink approximation. Once again the k-ω SST turbulence model was used, with the authors justifying use of this method by stating that it performs better in regions of adverse pressure gradients.…”

“…The final method used differs quite substantially from the other methods used and those used in the studies of Harrison et al [8], Lartiga & Crawford [9] and Bai et al [11]. It involved using the interface capabilities of Ansys CFX to constantly recirculate flow from the outlet of the Chilworth flume back to the inflow, in order to allow the flow to continually develop.…”