Validation of the actuator disc approach using small-scale model wind turbines

Abstract

Abstract. The aim of the present study is the validation of the implementation of an actuator disc (ACD) model in the computational fluid dynamics (CFD) code PHOENICS. The flow behaviour for three wind turbine cases is investigated numerically and compared to wind tunnel measurements: (A) the flow around a single model wind turbine, (B) the wake interaction between two in-line model wind turbines for a uniform inflow of low turbulence intensity and (C) the wake interaction between two in-line model wind turbines at different separation distances in a uniform or sheared inflow of high turbulence intensity. This is carried out using Reynolds-averaged Navier-Stokes (RANS) simulations and an ACD technique in the CFD code PHOENICS. The computations are conducted for the design condition of the rotors using four different turbulence closure models and five different thrust distributions. The computed axial velocity field as well as the turbulence kinetic energy are compared with hot-wire anemometry (HWA) measurements. For the cases with two in-line wind turbines, the thrust coefficient is also computed and compared with measurements. The results show that for different inflow conditions and wind turbine spacings the proposed method is able to predict the overall behaviour of the flow with low computational effort. When using the k-ε and Kato-Launder k-ε turbulence models the results are generally in closer agreement with the measurements.

Brief communication: Wind-speed-independent actuator disk control for faster annual energy production calculations of wind farms using computational fluid dynamics

“…The RANS-AD methodology used for our wind farm simulations are described in previous work (van der Laan et al, 2015b), and a brief summary is presented here. The RANS equation are solved in EllipSys3D, the in-house finite volume flow solver of DTU Wind Energy, initially developed by Sørensen (1994); Michelsen (1992). The k-ε-f P turbulence model is employed (van der Laan et al, 2015c), which has been developed for RANS-AD simulations using a neutral atmospheric surface layer.…”

“…This is generally not true for above-rated wind speeds, where the thrust force distribution is typically more uniform. However, Simisiroglou et al (2017) has shown that the effect of different thrust force distributions (with constant total thrust force) on the velocity deficit is mainly visible in the near wake, while the far wake is almost unaffected, especially when atmospheric turbulence is present. In addition, the wake effects above rated wind speeds are small due to the low thrust coefficient.…”

Faster wind farm AEP calculations with CFD using a generalized wind turbine model

“…As a result, the simple model without rotational speed limits does not compare well with the most of wind turbine data that do use rotational speed limits. This is not a significant issue for idealized AEP calculation using RANS, since the tip speed ratio only influences the blade force distributions (when employing the analytical rotor model [15]) and wake rotation magnitude, which are both small effects compared to the thrust coefficient, as shown in previous work [16] and by Simisiroglou et al ( 2017) [17]. In fact, one could remove the necessity of a tip speed ratio by employing an AD model based on a reference normalized thrust force distribution that is the same for all wind speeds, without tangential forces [18].…”

The flow in the induction and entrance regions of lab‐scale wind farms

“…Matching flow statistics around ADs and wind turbines, both for single wake generating objects (WGOs) and WGOs in wind farms, has received considerable attention in the literature. Examples of such validations in numerical simulations are found in previous studies 35–39 . Different studies have also compared AD wakes to wind turbine wakes experimentally.…”