Actuator line model development using simplified methods for force calculations

G. P. NAVARRO DIAZ; A. D. OTERO; HENRIK ASMUTH; JENS NØRKÆR SØRENSEN; STEFAN IVANELL

Hannover

Conferencia; Wind Energy Science Conference 2021; 2021

European Academy of Wind Energy

In this work, the capability to simulate transient wind turbine (WT) wake interaction problems is extended to the use of limited basic manufacturer WT information. For this purpose, two new variants of the actuator line model (ALM) are proposed in which the forces are computed locally using generic WT data, i.e. any combination of the thrust and power coefficient and the tip speed ratio. These implementations are introduced as extensions of the actuator disc model (ADM) in which the local adaptation is obtained by means of a numerical (Navarro Diaz et al., 2019) or an analytical (Sørensen et al., 2020) procedure. In this study, the new ALMs with analytical and numerical adaptations are compared  to the ADM, using the same data, as well as to the classical ALM based on blade element theory (Sørensen and Shen, 2002), which provides more detailed force distributions by using airfoil data. In order to evaluate the local force calculation, the analysis of a partial wake interaction case between two WTs is carried out, first for a uniform and then for a turbulent neutral atmospheric boundary layer inflow. Large Eddy Simulation is used, choosing the OpenFOAM software SOWFA (Churchfield et al., 2012) and the reference NREL 5MW wind turbine as a test case. For the uniform inflow case, below and above rated velocities are tested. Figure 1 shows the instantaneous volumetric blade force and vorticity for the below rated velocity on vertical planes at the position of the impacted WT, as well as instantaneous vorticity on horizontal planes at hub height. Four models are compared; the ADM with numerical local adaptation approach (ADM-num), two ALMs with numerical and analytical local adaptation (ALM-num and ALM-an, respectively) and the classical ALM build from airfoil data (ALM-airfoil). In this figure it can be noticed that all ALMs show that a better transient WT-airflow interaction and near wake description is obtained by applying the forces along each blade, as was also demonstrated by Martinez et al. (2012) and Marjanovic et al. (2017). For the same case, Figure 2 shows the time-averaged velocity on a horizontal line at hub height and 2 diameter upstream of the impacted WT. On this WT, also the time-averaged normal and tangential force distributions, with the blades in horizontal position, is shown. From the figure, it is noticed that the new ALMs give similar results in the force calculations at each side of the rotor. These results show that both ALMs and the ADM manage to capture the trend in the normal force distribution obtained with airfoil data. Bigger differences are found when comparing the tangential force distributions. The results demonstrate the possibility of simulating transient WT interaction problems in wind farms  by only using basic generic manufacturer data as input to the simulations.