The present work aims to investigate different methodologies for the numerical simulation of an upwind three-bladed wind turbine; which is supposed to be a base model to simulate icing in cold climate windmills. That is a model wind turbine for which wind tunnel tests have been completed at the Norwegian University of Science and Technology (NTNU). Using the assumption of axisymmetry, one-third of rotor has been modeled and periodic boundaries applied to include the effects of other blades. Then the full rotor was studied with transient simulation. To take in the effects of wind turbine wakes, the wind tunnel entrance and exit have been considered 4 and 5 diameters upstream and downstream of the rotor plane, respectively. Furthermore, the effects of tower and nacelle are included in a full-scale transient model of the wind tunnel. Structured hexa mesh has been created and the mesh is refined up to y+=1 in order to resolve the boundary layer. The simulations were performed using standard k-e, Shear Stress Transport (SST) model and a sophisticated model Scale-Adaptive Simulation (SAS)-SST to investigate the capability of turbulence models at design and off-design conditions The performance parameters, i.e., the loads coefficients and the wake behind the rotor were selected to analyze the flow over the wind turbine. The study was conducted at both design and offdesign speeds. The near wake profiles resulted from the transient simulation match well with the experiments at all the speed ranges. For the wake development modelling at high TSR, the present simulation needs to be improved, while at low and moderate TSR the results match with the experiments at far wake too. The agreement between the measurements and CFD is better for the power coefficient than for the thrust coefficient

To develop a numerical model of icing on wind turbine blades, a CFD simulation was conducted to investigate the effect of critical ice accretions on the aerodynamic characteristics of a 0.610 m chord NACA 0011 airfoil section. Aerodynamic performance coefficients and pressure profile were calculated and compared with the available measurements for a chord Reynolds number of 1.83x106. Ice shapes were simulated with flat plates (spoiler-ice) extending along the span of the wing. Lift, drag, and pressure coefficients were calculated in zero angle of attack through the steady state and transient simulations. Different approaches of numerical studies have been applied to investigate the icing conditions on the blades. The simulated separated flow over the sharp spoilers is challenging and can be seen as a worst test case for validation. It allows determining a reliable strategy to simulate real ice shapes [1] for which the detailed validation cannot easily be provided.

The main objective of this study is to estimate the dynamic loads acting over a glaze-iced airfoil. This work studies the performance of unsteady Reynolds-averaged Navier-Stokes (URANS) simulations in predicting the oscillations over an iced airfoil. The structure and size of time-averaged vortices are compared to measurements. Furthermore, the accuracy of a two-equation eddy viscosity turbulence model, the shear stress transport (SST) model, is investigated in the case of the dynamic load analysis over a glaze-iced airfoil. The computational fluid dynamic analysis was conducted to investigate the effect of critical ice accretions on a 0.610 m chord NACA 0011 airfoil. Leading edge glaze ice accretion was simulated with flat plates (spoiler-ice) extending along the span of the blade. Aerodynamic performance coefficients and pressure profiles were calculated and validated for the Reynolds number of 1.83 × 106. Furthermore, turbulent separation bubbles were studied. The numerical results confirm both time-dependent phenomena observed in previous similar measurements: (1) low-frequency mode, with a Strouhal number Sth≈0,013–0.02, and (2) higher frequency mode with a Strouhal number StL≈0,059–0.69. The higher frequency motion has the same characteristics as the shedding mode and the lower frequency motion has the flapping mode characteristics

Wind turbines installed in cold-climate regions are prone to the risks of ice accumulation which affects their aeroelastic behavior. The studies carried out on this topic so far considered icing in a few sections of the blade, mostly located in the outer part of the blade, and their influence on the loads and power production of the turbine are only analyzed. The knowledge about the influence of icing in different locations of the blade and asymmetrical icing of the blades on loads, power, and vibration behavior of the turbine is still not matured. To improve this knowledge, multiple simulation cases are needed to run with different ice accumulations on the blade considering structural and aerodynamic property changes due to ice. Such simulations can be easily run by automating the ice shape creation on aerofoil sections and two-dimensional (2-D) Computational Fluid Dynamics (CFD) analysis of those sections. The current work proposes such methodology and it is illustrated on the National Renewable Energy Laboratory (NREL) 5 MW baseline wind turbine model. The influence of symmetrical icing in different locations of the blade and asymmetrical icing of the blade assembly is analyzed on the turbine’s dynamic behavior using the aeroelastic computer-aided engineering tool FAST. The outer third of the blade produces about 50% of the turbine’s total power and severe icing in this part of the blade reduces power output and aeroelastic damping of the blade’s flapwise vibration modes. The increase in blade mass due to ice reduces its natural frequencies which can be extracted from the vibration responses of the turbine operating under turbulent wind conditions. Symmetrical icing of the blades reduces loads acting on the turbine components, whereas asymmetrical icing of the blades induces loads and vibrations in the tower, hub, and nacelle assembly at a frequency synchronous to rotational speed of the turbine.

Icing limits the performance of wind turbines in cold climates. The prediction of the aerodynamic performance losses and their distribution due to ice accretion is essential. Blade element momentum (BEM) is the basis of blade structural studies. The accuracy and limitations of this method in icing condition are assessed in the present study. To this purpose, a computational study on the aerodynamic performance of the full-scale NREL 5 MW rotor is performed. Three-dimensional (3D) steady Reynolds-averaged Navier–Stokes (RANS) simulations are performed for both clean and iced blade, as well as BEM calculations using two-dimensional (2D) computational fluid dynamics (CFD) sectional airfoil data. The total power calculated by the BEM method is in close agreement with the 3D CFD results for the clean blade. There is a 4% deviation, while it is underestimated by 28% for the iced one. The load distribution along the clean blade span differs between both methods. Load loss due to the ice, predicted by 3D CFD, is 32% in extracted power and the main loss occurs at the regions where the ice horn height exceeds 8% of the chord length.

Icing of wind turbine blades poses a challenge for the wind power industry in cold cli-mate wind farms. It can lead to production losses of more than 10%of the annual energyproduction. Knowledge of how the production is affected by icing is of importance. Com-plicating this reality is the fact that even a small amount of uncertainty in the shape ofthe accreted ice may result in a large amount of uncertainty in the aerodynamic perform-ance metrics. This paper presents a numerical approach using the technique of polyno-mial chaos expansion (PCE) to quantify icing uncertainty faster than traditionalmethods. Time-dependent bi-dimensional Reynolds-averaged Navier–Stokes computa-tional fluid dynamics (RANS-CFD) simulations are considered to evaluate the aerody-namic characteristics at the chosen sample points. The boundary conditions are based onthree-dimensional simulations of the rotor. This approach is applied to the NREL 5 MWreference wind turbine allowing to estimate the power loss range due to the leading-edgeglaze ice, considering a radial section near the tip. The probability distribution functionof the power loss is also assessed. The results of the section are nondimensionalized andassumed valid for the other radial sections. A correlation is found allowing to model theload loss with respect to the glaze ice horn height, as well as the corresponding probabil-ity distribution. Considering an equal chance for any of the ice profiles, load loss is esti-mated to be lower than 6.5%for the entire blade in half of the icing cases, while it couldbe roughly 4–6 times in the most severe icings.