CFD for Wind and Tidal Offshore Turbines

This work presents analytical estimates for various flow scales encountered in cross-flow turbines (i.e. Darrieus type or vertical axis) for renewable energy generation (both wind and tidal). These estimates enable the exploration of spatial or temporal interactions between flow phenomena and provide quantitative and qualitative bounds of the three main flow phenomena: the foil scale, the vortex scale and wake scale. Finally using the scale analysis, we show using an illustrative example how high order computational methods prove beneficial when solving the flow physics involved in cross-flow turbines.

This work presents a 2D study of vertical axis turbines with application to wind or tidal energy production. On the one hand, a degree-adaptive Hybridizable Discontinuous Galerkin (HDG) method is used to solve this incompressible Navier–Stokes problem. The HDG method allows to drastically reduce the coupled degrees of freedom (DOF) of the computation, seeking for an approximation of the solution that is defined only on the edges of the mesh. The discontinuous character of the solution provides an optimal framework for a degree-adaptive technique. Degree-adaptivity further reduces the number of DOF in the HDG computation by means of degree-refining only where more precision is needed. On the other hand, the finite volume method of ANSYS^® is used to validate and compare the obtained results.

The sliding mesh approach is widely used in numerical simulation of turbomachinery flows to take in to account the rotor/stator or rotor/rotor interaction. This technique allows relative sliding of one grid adjacent to another grid (static or in motion). However, when a high-order method is used, the interpolation used in the sliding mesh model needs to be of, at least, the same order than the numerical scheme, in order to prevent loss of accuracy. In this work we present a sliding mesh model based on the use of Moving Least Squares (MLS) approximations. It is used with a high-order ( > 2) finite volume method that computes the derivatives of the Taylor reconstruction inside each control volume using MLS approximants. Thus, this new sliding mesh model fits naturally in a high-order MLS-based finite volume framework (Cueto-Felgueroso et al., Comput Methods Appl Mech Eng 196:4712–4736, 2007; Khelladi et al., Comput Methods Appl Mech Eng 200:2348–2362, 2011) for the computation of acoustic wave propagation into turbomachinery.

Vertical axis wind turbine (VAWT) start-up is a highly non-linear process, with turbines experiencing a long idling period of low acceleration before a sudden increase in rotational velocity to a final equilibrium state. The physics of start-up behaviour is not well understood, with some analyses showing VAWTs to be incapable of self-start, in spite of experimental and field evidence to the contrary. This study attempts to assess the impact of blade–wake interactions on start-up behaviour.

A high-order discontinuous Galerkin (DG) solver was used to simulate the flow around a VAWT, with experimentally obtained acceleration applied to the turbine, forcing it from standing to a tip-speed ratio (TSR) of 2 in a typically non-linear pattern. Additionally, the DG code was run with a constant TSR of 1, a TSR at which blade–wake interactions are known to occur. A blade-element momentum (BEM) method was used to provide an additional simulation of the turbine at this speed. Since the DG code captures blade–wake interactions while the BEM does not, a comparison of the outputs of the two was made to assess the impact of the interactions on blade forces.

Vertical Axis Tidal Turbines (VATTs) are an innovative way of harnessing renewable energy from tidal streams. Herein a novel numerical approach using a refined Large Eddy Simulation (LES) code to simulate the performance of a VATT is presented. The turbine blades are modelled with Lagrangian markers using the Immersed Boundary Method which offers several advantages especially concerning computational effort. Comparisons of the LES results with experimental and numerical data suggest reasonably good accuracy of the code. In addition, the stability of the method for high Reynolds number flows is also discussed.

The present study discusses two-dimensional numerical simulations of a cross-flow vertical-axis marine (Water) turbine (straight-bladed Darrieus type) with particular emphasis on the turbine unsteady behavior. Numerical investigations of a model turbine were undertaken using commercial computational solvers. The domain and mesh were generated using a glyph script in POINTWISE-GRIDGEN, while the simulations were performed in ANSYS-FLUENT v14. For the simulation, a sliding mesh technique was used in order to model the rotation of the turbine; a shear stress transport k–ω turbulence model was used to model the turbulent flow. In order to simulate the interaction between the dynamics of the flow and the Rigid Body Dynamics (RBD) of the turbine a User Define Function (UDF) was generated. The primary turbine operational variables of interest were the evolution of torque, power, and runaway speed. Numerical results show that as the freestream velocity is increased, the runaway angular speed of the turbine increases, which is consistent with the observation that the frequency of oscillation of the angular velocity (in the quasi steady-state) increases as the freestream velocity also increases. For a given turbine, it was observed that the increment in the moment of inertia of the turbine does not influence the average value of the runaway angular velocity (quasi-steady state) but causes an increase in the time taken for achieving this quasi steady-state.

In recent years there is a renewed interest in both large-scale and small-scale vertical axis wind turbines (VAWT). The development of multi-megawatt floating offshore VAWT is mainly a response to a plateau in the improvement of the aerodynamic performance of horizontal axis wind turbines. The research in small rotors (<100 kW) is motivated by the future demand for a decentralized sustainable energy supply in remote areas. However, such designs have received much less attention than the more common propeller-type designs and the understanding of some aspects of their operation remains, to this day, incomplete. This holds, because after some authors the starting operation is difficult, if not impossible, to induce the self-start capabilities without external assistance. This paper reviews the cause of the inability of the low solidity fixed pitch vertical axis wind turbines to self-start, and investigates the flow physics of dynamic stall in order to comprehend, interpret, understand and explain-all this comprising the problem of start-up.

This paper describes a new development aiming to deform multi-block structured viscous meshes during fluid–solid interaction simulations. The focus is put on the deformation of external aerodynamic configurations accounting for large structural displacements and 3D multi-million cells meshes. In order to preserve the quality of the resulting mesh, it is understood as a fictitious continuum during the deformation process. Linear elasticity equations are solved with a multigrid and parallelized solver, assuming a heterogeneous distribution of fictitious material Young modulus. In order to improve the efficiency of the system resolution an approximate initial solution is obtained prior to the elastic deformation, based on Radial Basis Functions and Transfinite interpolators. To validate the performances of the whole algorithm, the DTU-10MW reference offshore wind turbine described by Bak et al. is analyzed (Description of the DTU 10 MW reference wind turbine. Technical report. Technical University of Denmark Wind Energy, Roskilde, 2013).

This chapter presents numerical simulations of water waves using the Finite Volume Method. Wave loads exerted on a truncated circular cylinder are calculated and compared to experimental data.

The method is validated on two test cases, both regarding truncated circular cylinder. The first test case considers maximum regular wave loads with different frequencies and wave heights. The second test case simulates phase focused freak wave and its impact on the cylinder.

Finite Volume Methods were successfully used in the last years to solve differential hyperbolic problems (Leveque, Finite-volume methods for hyperbolic problems. Cambridge University Press, Cambridge, 2005). Our research is focused here on the use of the Moving Least Squares (MLS) approximations for the development of a selective limiting technique to keep the accuracy of high-order methods in non-smooth flows. Following (Nogueira et al., Comput Methods Appl Mech Eng 199:2544–2558, 2010) we use the multiresolution properties of the MLS methodology and we define a shock-detection technique to act as a smoothness indicator. This sensor is used to detect shock waves present in the flow problem. The use of this technique combined with slope limiters improves the accuracy of the resulting TVD scheme. In this work we present the first results obtained with this technique applied to the resolution of the shallow waters equations with a high-order FV-MLS scheme (Cueto-Felgueroso et al., Comput Methods Appl Mech Eng 196:4712–4736, 2007). We present several 1D results and we compare them with those obtained with other high-order schemes.

In this work, a comparison between results of a panel method and a RANS solver is made for a horizontal axis marine current turbine in uniform inflow conditions. The panel method calculations were made with panel code PROPAN (Baltazar, On the modelling of the potential flow about wings and marine propellers using a boundary element method. Ph.D. thesis, Instituto Superior Técnico, 2008). A vortex pitch wake alignment model is considered for the blade wake. The RANS calculations were carried out with RANS code ReFRESCO (http://​www.​marin.​nl/​web/​Facilities-Tools/​CFD/​ReFRESCO.​htm). A comparison of the blade pressure distributions, wake geometry and thrust and power coefficients is made. A reasonable to good agreement of the blade distribution and turbine forces is seen between the codes. Comparison of the numerical calculations with experimental performance data available in the literature is also presented. In general, the trend of the thrust and power coefficients is well captured by the numerical methods.