Three dimensional actuator disc modelling of wind farm wake interaction S Ivanell 1,2, R Mikkelsen 3, J N Sørensen 3, D Henningson 1 1 Linné Flow Centre, KTH Mechanics, Stockholm, Sweden 2 Gotland University, Visby, Sweden 3 Technical University of Denmark, Lyngby 1(20)
Background Wake structure Experiments Simulation methods Park simulations 2(20)
Project background Project layout Numerical simulation of the wake behind one turbine. Stability study, study of basic mechanisms behind breakdown of the wake Park simulations 3(20)
Wake structure [Montgomerie, Dahlberg] 4(20)
Available Wake Measurement Data [Risø] DNW WT i Holland 5(20)
CFD Simulation EllipSys3D a 3D solver developed by DTU and RISØ. (Cooperation with DTU - The Technical University of Denmark) A general purpose 3D solver developed by: Jess Michelsen (DTU/Risø) & Niels Sørensen (Risø) Multi block, Finite volume, Collocated storage arrangement, MPI Special technique for wind power simulation The Actuator Line Method (ACL) or the Actuator Disc Method (ACD) LES (Mixed sub-grid-scale model by Ta Phuoc) 6(20)
ACL/ACD Metod 7(20)
Park simulations with the ACD method 8(20)
Mesh 9(20)
Prescribed wind profile y Power law profile Δ Parabolic profile h hub 10(20)
Pre-generated turbulens 11(20)
9 turbine park simulation 12(20)
Velocity at hub height 13(20)
Wake interaction 0 degree inflow angle 15 degree inflow angle 14(20)
Mesh size dependence 15(20)
Difference between simulation with and without turbulence. 16(20)
Conclusions The result clearly indicates that turbulent inflow has a strong impact on the result and leads to a more complex flow structure. The result also indicates a more realistic flow structure, i.e. realistic local velocities. Based on this study we are however far from being able to conclude how dependent the level of production is on the turbulence intensity. Further investigations on how accurately this method captures the complex flow in real wind farms are needed. Future work will concentrate on further development of the method and setting up simulations corresponding to a real wind turbine farm where measurement data is available for verification. 17(20)
Off shore 18(20)
Off shore Perioidic b.c. Perioidic b.c. 19(20)
On shore 20(20)
PART 2 - Stability study by introducing harmonic perturbations together with the ACL method 21(20)
40 block 120 periodic mesh, ~5 10^6 nodes 22(20)
Vortex spiral 23(20)
Iso Omega; 2Hz, 6Hz 24(20)
Extracting fields 25(20)
Fourier analysis 26(20)
Identifying spiral 27(20)
Development of disturbance 28(20)
Development of disturbance 29(20)
Disturbance amplitude 30(20)
Growth rate 31(20)
Re(Phi); 2Hz, 5Hz 32(20)
Modes 33(20)
Modes 34(20)
Modes 35(20)
Modes 2Hz 5Hz 36(20)
Modes, 2Hz 37(20)
Unfolded view, 2Hz, 5Hz 38(20)
Direction of velocity component 39(20)
Conclusions part 2 The result show that the instability is dispersive and that growth arise only for some specific frequencies and type of modes. The result indicated two types of modes, one where oscillations of every other vortex spiral is out of phase, one with oscillations in every vortex spiral in phase. The result show that the mode with spirals out of phase results in largest growth. Main extension is in radial and flow direction (phase shift of 180 ). Downstream the non-linear development of the instability result in vortex pairing. 40(20)