A Comparative Study of the Wake Dynamics during Yaw and Curtailment. - PowerPoint PPT Presentation

A Comparative Study of the Wake Dynamics during Yaw and Curtailment. Søren Juhl Andersen 1 June 20, 2019 Email: 1 sjan@dtu.dk Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 1 / 22

Motivation Two main strategies are generally be- ing investigated for improved wind farm control to mitigate the adverse wake ef- fects. Yawing a turbine, which aim to deflect the wake away from the next turbine. Derating a turbine, which aims to decrease the wake deficit. Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 2 / 22

Motivation The present work aims to provide a di- rect and completely equivalent compar- ison of the two operating conditions. The study is focussing on: Turbine operation Wake dynamics Potential power gain with two turbines Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 2 / 22

Flow Solver and Turbine Modeling EllipSys3D 1 , 2 Actuator Line 3 Finite volume and incompressible Apply body forces along rotating lines Multigrid and multiblock Tabularised lift and drag coefficients Parallelized with MPI Fully coupled to Flex5 4 , which is a Large Eddy Simulation(LES) modal based aero-elastic tool Dynamic turbine controller. 1 Michelsen, 1992, 2 Sørensen, 1995, 3 Sørensen and Shen, 2002, 4 Øye, 1996 Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 3 / 22

Turbine and Inflow Wind Turbine Inflow V27, comparable to the V27 at U hub = 8 . 3283 m / s SWiFT 1 TI hub = 7 . 7% R = 13 . 5 m Shear: α = 0 . 18 Z hub = 32 m 1 Resor and LeBlanc, 2014 Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 4 / 22

Effect of Control Strategies Using Flex to match streamwise thrust force for derating and yawing. a) a) 25 25 20 20 T [ kN ] T [ kN ] 15 15 Yaw5 EllipSys3D Derate5 Flex5 10 10 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 t [ s ] t [ s ] b) b) 25 25 20 20 T [ kN ] T [ kN ] 15 15 Yaw1 EllipSys3D Derate1 Flex5 10 10 0 200 400 600 800 1000 1200 0 200 400 600 800 1000 1200 t [ s ] t [ s ] Comparison of thrust force computed in Flex5 for yaw and Comparison of thrust force computed in Flex5 and EllipSys3D as function of time for a) φ = 0 ◦ and b) derating. φ = − 35 ◦ . Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 5 / 22

Effect of Control Strategies Yawing is penalised more serverly Matching( < 1%) mean thrust force: in terms of power production: 21 1 Yaw Derate Flex5 Derate Yaw Flex5 Normal Operation Derate EllipSys Yaw EllipSys 1:1 20 0.9 19 P T [ kN ] P 0 [] 0.8 18 0.7 17 16 0.6 -40 -30 -20 -10 0 10 20 30 40 0.6 0.7 0.8 0.9 1 T γ [ ◦ ] T 0 [] Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 6 / 22

Effect of Control Strategies Comparing Power of 1st Turbine from Flex5 and EllipSys3D. Total simulation time is 20 min. a) 160 120 P [ kW ] 80 EllipSys3D Flex5 40 0 200 400 600 800 1000 1200 t [ s ] b) 160 120 P [ kW ] 80 EllipSys3D Flex5 40 0 200 400 600 800 1000 1200 t [ s ] Comparison of electrical power computed in Flex5 and EllipSys3D as function of time for a) φ = 0 ◦ and b) φ = − 35 ◦ . Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 7 / 22

Average Wakes Comparing normal operation and yawing − 35 ◦ and the corresponding derating. U U hub [ R ] Yawing Normal [ R ] operation Derating [ R ] X [ R ] Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 8 / 22

Wake Dynamics Comparing wake at 6 R downstream Derating Normal Operation Yawing Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 9 / 22

Center of Wake Movement Comparing normal operation and yawing − 35 ◦ and the corresponding derating. y C [ R ] X [ R ] Subtracting median. y ∗ C [ R ] X [ R ] Large meandering for yawing and derated wakes in near wake Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 10 / 22

Wake Breakdown Wake breakdown location 1 : � l � �� 16 � � �� ln � � + 5 . 5 ln � U 3 c = − 0 . 3 T i T i . R N b λ C T nw » However, what dominates: Breakdown or smaller deficit? Using Proper Orthogonal Decomposition 1 to assess wake breakdown location from minima of 1st mode. 20 Derate Baseline 19 Yaw 18 � l � 17 nw [ R ] R 16 15 14 13 12 16 16.5 17 17.5 18 18.5 19 19.5 20 20.5 21 T [ kN ] Yawed wakes break down faster 1 Sørensen et al., Royal Society, 2015 Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 11 / 22

Power Gain of 2 Turbines - Example 1st turbine in EllipSys3D, while 2nd turbine is Flex5. X = 5 R and yawing 15 ◦ Baseline Operation: P tot = 109 + 36 = 145kW 150 100 P [kW] 50 0 0 200 400 600 800 1000 1200 Yaw Operation: P tot = 84 + 56 = 140kW 150 100 P [kW] 50 0 0 200 400 600 800 1000 1200 P 1 + P 2 250 200 P [kW] 150 100 50 0 200 400 600 800 1000 1200 t [s] Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 12 / 22

Mean Power Gain of 2 Turbines Estimating power gain of two turbines by summing the difference in instantaneous power production of controlled turbine and baseline turbine. Normalising by sum of instantaneous power production of baseline turbines. Then taking the time average to get the mean power gain. � � ( P 1 , E , control − P 1 , E , base ) + ( P 2 , F , control − P 2 , F , base ) < ∆ P tot > = 100 · ( P 1 , E , base + P 2 , F , base ) 2 2 0 0 -2 -2 P tot > [%] P tot > [%] -4 -4 -6 -6 -8 -8 < < -10 -10 -12 -12 -14 -14 0 5 10 15 20 0 5 10 15 20 X [R] X [R] Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 13 / 22

Distribution of Total Power Production φ = − 35 ◦ φ = − 30 ◦ φ = − 15 ◦ φ = − 5 ◦ X = 5 R X = 10 R Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 14 / 22

Distribution of Total Power Production Conclusions and Discussions Yawing is penalized harder in terms of power production than derating. Reducing streamwise thrust increases wake movement/meandering, particular for short distances behind the turbine Additionally, yawing does not only deflect the wake, but potentially also initiates the beneficial breakdown sooner. Neither yawing nor derating appears beneficial for this scenario. However, yawing has a distinct peak around 3 − 7 R downstream in terms of potential power increase, i.e. the wake needs to deflect sufficiently before and at the same time exploit earlier recovery Derating has a more constant potential in the far wake, but larger at closer distances( X < 10 R ) However, the overall benefit for a two turbine system is still largely uncertain. Particular as yawing involves significantly higher uncertainties than derating. Narrow distributions when derating, and very wide distributions for yaw. Is this actual flow control? Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 15 / 22

Distribution of Total Power Production Acknowledgements CCA on Virtual Atmosphere PossPow and CONCERT TotalControl Thanks for your attention. Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 16 / 22

Distribution of Total Power Production Acknowledgements CCA on Virtual Atmosphere PossPow and CONCERT TotalControl Thanks for your attention. Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 17 / 22

Distribution of Total Power Production Extra Slide: Comparing EllipSys3D and Flex5 2nd Turbine a) 200 EllipSys3D 160 Flex5 P [ kW ] 120 80 40 0 0 200 400 600 800 1000 1200 t [ s ] b) 200 EllipSys3D 160 Flex5 P [ kW ] 120 80 40 0 0 200 400 600 800 1000 1200 t [ s ] Comparison of electrical power of 2nd turbine at a downstream distance of 10 R computed in Flex5 and EllipSys3D for a) φ = 0 ◦ and b) φ = − 30 ◦ . For EllipSys3D, it is the 2nd turbine of 4, i.e. added blockage. Mean difference in power is a) approx. 4% and b) approx. 6%. Could correct using Troldborg and Forsting, Wind Energy, 2017, 20:2011-2020 and perhaps Mann et al. Wind Energy Science, 3, 293-300, 2018. Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 18 / 22

Power Gain 2nd Turbine Power gain of 2nd turbine P2 P2 30 14 12 25 10 20 8 15 6 4 10 2 5 0 0 -2 0 5 10 15 20 0 5 10 15 20 Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 19 / 22

Potential Power Gain Estimating power gain of 2nd turbine using time and rotor averaged velocity in wake( < U > ) compared to normal operation( U NO ) normalized by flow with no first turbine U 0 . ∆ P 1 is power loss on 1st turbine. P 2 = 100 · < U > 3 − < U NO > 3 ∆ � < U 0 > 3 γ = − 35 ◦ γ = − 30 ◦ γ = − 25 ◦ γ = − 15 ◦ γ = − 5 ◦ ∆ � P 2 [%] X [ R ] X [ R ] X [ R ] X [ R ] X [ R ] Andersen A Comparative Study of the Wake Dynamics during Yaw and Curtailment. June 20, 2019 20 / 22