Wind conditions based on coupling between a mesoscale and microscale model José Laginha Palma and Carlos Veiga Rodrigues CEsA Centre for Wind Energy and Atmospheric Flows Faculty of Engineering, University of Porto (Portugal) Presented at EWEA Technology Workshop: Resource Assessment 213 Trinity College, Dublin, (Ireland) 25 & 26 June 213 José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (1 of 15)
Contents Sites Contents 1 Wind conditions for wind resource assessment and wind power forecasting 2 Model chain comprised of: general circulation model ( 1 km) WRF, regional model (27 to 3 km) VENTOS /M, microscale model (3 2 m) 3 Comparison with measurements at 2 complex sites (14 anemometers over 7 masts) wind speed wind direction turbulence intensity IEC 614-1: Wind turbines Part 1: Design requirements shear factor José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (2 of 15)
Introduction Contents Sites Sites 1 and 2 4 masts, 2 & 4 m, 3 & 6 m 3 masts, 2 & 4 m, 3 & 6 m surface height: 3 to 115 m surface height: 12 to 132 m Jose L. Palma & Carlos V. Rodrigues Meso and microscale coupling (3 of 15)
Mesoscale Microscale Mesoscale: regional model WRF Weather Research and Forecasting model (version 3.2.1) ARW solver Initial/Boundary conditions: NCEP GDAS operational analyses NOAA RTG-SST Parameterization schemes: Microphysics: WRF Single-Moment 6-class Atmospheric radiation: Dudhia scheme for SW, RRTM for LW Cumulus parameterization: Kain-Fritsch scheme Planetary Boundary layer: ACM2 non-local closure Surface Layer: Pleim-Xiu scheme Land-Surface Model: Pleim-Xiu (2 layers) José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (4 of 15)
Mesoscale Microscale Mesoscale: domains and nesting configuration Northings ED5 UTM 29N (m) 5 48 46 44 42 4 38 d1 d2 d3 d4 d5 2 4 6 8 1 Eastings ED5 UTM 29N (m) z (m) 175 125 75 25 75 25 5 Domain, grid, resolution: d1: 24 44, 27 km d2: 43 61, 9 km d3: 46 4, 3 km d4: 34 34, 3 km d5: 34 34, 3 km Vertical levels: 49 Output: 3 minutes Time-step (seconds): 18, 6, 2 Time periods: 25, July, 1 st to 15 th Summer 25, Nov. 19 th to Dec. 3 rd Autumn José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (5 of 15)
Mesoscale Microscale Microscale: surface grid José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (6 of 15)
Mesoscale Microscale Microscale: VENTOS /M SITE 1 Grid: 47 47 44 Domain: 18 18 km 7 x : 2 to 665 m z : 3 to 922 m Time-step: 2 s Partitions: 6 Speed-up: 4. / 3.8 Resolution masts: 211 25 m SITE 2 Grid: 42 49 44 Domain: 18 21 km 7 x : 15 to 984 m z : 3 to 919 m Time-step: 2 s Partitions: 4 Speed-up: 3.1 / 5.2 Resolution masts: 166 31 m - VENTOS is a software library (EC trade mark nº 476438), used since 22 by RES and 24 by NPC. - VENTOS /M, part of the VENTOS software library, is a new addition (212), with the ability to mimic the whole range of atmospheric phenomena at a micro-scale level with a lower number of physical simplifications, compared to either meso- or microscale conventional approaches. José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (7 of 15)
Error measures Introduction Error measures Summer period Autumn period High wind speeds (14 days: 19 Nov 3 Dec 25) Turbulence intensity Shear factor Mean square error MSE = 1 N (Φ pre Φ m) 2 (1) N n=1 where Φ is wind speed or wind direction and pre and m refer to predicted and measured values Root mean square error RMSE = 1 N N (Φ pre Φ m) 2 (2) n=1 Skill score SS = 1 MSE micro MSE meso (3) < SS 1, the coupled (micro) approach is superior to meso only. SS <, the coupled (micro) approach is worse than meso only. José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (8 of 15)
Summer period: results for site 1 Error measures Summer period Autumn period High wind speeds (14 days: 19 Nov 3 Dec 25) Turbulence intensity Shear factor Site 1 micro meso 4.5 RMSE for 4. 3.5 3. 2.5 2. 1.5 1..5 3 6 9 12 15 18 21 Time of day (Hour) M11 6 m micro meso cups 16 M11 6 m 33 5 6 3 micro meso cups 14 3 4 6 3 Records (%) 12 1 8 6 4 27 24 2 1 12 9 2 2 4 6 8 1 12 14 16 18 2 Wind speed (m s 1 ) 21 Azimuth ( ) 18 15 1 2 3 4 5 6 Records (%) José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (9 of 15)
Autumn period: results for site 1 Error measures Summer period Autumn period High wind speeds (14 days: 19 Nov 3 Dec 25) Turbulence intensity Shear factor Site 1 micro meso 6 RMSE for 5 4 3 2 1 3 6 9 12 15 18 21 Time of day (Hour) M11 6 m micro meso cups 16 M11 6 m 33 35 3 3 micro meso cups 14 3 25 2 6 Records (%) 12 1 8 6 4 27 24 5 1 15 12 9 2 2 4 6 8 1 12 14 16 18 2 22 24 Wind speed (m s 1 ) 21 Azimuth ( ) 18 15 5 1 15 2 25 3 35 Records (%) José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (1 of 15)
High wind speeds (14 days: 19 Nov 3 Dec 25) 3 25 Error measures Summer period Autumn period High wind speeds (14 days: 19 Nov 3 Dec 25) Turbulence intensity Shear factor M23 4 m (1h average) micro meso cups 2 15 1 Azimuth ( ) 5 36 33 3 27 24 21 18 15 12 9 6 3 25-11-19 25-11-2 25-11-21 25-11-22 25-11-23 25-11-24 25-11-25 25-11-26 25-11-27 Time (days) 25-11-28 25-11-29 25-11-3 25-12-1 25-12-2 25-12-3 25 M14 6 m (1h average) micro meso cups 2 15 1 5 Azimuth ( ) 36 33 3 27 24 21 18 15 12 9 6 3 25-11-19 25-11-2 25-11-21 25-11-22 25-11-23 25-11-24 25-11-25 25-11-26 25-11-27 Time (days) 25-11-28 25-11-29 25-11-3 25-12-1 25-12-2 25-12-3 José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (11 of 15)
Error measures Summer period Autumn period High wind speeds (14 days: 19 Nov 3 Dec 25) Turbulence intensity Shear factor Turbulence intensity and shear factor (IEC 614-1) Neutral flow simulations: TI and α nearly independent of wind speed ( presented at EWEC 21) TI @ 65 m 3 25 2 15 1 5 all data summer day winter night cfd : σ u 2/3 k cfd : σ V 4/3 k 5 1 15 { 2 25 3 @ 65 m } 33 α [3 m, 65 m].4.3.2.1..1.2.3.4 all data summer day winter night cfd cfd mean 5 1 15 { 2 25 3 @ 65 m } 33 Turbulence intensity and shear factor have become two indicators of increasing importance, given their impact of WT lifetime and operational costs. José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (12 of 15)
Turbulence intensity: effects due to stratification Error measures Summer period Autumn period High wind speeds (14 days: 19 Nov 3 Dec 25) Turbulence intensity Shear factor Summer Site 1 M14 6 m 25-7-1T: -- 25-7-15T: 8 day Autumn Site 1 M14 6 m 25-11-19T: -- 25-12-3T: 8 day 7 night 7 night 6 5 micro [ 2/3 k, σ V ], day micro [ 2/3 k, σ V ], night 6 5 micro [ 2/3 k, σ V ], day micro [ 2/3 k, σ V ], night TI (%) 4 TI (%) 4 3 3 2 2 1 1 5 1 15 2 8 Summer Site 2 M22 6 m 25-7-1T: -- 25-7-15T: day 5 1 15 2 25 3 8 Autumn Site 2 M22 6 m 25-11-19T: -- 25-12-3T: day 7 night 7 night 6 5 micro [ micro [ 2/3 k, σ V ], day 2/3 k, σ V ], night 6 5 micro [ micro [ 2/3 k, σ V ], day 2/3 k, σ V ], night TI (%) 4 TI (%) 4 3 3 2 2 1 1 5 1 15 2 5 1 15 2 25 3 José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (13 of 15)
Shear factor: effects due to stratification Error measures Summer period Autumn period High wind speeds (14 days: 19 Nov 3 Dec 25) Turbulence intensity Shear factor Summer Site 1 M14 25-7-1T: -- 25-7-15T: Autumn Site 1 M14 25-7-1T: -- 25-7-15T:.4.4 α (3m, 6m).2..2 α (3m, 6m).2..2 cup day cup night cup day cup night.4 meso day micro day meso night micro night.4 meso day micro day meso night micro night 5 1 15 2 Summer Site 2 M22 25-7-1T: -- 25-7-15T: 5 1 15 2 Autumn Site 2 M22 25-11-19T: -- 25-12-3T:.4.4 α (3m, 6m).2..2 α (3m, 6m).2..2 cup day cup night cup day cup night.4 meso day micro day meso night micro night.4 meso day micro day meso night micro night 5 1 15 2 5 1 15 2 José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (14 of 15)
Wind speed SS > for 1 out of 14 masts Autumn has higher RMSE (3.4 against 2.3 m s 1 ) Histograms: microscale captures high velocities and shows good agreement Wind direction Microscale performs worse on overall: mostly SS =.2 1.7, higher variability, higher bias and phase error. Wind roses: microscale captures most frequent sector are better in Autumn than Summer Turbulence intensity and shear factor Stratified flow: simulations able to capture increase of TI with decrease of V h Diurnal TI agreement is better than in nocturnal period Microscale shows better agreement than mesoscale and capture nocturnal negative shear factor trend José L. Palma & Carlos V. Rodrigues Meso and microscale coupling (15 of 15)