Wind Farm Flow Modelling Using CFD
2012 Update Webinar Christiane Montavon, Ian Jones ANSYS
1 2011 ANSYS, Inc. May 8, 2012
Agenda
15:00 Introduction to the webinar and ANSYS
15.10 Presentation
Why CFD Simulation? Validation and technical advances Overview of ANSYS CFD tools for wind farm flow modelling
15.45 Software Demonstration
Automated workflow for wind farm modelling
16.05 Question and Answers 16.15 Close
2011 ANSYS, Inc.
May 8, 2012
Acknowledgements
Carbon Trust, project CT090-091, Power from Onshore Wind Farms, and partners,
RWE, Scottish Power, SgurrEnergy
Carbon Trust OWA Phase 1 and partners
(Dong, RWE, Scottish Power, SSE Renewables, Statoil, Frazer-Nash).
SSE Renewables
E.ON EU BREIN project Loughborough University University of Strathclyde... and many more
3 2011 ANSYS, Inc. May 8, 2012
ANSYS: A Comprehensive Simulation Platform
Electric Machine Blade design
Generator and shaft design
Speed Sensor
Power Electronic
Wind farm configuration for optimal power generation Transformer
Rotor sizing and acoustics
Power Distribution
Tower design and FSI
4 2011 ANSYS, Inc. May 8, 2012
Electromechanical Component
Offshore Installation and certification
Site selection, land and sea
Background
More and more onshore sites developed in complex terrain and complex forestry environment. Associated risks:
Separation, negative shear exponent factors Increased turbulence implications for turbine longevity and energy output
On such sites, standard industry tools (linearised models) used outside of the envelope where they are meant to operate
Large array losses, particularly so offshore
Empirical models tend to underestimate the losses for large arrays Atmospheric stability significantly affects array efficiency (e.g. L. Jensen, EWEC proceedings, Milan 2007)
CFD models increasingly advocated to address these issues.
Need for fidelity of solution and reliability validation! Need for automated solution for users without CFD background
5 2011 ANSYS, Inc. May 8, 2012
N-S Solvers vs. Linearised models
Ref. 2
Advantages of Navier-Stokes solvers as compared to linearized models:
Accurate prediction of turbulence: - flow turbulence is modeled or resolved using RANS/LES Better prediction of multiple-wake effects - accurate geometry description and wake prediction from multiple installations - no limit to number of wind turbines considered Separation/shade effects due to complex terrain - complex terrain is resolved - shading effects, recirculation and separation are captured
Ref. 1
Ref. 1
1. Barthelmie, R.J et al., Modelling and measurements of wakes in large wind farms, Journal of Physics: Conference Series 75 (2007) 012049. 2. Barthelmie, R.J et al., Modelling Uncertainties in power prediction offshore, IEA, Risoe, March 2004. 6 2011 ANSYS, Inc. May 8, 2012
Validation material
On range of sites
Onshore: Blacklaw, An Suidhe, Nant y Moch, Harestanes Offshore: Horns Rev, North Hoyle
On range of issues
Complex terrain Complex forestry Stability Wake interaction
Some done by our users
Offshore: Burbo Bank, Gunfleet Sands, Barrow (DONG energy) Forestry: Loughborough University
CFD delivers increased accuracy and insight in flow conditions
7 2011 ANSYS, Inc. May 8, 2012
Example Blacklaw Power Prediction
Complex forestry Significant wake effects Good prediction of normalised power output RMSE for power prediction over all turbines and over both masts is 8.5%
R. Spence, C. Montavon, I. Jones, C. Staples, C. Strachan, D. Malins, 2010, Wind modelling evaluation using an operational wind farm site, http://www.ewec2010proceedings.info/allfiles2/517_EWEC 2010presentation.pdf
2011 ANSYS, Inc.
May 8, 2012
Technical Advances
Forestry model (resistive model)
Variable forestry height Variable loss coefficient
Wake Model, large array losses
Horns Rev, North Hoyle
Atmospheric stability accounted for via equation for potential temperature, buoyancy effect in turbulence model
Harestanes, An Suidhe Horns Rev
2011 ANSYS, Inc.
May 8, 2012
Forest Canopy Model
Flow separation off forested region
Resistive Canopy Model available:
Galion Lidar data
Svensson Lopes da Costa Katul Resistance in momentum only
CFD
Canopy Input Data
From roughness data CFX Interpolation Table
variable tree heights
Height Max
Forestry loss coefficient Constant or variable with height
10
Max Forestry Loss Coef
C. Montavon, I. Jones, D. Malins, C. Strachan, R. Spence, R. Boddington, 2012, Modelling of wind speed and turbulence intensity 2011 ANSYS, Inc. May 8, 2012 EWEA 2012, Copenhagen. for a forested site in complex terrain,
Horns Rev Results at Hub Height Sector 280
Uref = 8 m/s at 70m, z0 = 0.0002m Wind direction: sector 280 Horizontal velocity Turbulence intensity
11
2011 ANSYS, Inc.
May 8, 2012
Horns Rev Normalised Power Down a Row
10m/s 270 2 bin
1.2 Normalised Power 1 0.8 0.6 0.4 0.2 0 1 2 3 4 5 6 7 8 Turbine Group 9 10 ANSYS CFD
Simulations by step of 1 degree, sector 270 285, averaged for three different bin sizes. Reasonably good prediction
Tendency for over-estimation of array losses Good prediction of slope down the row
Consistent for various bin sizes
10m/s 270 10 bin
1.2 ANSYS CFD Normalised Power Normalised Power 1 Data - UpWind 1 1.2
10m/s 270 30 bin
ANSYS CFD Data - UpWind
0.8
0.6 0.4 0.2 1 2 3 4 5 6 7 8 9 10 Turbine Group
0.8
0.6 0.4 1 2 3 4 5 6 7 8 9 10 Turbine Group
Upwind data from Wake Measurements Used in the Model Evaluation. K.S. Hansen, R. Barthelmie, D. Cabezon and E. Politis. Upwind Wp8: Flow; Deliverable D8.1 Data. 18 June 2008.
12 2011 ANSYS, Inc. May 8, 2012
North Hoyle Normalised Power Down a Row
Uref = 10 m/s at 67m, z0 = 0.0001m, upstream TI = 7% Wind direction: sector 260
Very good agreement with power data for both bin sizes Absolutely blind test case!
10 deg bin
1.2 1.2 1 0.8 0.6 ANSYS CFD 0.4 0.2 0 1 2 3 4 5 Measured Data Upper 25% Lower 25% 0 1 2 3 4 5 1 0.8 0.6 ANSYS CFD 0.4 0.2 Measured Data Upper 25% Lower 25%
30 deg bin
Normalised Power
Column
Normalised Power
Column
C. Montavon, S.-Y. Hui, J. Graham, D. Malins, P. Housley, E. Dahl, P. de Villiers, B. Gribben, 2011, Offshore Wind Accelerator: wake modelling using CFD, EWEA 2011, Brussels.
13 2011 ANSYS, Inc. May 8, 2012
Improved results with atmospheric stability
Atmosphere on average stable above boundary layer Including surface stability effects significantly affects the prediction of array losses
Including this effect
Changes the relative distribution of
the wind speed between hill tops and valleys improves mast to mast cross prediction when masts located in different type of positions (i.e. hill tops vs valleys) Improves the prediction of the relative TI on site
C. Montavon, C. Staples, C. Weaver, 2011, Simulating the flow conditions over complex terrain with RANS models: sensitivity to a selection of parameters including atmospheric stability , EWEA 2011, Brussels. C. Montavon, I. Jones, D. Malins, C. Strachan, R. Spence, R. Boddington, 2012, Modelling of wind speed and turbulence intensity for a forested site in complex terrain, EWEA 2012, Copenhagen.
14 2011 ANSYS, Inc. May 8, 2012
Harestanes, Masts on site
4 masts on site, 3 near hill tops, 1 in a location with difficult flow conditions (valley, proximity to forestry)
Cross prediction done for 3 masts with long concurrent time series
All masts with data at 70m, 60m, 40m, 30m Hill top masts: Holehouse Hill Hareshaw Rig Valley mast: Bran Rig Bran Rig
15 2011 ANSYS, Inc. May 8, 2012
Hareshaw Rig
Holehouse Hill
Harestanes Mast to Mast Cross Prediction (wind speed)
Maximum relative errors in wind speed cross predictions for three model configurations:
1. Purely neutral
2. Conventionally neutral i.e. stable conditions in free stream, with potential temperature gradient of US standard atmosphere, and neutral conditions at ground (adiabatic).
Model neutral stable
3 masts (70m only) 3 masts (all heights) BRAN-HOL BRAN-HAR HOL-HAR BRAN-HOL BRAN-HAR HOL-HAR 11.8% 0.4% 13.3% 4.1% 1.8% 5.4% 24.0% 2.9% 24.0% 6.5% 3.2% 6.2%
Improved results with atmospheric stability
16 2011 ANSYS, Inc. May 8, 2012
Validated Tools
Ongoing validation exercise together with end users, see e.g. our joint EWEA publications at Copenhagen 2012, available from:
https://docs.google.com/open?id=0B6Cp_fvx8o5Dei0tenVqelZZQzQ
Brussels 2011, available from:
https://docs.google.com/leaf?id=0B6Cp_fvx8o5DMjBkMzQxNWMtOTk xYy00ZmFiLWFlNzMtZGFhM2U0NWQ1ODFh&hl=en_GB
Previous years, available from:
https://docs.google.com/leaf?id=0B6Cp_fvx8o5DZmEzYzYxMDAtZGYxZi 00YmI0LWIwMjYtZWM3NmViYWM2NDI3&hl=en_GB
17
2011 ANSYS, Inc.
May 8, 2012
ANSYS CFD Featured In Independent Publications
Beaucage, P., Robinson, N., Brower, M., Alonge, C., Overview of six commercial and research wake models for large offshore wind farms, Proceedings EWEA 2012, Copenhagen. Garza, J., A. Blatt, R. Gandoin and S.-Y. Hui (2011) Evaluation of two novel wake models in offshore wind farms . Proceedings from the EWEA Offshore conference, 29 Nov. - 1 Dec 2011. Desmond, C., Sayer, A., Watson, S., Hancock, P., Forest Canopy Flows in Non-Neutral Stability , EWEA 2012, Copenhagen. (poster award).
Clive, P., Dinwoodie, I., Quail, F., Direct measurement of wind turbine wakes using remote sensing, Proceedings EWEA 2011, Brussels.
18 2011 ANSYS, Inc. May 8, 2012
WindModeller Tools for Automated Solution
WindModeller: set of tools wrapped around ANSYS standard CFD products:
Allow non-CFD experts to perform wind farm analyses in
automated way
Drive ANSYS CFX or FLUENT flow solver Allow advanced user to encapsulate their own expertise
(access to customised setup and post-processing scripts which can easily be altered by the user to further develop the tools )
19
2011 ANSYS, Inc.
May 8, 2012
Tools for Automated Solution
Objective Automation of Analysis and Data
extraction Map Mesh CFD Report in
one step CFD solution + automated post-processing
20
2011 ANSYS, Inc.
Wind dataMay transposition module (cross prediction and energy assessment) 8, 2012
WindModeller: Simulation Process
Wind farm simulation process from user perspective
Set up analysis on desktop computer (either via GUI or command line) Submit job to:
Run possible large number of cases on the local machine or on a remote server Postprocess results to automatically generate reports/summary data files
Possibility to perform additional post-processing on individual results files using CFD
Post
Setup on desktop
21 2011 ANSYS, Inc. May 8, 2012
Run on local or remote computer
Report as html file
Meshing Approach
Current recognised terrain format SRTM, Shuttle Radar Topography
Mission, freely available, 90m resolution (finer resolution in the US) NTF, National Transfer Format, contour data (UK) .map files (WAsP format) Generic point data file (.csv)
Terrain converted to tesselated format (STL) Meshing with custom tools Fixed mesh structure, hexahedral mesh
22
(5 or 9 blocks), aimed at process automation Template mesh morphed onto STL terrain representation Variable mesh topology for elongated/twin wind farms
May 8, 2012
2011 ANSYS, Inc.
Varying Mesh Topologies
Compact wind farm
Elongated wind farm
Twin wind farms
23 2011 ANSYS, Inc. May 8, 2012
Meshing Controls
User can prescribe:
horizontal resolution in central
region Rate of horizontally expansion outside first layer cell heights in vertical Good control of mesh resolution in lower heights (ensures appropriate resolution is achieved in forested regions) Smooth vertical expansion above
24
2011 ANSYS, Inc.
May 8, 2012
Mesh Adaption on Wind Turbine Rotor
Improve resolution by automatically refining mesh around the turbine location, from the specification of the rotor location and actuator disk parameters only
Automatically enabled if wake model is used.
Initial mesh
1st refinement 2nd refinement Final mesh
25
2011 ANSYS, Inc.
May 8, 2012
Setup
Outer surface divided into 24 regions
12 for inlet b.c. (Dirichlet on
velocity) 12 for outlet b.c. (entrainment conditions with prescribed static pressure)
Setup automated to run for e.g. 12 wind directions Selection of surfaces defining inlet/outlet automated in script running cases for various wind directions meshing done only once
26 2011 ANSYS, Inc. May 8, 2012
Flow Modelling in WindModeller
Atmosphere modelled as:
incompressible fluid (Air at 15C) assuming neutral stability solving for steady state RANS Shear Stress Transport (SST) turbulence model or k- .
Turbulence modelled via two-equation model
Ground modelled as rough wall (spatially variable roughness) Inlet boundary conditions
Classical constant-shear ABL profiles (Durbin & Petterson Reif ):
z u ln( ) z0 u*
2 u* k 1/ 2 C
3 u* z
Additional physics:
Forest canopy model (resistive term in momentum equation + additional source
terms in turbulence model) Multiple wake model (actuator disk model) Atmospheric stability as beta feature
27
2011 ANSYS, Inc.
May 8, 2012
Wake Modelling
Hierarchy of Wake Models available in ANSYS CFD
Resolved blade models
Virtual Blade Models Simple Actuator Disk Models
Provide practical model for
calculations with many turbines Input is turbine thrust curve, turbine diameter, turbine hub height Provides momentum sink in cylindrical volume surrounding each turbine Basis of Models for WindModeller
28
2011 ANSYS, Inc.
May 8, 2012
Simple Wake Model
Wind turbine represented by momentum sink constant thrust per volume within identified rotor disk. Wind turbine orientation parallel to wind direction at inlet Works on any type of mesh, although it is expected that the best results will be obtained with resolution that captures the wind turbine disk reasonably well User input: Coordinates of hub location WT diameter WT thrust and power curve
29
2011 ANSYS, Inc.
May 8, 2012
Post-Processing in WindModeller
As part of the automated approach WindModeller can generate:
Plots of streamlines Identification of recirculation zones Plots at constant height AGL and
profiles at wind turbine/mast locations for quantities such as normalised velocity, turbulence intensity, shear exponent factor Exported data tables of similar quantities at wind turbine/mast locations Export to Google Earth (.kml files) Automated report in html format
30
2011 ANSYS, Inc.
May 8, 2012
Post-Processing in WindModeller
As part of the automated approach WindModeller can generate:
Plots of streamlines Identification of recirculation zones Plots at constant height AGL and
profiles at wind turbine/mast locations for quantities such as normalised velocity, turbulence intensity, shear exponent factor Exported data tables of similar quantities at wind turbine/mast locations Export to Google Earth (.kml files) Automated report in html format
31
2011 ANSYS, Inc.
May 8, 2012
Post-Processing in WindModeller
As part of the automated approach WindModeller can generate:
Plots of streamlines Identification of recirculation zones Plots at constant height AGL and
profiles at wind turbine/mast locations for quantities such as normalised velocity, turbulence intensity, shear exponent factor, flow angles Exported data tables of similar quantities at wind turbine/mast locations Export to Google Earth (.kml files) Automated report in html format
32
2011 ANSYS, Inc.
May 8, 2012
Post-Processing in WindModeller
As part of the automated approach WindModeller can generate:
Plots of streamlines Identification of recirculation zones Plots at constant height AGL and
profiles at wind turbine/mast locations for quantities such as normalised velocity, turbulence intensity, shear exponent factor , flow angles Exported data tables of similar quantities at wind turbine/mast locations Export to Google Earth (.kml files) Automated report in html format
33
2011 ANSYS, Inc.
May 8, 2012
Post-Processing in WindModeller
As part of the automated approach WindModeller can generate:
Plots of streamlines Identification of recirculation zones Plots at constant height AGL and
profiles at wind turbine/mast locations for quantities such as normalised velocity, turbulence intensity, shear exponent factor , flow angles Exported data tables of similar quantities at wind turbine/mast locations Export to Google Earth (.kml files) Automated report in html format
34
2011 ANSYS, Inc.
May 8, 2012
Post-Processing in WindModeller
As part of the automated approach WindModeller can generate:
Plots of streamlines Identification of recirculation zones Plots at constant height AGL and
profiles at wind turbine/mast locations for quantities such as normalised velocity, turbulence intensity, shear exponent factor , flow angles Exported data tables of similar quantities at wind turbine/mast locations Export to Google Earth (.kml files) Automated report in html format
35
2011 ANSYS, Inc.
May 8, 2012
Post-Processing in WindModeller
As part of the automated approach WindModeller can generate:
Plots of streamlines Identification of recirculation zones Plots at constant height AGL and
profiles at wind turbine/mast locations for quantities such as normalised velocity, turbulence intensity, shear exponent factor , flow angles Exported data tables of similar quantities at wind turbine/mast locations Export to Google Earth (.kml files) Automated report in html format, including the above
May 8, 2012
36
2011 ANSYS, Inc.
Post-Processing in WindModeller
As part of the automated approach WindModeller can generate:
Plots of streamlines Identification of recirculation zones Plots at constant height AGL and
profiles at wind turbine/mast locations for quantities such as normalised velocity, turbulence intensity, shear exponent factor , flow angles Exported data tables of similar quantities at wind turbine/mast locations Export to Google Earth (.kml files) Automated report in html format, including the above
May 8, 2012
37
2011 ANSYS, Inc.
Wind Data Transposition Module
Simulations establish climatological relationships between wind conditions at mast (reference site) and WTG (predicted site). Simulations are performed independently from the data collection at the mast. Wind conditions predicted at WTG
Data collected at mast
Wind data transposition module
38 2011 ANSYS, Inc. May 8, 2012
Energy Assessment/Cross Prediction
Wind data for input: Time series or Frequency tables Allows for multiple masts and with / without wake calculations Can deal with heterogeneous wind farm Masts data collected before or after wind turbines installed Possibility to do a postmortem on an existing wind farm to understand performance issues of single WT if mast still measuring when wind farm is operational
Note: masts must be within simulation domain (no MCP)
39 2011 ANSYS, Inc. May 8, 2012
Output from Data Transposition
Tables of Capacity Factors (by directions and overall) summarising the average annual energy output at each WT. Wind speed distributions (WAsP .tab files) at WT and masts. Resource file (WAsP .rsf) at WT locations. Summary table with average wind speed at masts from cross prediction. Summary tables of mean and representative turbulence intensity by wind speed classes at masts and WT locations. (when working from time series, including the wind speed standard deviation as input).
40 2011 ANSYS, Inc. May 8, 2012
Availability of WindModeller Tools
Customised tools based on standard ANSYS CFD software under continuous development Driven by customer and project demands Used on many cases Made available to customer on service based approach via two stage process: 1st phase: demonstration of capability on terrain chosen by customer 2nd phase: Technology Transfer. Tools made available to customer, also includes one-to-one training, support and maintenance of the tools after delivery. Develop features on request.
41
2011 ANSYS, Inc.
May 8, 2012
Ongoing Projects
OWA Phase II: integration of atmospheric stability and associated effects on large array losses Carbon Trust POWFARMM project: complex terrain, forestry, wakes, comparison with masts and Galion LIDAR data Various consultancy projects recently completed for customers
Pollution transport Integration of buildings More complex stability conditions (e.g. strong inversions,
coastal low level jets)
Diversification into modelling of marine arrays (TideModeller)
42 2011 ANSYS, Inc. May 8, 2012
Summary
CFD delivers increased accuracy for
Complex terrain/complex forestry cases Estimation of large array losses
Atmospheric stability improves accuracy compared to
neutral cases, which tend to be used by other CFD packages in the industry ANSYS has a suite of tools (WindModeller) that helps you automate the simulation process, with state of the art models for
Forestry Wakes Atmospheric stability
Validation material available to attest this.
43 2011 ANSYS, Inc. May 8, 2012
Upcoming ANSYS Events
ANSYS Events
Webinar - Advanced Multi-body Hydrodynamics and Motion Analysis Using AQWA Software - 2012 Update
Scheduled for 2012 Information: http://www.ansys.com/events
Friday, 11th May 2012 http://www.ansys.com/aqwawebinar2012
All Energy in Aberdeen on May 23 & 24 stand C111.
Question, Comments, Inquiries
[email protected]
44
2011 ANSYS, Inc.
May 8, 2012
