wind farm flow modelling using cfd - ansys … uk...wind farm flow modelling using cfd 2012 update...

© 2011 ANSYS, Inc. May 8, 2012 1

Wind Farm Flow Modelling Using CFD

2012 Update Webinar Christiane Montavon, Ian Jones

ANSYS

© 2011 ANSYS, Inc. May 8, 2012 2

• 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

Agenda

© 2011 ANSYS, Inc. May 8, 2012 3

• 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

Acknowledgements

© 2011 ANSYS, Inc. May 8, 2012 4

Site selection, land and sea

Generator and shaft design

Wind farm configuration for optimal power generation

Blade design

Rotor sizing and acoustics

Tower design and FSI

Offshore Installation and certification

ANSYS: A Comprehensive Simulation Platform

Electric Machine

Transformer

Power Distribution

Speed Sensor

Electromechanical Component

Power Electronic

© 2011 ANSYS, Inc. May 8, 2012 5

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

Background

© 2011 ANSYS, Inc. May 8, 2012 6

N-S Solvers vs. Linearised models

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. 2

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.

© 2011 ANSYS, Inc. May 8, 2012 7

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

Validation material

CFD delivers increased accuracy and insight in flow conditions

© 2011 ANSYS, Inc. May 8, 2012 8

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_EWEC2010presentation.pdf

© 2011 ANSYS, Inc. May 8, 2012 9

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

Technical Advances

© 2011 ANSYS, Inc. May 8, 2012 10

Resistive ‘Canopy Model’ available:

– Svensson

– Lopes da Costa

– Katul

– Resistance in momentum only

Canopy Input Data

– From roughness data

– CFX Interpolation Table

variable tree heights

Forestry loss coefficient – Constant or variable with height

Forest Canopy Model

Max Forestry Loss Coef

Height Max

Galion Lidar data

CFD

Flow separation off forested region

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.

© 2011 ANSYS, Inc. May 8, 2012 11

Horns Rev Results at Hub Height – Sector 280

Horizontal velocity Turbulence intensity

Uref = 8 m/s at 70m, z0 = 0.0002m Wind direction: sector 280

© 2011 ANSYS, Inc. May 8, 2012 12

0

0.2

0.4

0.6

0.8

1

1.2

1 2 3 4 5 6 7 8 9 10

Norm

alis

ed P

ow

er

Turbine Group

10m/s 270° 2° bin

ANSYS CFD

0.2

0.4

0.6

0.8

1

1.2

1 2 3 4 5 6 7 8 9 10

Norm

alis

ed P

ow

er

Turbine Group

10m/s 270° 10° bin

ANSYS CFD

Data - UpWind

0.4

0.6

0.8

1

1.2

1 2 3 4 5 6 7 8 9 10

Norm

alis

ed P

ow

er

Turbine Group

10m/s 270° 30° bin

ANSYS CFD

Data - UpWind

Horns Rev Normalised Power Down a Row

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

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.

© 2011 ANSYS, Inc. May 8, 2012 13

0

0.2

0.4

0.6

0.8

1

1.2

1 2 3 4 5

No

rma

lis

ed

Po

we

r

Column

10 deg bin

ANSYS CFD

Measured Data

Upper 25%

Lower 25%0

0.2

0.4

0.6

0.8

1

1.2

1 2 3 4 5

No

rma

lis

ed

Po

we

r

Column

30 deg bin

ANSYS CFD

Measured Data

Upper 25%

Lower 25%

North Hoyle Normalised Power Down a Row

Very good agreement with power data for both bin sizes

Absolutely blind test case!

Uref = 10 m/s at 67m, z0 = 0.0001m, upstream TI = 7% Wind direction: sector 260

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.

© 2011 ANSYS, Inc. May 8, 2012 14

Improved results with atmospheric stability

Atmosphere on average stable above boundary layer

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

Including surface stability effects significantly affects the prediction of array losses

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.

© 2011 ANSYS, Inc. May 8, 2012 15

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

Harestanes, Masts on site

All masts with data at 70m, 60m, 40m, 30m Hill top masts: Holehouse Hill Hareshaw Rig Valley mast: Bran Rig

Hareshaw Rig

Holehouse Hill

Bran Rig

© 2011 ANSYS, Inc. May 8, 2012 16

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).

Harestanes Mast to Mast Cross Prediction (wind speed)

3 masts (70m only) 3 masts (all heights)

Model BRAN-HOL BRAN-HAR HOL-HAR BRAN-HOL BRAN-HAR HOL-HAR

neutral 11.8% 13.3% 1.8% 24.0% 24.0% 3.2%

stable 0.4% 4.1% 5.4% 2.9% 6.5% 6.2%

Improved results with atmospheric stability

© 2011 ANSYS, Inc. May 8, 2012 17

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_fvx8o5DMjBkMzQxNWMtOTkxYy00ZmFiLWFlNzMtZGFhM2U0NWQ1ODFh&hl=en_GB

Previous years, available from:

https://docs.google.com/leaf?id=0B6Cp_fvx8o5DZmEzYzYxMDAtZGYxZi00YmI0LWIwMjYtZWM3NmViYWM2NDI3&hl=en_GB

Validated Tools







© 2011 ANSYS, Inc. May 8, 2012 18

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.

ANSYS CFD Featured In Independent Publications

© 2011 ANSYS, Inc. May 8, 2012 19

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 )

WindModeller Tools for Automated Solution

© 2011 ANSYS, Inc. May 8, 2012 20

Tools for Automated Solution

Objective – Automation of Analysis and Data

extraction

– Map Mesh CFD Report in

one step

CFD solution + automated post-processing

Wind data transposition module (cross prediction and energy assessment)

© 2011 ANSYS, Inc. May 8, 2012 21

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 Run on local or remote computer Report as html file

© 2011 ANSYS, Inc. May 8, 2012 22

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

(5 or 9 blocks), aimed at process automation

– Template mesh morphed onto STL terrain representation

– Variable mesh topology for elongated/twin wind farms

Meshing Approach

© 2011 ANSYS, Inc. May 8, 2012 23

Varying Mesh Topologies

Compact wind farm

Elongated wind farm

Twin wind farms

© 2011 ANSYS, Inc. May 8, 2012 24

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

Meshing Controls

© 2011 ANSYS, Inc. May 8, 2012 25

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

© 2011 ANSYS, Inc. May 8, 2012 26

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

Setup

© 2011 ANSYS, Inc. May 8, 2012 27

Flow Modelling in WindModeller

Atmosphere modelled as: – incompressible fluid (Air at 15C) – assuming neutral stability – solving for steady state RANS

Turbulence modelled via two-equation model – Shear Stress Transport (SST) turbulence model or k- ε.

Ground modelled as rough wall (spatially variable roughness)

Inlet boundary conditions – Classical constant-shear ABL profiles (Durbin & Petterson Reif ):

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

2/1

2

*

C

uk u

uln

z

z * ( ) 0 z

u

3

*

© 2011 ANSYS, Inc. May 8, 2012 28

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

Wake Modelling

© 2011 ANSYS, Inc. May 8, 2012 29

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

© 2011 ANSYS, Inc. May 8, 2012 30

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

Post-Processing in WindModeller

© 2011 ANSYS, Inc. May 8, 2012 31



– Plots at constant height AGL and profiles at wind turbine/mast locations for quantities such as normalised velocity, turbulence intensity, shear exponent factor





© 2011 ANSYS, Inc. May 8, 2012 32



– 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





© 2011 ANSYS, Inc. May 8, 2012 33



– 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





© 2011 ANSYS, Inc. May 8, 2012 38

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.

Data collected at mast

Wind conditions predicted at WTG

Wind data transposition module

© 2011 ANSYS, Inc. May 8, 2012 39

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 post-mortem 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)

© 2011 ANSYS, Inc. May 8, 2012 40

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).

© 2011 ANSYS, Inc. May 8, 2012 41

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.

Availability of WindModeller Tools

© 2011 ANSYS, Inc. May 8, 2012 42

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)

Ongoing Projects

© 2011 ANSYS, Inc. May 8, 2012 43

• 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.

Summary

© 2011 ANSYS, Inc. May 8, 2012 44

Upcoming ANSYS Events

• Webinar - Advanced Multi-body Hydrodynamics and Motion Analysis Using AQWA Software - 2012 Update – Friday, 11th May 2012

– http://www.ansys.com/aqwawebinar2012

• All Energy in Aberdeen on May 23 & 24 - stand C111.

ANSYS Events

Scheduled for 2012

Information: http://www.ansys.com/events

• [email protected]

Question, Comments, Inquiries…

http://www.ansys.com/aqwawebinar2012

http://www.ansys.com/events

wind farm flow modelling using cfd - ansys … uk...wind farm flow modelling using cfd 2012 update...

Documents