SlideShare a Scribd company logo
This Document Contains the Following
Data Types
GE Confidential
Wind Plant Reliability
Complex Flow Research Perspectives
Stefan Kern, GE Global Research, Munich
August 14, 2013
@#GECON&*
Sandia Wind Plant Reliability Workshop
August 14 2013
Focus on Wind Plant
Holistic plant-level viewTurbine focus
Design
Siting
Operation
Reduce LCOE by improved design, operation & micrositing
Opt. performance w/ clean
blades in free-stream
Opt. plant power curve
Identical units according to
site wind class
Optimized product- &
component mix
„Cooperative“ control;
use excess margins
As independent units
@#GECON&*
Sandia Wind Plant Reliability Workshop
August 14 2013
Complex Flow Challenges
Design Siting Control
• Incomplete physics understanding
• Lack of data / limited exploitation of existing fleet data
• Limited modeling capabilities, standards lagging behind
@#GECON&*
Sandia Wind Plant Reliability Workshop
August 14 2013
Needs
Understand the problem & entitlement
• How much wake losses can we recover?
• Excess margins or underestimating loads?
Improved engineering practices & models
• Hierarchy of models - modular fidelity
• Proven & validated against field data
• Fleet data mining
New technology
• Rethink turbine design & siting
• Park level control strategies
Both internal R&D and collaborations with
external partners
@#GECON&*
Sandia Wind Plant Reliability Workshop
August 14 2013
Learning from Field Data
Wake behind single turbine
Wind velocity [m/s]
Std dev
95% conf int
mean
Heightaboveground[m]
Buffalo Gap wind park - one out
of hundreds in the GE fleet
Full scale flow field measurements
• Averaged velocity profiles of isolated/interacting wind
turbines pairs & wake turbulence
• Sparse data
Operational data of GE fleet
• Recorded 10min data of hundreds of wind parks
over many years
• Data quality can be challenging
• How to define/measure ideal farm performance?
Gaps: - Best practices for fleet data analysis
- Exploiting large data to full extent
- Limited load data inside plant along with
detailed inflow conditions
@#GECON&*
Sandia Wind Plant Reliability Workshop
August 14 2013
Learning from Scaled Experiments
Gaps: - Scalability questions for small scale models
- Medium scale experiments in largest facilities
Single turbines
• Validation data for predictions of blade aerodynamic
performance with controlled inflow
• Typically focussed on near wake, limited far wake data
Interacting turbines
• Typically models with D<50cm in ABL wind tunnels
• Limited flexibility in controlling turbine operation
Source http://wire.epfl.ch
Source http://wind.nrel.gov
NREL turbine in
NASA AMES wind tunnel
Wind farm in ABL wind tunnel
MEXICO experiment, DNW wind
tunne
Source http://www.ecn.nl
@#GECON&*
Sandia Wind Plant Reliability Workshop
August 14 2013
Engineering Models
Rotor & wake
• Largely focused on optimizing standalone unit performance
• Limited capability to capture complex inflow
Farm level
• Limited to steady state, assumption of constant wind
direction and wind speed
• Following IEC certification requirements/common practice
Gaps: - Comprehensive calibration of farm models to field
data/high fidelity models
- Capturing of unsteady effects & complex terrain
@#GECON&*
Sandia Wind Plant Reliability Workshop
August 14 2013
High Fidelity Simulation & HPC
Gaps: - Comprehensive validation
- Largely limited to simple terrain
- Flowdown to engineering tools
- Exploiting large data to full extent
PSU Cyber Wind Facility
Rotor & wake
• Capture complex inflow for detailed rotor aero & wake
• Long term vision: replace expensive field tests
• Assessing modeling approaches & numerical methods
Farm level
• Assess wake impact, control strategies & new architectures
• Reduce wake losses and fatigue loading
Partners: Penn State, NREL, LLNL, UC Louvain
@#GECON&*
Sandia Wind Plant Reliability Workshop
August 14 2013
Concerted Field Tests & Simulation
Source: energy.sandia.gov
…key to advance and validate modeling capabilities
Full scale
Wind farm
Red. scale
Few turbines
Test setup Simulation
Wind tunnel
Contr. cond
Measmt tech
Source: mexnext.org
Source: NREL
Source: Texas Tech University
Distance from rotor plane (m)
Source: mexnext.org
@#GECON&*
Sandia Wind Plant Reliability Workshop
August 14 2013
Conclusions
• Great momentum & progress in complex flow research
− Experimental facilities
− Measurement technology
− Simulation
• Overcome barriers to develop & deploy new technologies
− Demonstrate business value
− Flow down to engineering practices and tools

More Related Content

Stefan Kern: 2013 Sandia National Laboratoies Wind Plant Reliability Workshop

  • 1. This Document Contains the Following Data Types GE Confidential Wind Plant Reliability Complex Flow Research Perspectives Stefan Kern, GE Global Research, Munich August 14, 2013
  • 2. @#GECON&* Sandia Wind Plant Reliability Workshop August 14 2013 Focus on Wind Plant Holistic plant-level viewTurbine focus Design Siting Operation Reduce LCOE by improved design, operation & micrositing Opt. performance w/ clean blades in free-stream Opt. plant power curve Identical units according to site wind class Optimized product- & component mix „Cooperative“ control; use excess margins As independent units
  • 3. @#GECON&* Sandia Wind Plant Reliability Workshop August 14 2013 Complex Flow Challenges Design Siting Control • Incomplete physics understanding • Lack of data / limited exploitation of existing fleet data • Limited modeling capabilities, standards lagging behind
  • 4. @#GECON&* Sandia Wind Plant Reliability Workshop August 14 2013 Needs Understand the problem & entitlement • How much wake losses can we recover? • Excess margins or underestimating loads? Improved engineering practices & models • Hierarchy of models - modular fidelity • Proven & validated against field data • Fleet data mining New technology • Rethink turbine design & siting • Park level control strategies Both internal R&D and collaborations with external partners
  • 5. @#GECON&* Sandia Wind Plant Reliability Workshop August 14 2013 Learning from Field Data Wake behind single turbine Wind velocity [m/s] Std dev 95% conf int mean Heightaboveground[m] Buffalo Gap wind park - one out of hundreds in the GE fleet Full scale flow field measurements • Averaged velocity profiles of isolated/interacting wind turbines pairs & wake turbulence • Sparse data Operational data of GE fleet • Recorded 10min data of hundreds of wind parks over many years • Data quality can be challenging • How to define/measure ideal farm performance? Gaps: - Best practices for fleet data analysis - Exploiting large data to full extent - Limited load data inside plant along with detailed inflow conditions
  • 6. @#GECON&* Sandia Wind Plant Reliability Workshop August 14 2013 Learning from Scaled Experiments Gaps: - Scalability questions for small scale models - Medium scale experiments in largest facilities Single turbines • Validation data for predictions of blade aerodynamic performance with controlled inflow • Typically focussed on near wake, limited far wake data Interacting turbines • Typically models with D<50cm in ABL wind tunnels • Limited flexibility in controlling turbine operation Source http://wire.epfl.ch Source http://wind.nrel.gov NREL turbine in NASA AMES wind tunnel Wind farm in ABL wind tunnel MEXICO experiment, DNW wind tunne Source http://www.ecn.nl
  • 7. @#GECON&* Sandia Wind Plant Reliability Workshop August 14 2013 Engineering Models Rotor & wake • Largely focused on optimizing standalone unit performance • Limited capability to capture complex inflow Farm level • Limited to steady state, assumption of constant wind direction and wind speed • Following IEC certification requirements/common practice Gaps: - Comprehensive calibration of farm models to field data/high fidelity models - Capturing of unsteady effects & complex terrain
  • 8. @#GECON&* Sandia Wind Plant Reliability Workshop August 14 2013 High Fidelity Simulation & HPC Gaps: - Comprehensive validation - Largely limited to simple terrain - Flowdown to engineering tools - Exploiting large data to full extent PSU Cyber Wind Facility Rotor & wake • Capture complex inflow for detailed rotor aero & wake • Long term vision: replace expensive field tests • Assessing modeling approaches & numerical methods Farm level • Assess wake impact, control strategies & new architectures • Reduce wake losses and fatigue loading Partners: Penn State, NREL, LLNL, UC Louvain
  • 9. @#GECON&* Sandia Wind Plant Reliability Workshop August 14 2013 Concerted Field Tests & Simulation Source: energy.sandia.gov …key to advance and validate modeling capabilities Full scale Wind farm Red. scale Few turbines Test setup Simulation Wind tunnel Contr. cond Measmt tech Source: mexnext.org Source: NREL Source: Texas Tech University Distance from rotor plane (m) Source: mexnext.org
  • 10. @#GECON&* Sandia Wind Plant Reliability Workshop August 14 2013 Conclusions • Great momentum & progress in complex flow research − Experimental facilities − Measurement technology − Simulation • Overcome barriers to develop & deploy new technologies − Demonstrate business value − Flow down to engineering practices and tools