The document discusses NREL's wind energy systems engineering initiative, which aims to develop an integrated software framework for modeling wind plant performance and costs from a systems perspective. The framework integrates various turbine and plant models from NREL's WISDEM suite using OpenMDAO. Initial analyses demonstrate the impact of system-level coupling on sensitivity and optimization studies compared to previous simplified models. The framework allows flexible configuration and improvement of models to better capture interactions across the full wind energy system over its lifetime.
Driving Behavioral Change for Information Management through Data-Driven Gree...
Katerine Dykes: 2013 Sandia National Laboratoies Wind Plant Reliability Workshop
1. NREL is a national laboratory of the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, operated by the Alliance for Sustainable Energy, LLC.
Wind Energy System
Design & Engineering
August 13, 2013
Katherine Dykes, Andrew Ning, Peter Graf, George Scott,
Rick Damiani, Derek Berry, Ben Maples, Ryan King,
Patrick Moriarty Maureen Hand and Paul Veers
2. 2
Outline
• NREL NWTC Wind Energy Systems
Engineering Program and Software
o A Systems Approach to Wind Energy
Research, Design and Development
o Model and Tool Development
o Initial Analysis Work
3. 3
Outline
• NREL NWTC Wind Energy Systems
Engineering Program and Software
o A Systems Approach to Wind Energy
Research, Design and Development
o Model and Tool Development
o Initial Analysis Work
4. 4
Motivation
• Wind Energy Technology is a complex system
o Many disciplines – aerodynamics, structures, electrical, etc
o Many stakeholders – supply chain, developers, financiers,
environmentalists, communities
o Long time scales – operation over several decades
o Large scope – activity within a single component to interaction of
turbines within the plant to interaction of plant with the grid
Environmental
Impacts Grid Integration
Impacts
Wind Cost of
Energy
Community
Impacts
5. 5
Wind Energy System Cost of Energy
• Often use simplified Cost of Energy representation as
global system objective:
o Where COE is cost of energy, CAPEX is the sum of BOS is
Balance of Station cost, TCC is Turbine Capital Cost (for full
project), F is the financing rate to annualize investment costs,
OPEX are the annual operating expenses and AEP is the net
annual energy production
7. 7
• Current “NREL Cost and
Scaling Model” uses
parameterized functional
relationships calibrated to
historical trends
o Originated with detailed design
studies in early 2000s
(WindPACT)
o Abstraction to simple parametric
relationships
o Useful for two primary types of
analyses on system costs:
– Changing input factor prices
over time
– Scaling of conventional
technology within a limited
range
o Publically available model
Current structure of cost model
Determining Wind Cost of Energy
8. 8
NREL Wind Energy System Engineering
• The National Wind Technology Center (NWTC) wind
energy systems engineering initiative has developed an
analysis platform and research capability to capture
important system interactions to achieve a better
understanding of how to improve system-level
performance and achieve system-level cost reductions.
• NREL Initiative in Systems Engineering goal is to
develop, maintain, and apply a software platform and
to leverage its research capabilities to:
o Integrate wind plant engineering performance and cost software modeling
to enable full system analysis.
o Apply of a variety of advanced analysis methods in multidisciplinary design
analysis and optimization (MDAO) and related fields.
o Develop a common platform and toolset to promote collaborative research
and analysis among national laboratories, industry, and academia.
9. 9
Outline
• NREL NWTC Systems Engineering
Program Overview
o A Systems Approach to Wind Energy
Research, Design and Development
o Software Development
o Initial Analysis Work
11. 11
Systems Engineering Software Framework
OpenMDAO:
• Work flows integrate models
together in structured ways (use
of NASA’s OpenMDAO software)
• Easily reconfigured (model
selection and analysis structure)
FUSED-Wind:
• Support software for structured
interface of wind turbine and
plant modeling tools into
OpenMDAO
• Allows interchange of models,
sharing of common turbine and
plant input descriptions
NREL WISDEM:
• Collection of NREL engineering
and cost turbine and plant
analysis tools integrated into
FUSED-Wind and OpenMDAO
• Full wind turbine and plant
model set planned for release
12. 12
NREL System Engineering Program Objectives
• Integration of models into framework includes several areas:
1. Turbine component structure and cost models
2. Structural models with dynamic models of turbine
performance
3. Integration of turbine models with physical plant models for
turbine interactions affecting loads and energy production
• At each level, a range of model fidelity is possible/needed
for analysis flexibility (highly modular)
NREL Cost
and Scaling
Model
(Blade Cost)
Example transition from Cost and Scaling Model to Systems Engineering Framework
Blade
Cost and
Sizing
Tool
PreComp
Sandia
NuMAD
pBEAM
Blade MCM
FAST
AWS
Truepower
OpenWind
13. 13
Turbine Engineering Analysis Modules: Rotor Example
AIRFOIL
• Radial location, chord, twist
• Profile shape
• Lift curves and drag polars
(Parameterized by Re and t/c)
• Web locations
ATMOSPHERE
• Density
• Viscosity
• Shear exponent
• Wind speed distribution
LAMINATE STACK
• Sequence of lamina
• Number of plies
• Ply thickness
• Ply orientation
• Ply material
TURBINE INPUTS
• Hub Height
• Cut-in and cut-out speed
• Min/max rotation speed
• Rated power
• Machine type
(Fixed/Variable Speed/Pitch)
• Drivetrain Efficiency
ROTOR INPUTS
• Number of blades
• Precone, tilt ROTOR STRUCTURE &
PERFORMANCE
• Blade mass and moments of
inertia
• Cost
• Power curve
• Turbine AEP
• Hub loads
• Sound power level
• Natural frequencies
• Blade displacements
• Panel buckling loads
• Axial and shear stress
<<< INPUTS
OUTPUTS >>>
• Geometry and
Materials are
pre-defined
• Outputs
include blade
mass
properties
and rotor
performance
14. 14
Plant Level Models: Turbine Capital Costs
• Structure and cost models aggregate together for full
turbine structure model preserving interactions
between sub-systems
•Rotor Speed
•Max Thrust
•Torque
•Moment
•Mass/Modal
Properties
•RNA Forces & Moments
•RNA Mass/Modal
Properties
Turbine Model
Component Costs
and Properties:
• Mass
• Cost
• natural
frequencies
• Deflection
• critical buckling
load
• axial stress
Turbine Capital Costs
and Properties:
• Mass
• Cost
TURBINE MODEL
INPUT DESIGN
PARAMETERS:
• Wind, water and
soil statistics
• Rotor Diameter
• Rotor Airfoil Class
• Rotor Geometry
• Drivetrain
Gearbox /
Generator
Configuration
• Machine Rating
• Hub Height
• Tower Geometry
<<< INPUTS OUTPUTS >>>
Drivetrain CST
LSS
• Length
• Outer Diameter
• Inner Diameter
• Weight
Bearing Positions
Main bearings
• Weight
• Housing Weight
Gearbox Configuration
• Gear Number
• Gear Type(s)
• Overall Stage Ratio
Gearbox
• Weight
• Stage Weights
• Stage Ratios
HSS &
Coupling
• Brake
Weight
• HSS
Weight
Generator Rating
Generator
• Weight
Bedplate
• Length
• Area
• Weight
Yaw
• Weight
ROTOR Outputs
• Diameter
• Mass
• Max Thrust
• Speed at Rated
• Torque at Rated
<<< INPUTS
OUTPUTS >>>
Drivetrain Configuration
Tower Outputs
• Top Diameter
Rotor CST
Tower/Monopile CST
RNA Assembly
15. 15
Plant Level Models: Finance / Cash Flow
• Model takes inputs from all other plant models on costs and energy
production
• Model determines values of a variety of financial indicators of interest
Cash Flow Finance Model
CASH FLOW MODEL
OUTPUTS:
• Project NPV
• Project IRR
• Project COE /
LCOE
• Project PPB
CASH FLOW MODEL
INPUTS:
• Electricity Price
• Debt / equity ratio
• Debt Interest Rate
• Equity Rate
• Tax Rate
• Incentives
• Project Timeline
<<< INPUTS
OUTPUTS >>>
OTHER MODEL
OUTPUTS:
• Turbine Capital
Costs
• Balance of Station
Costs
• Annual Energy
Production
• Operations &
Maintenance
Costs
17. 17
Outline
• NREL NWTC Systems Engineering
Program Overview
o A Systems Approach to Wind Energy
Research, Design and Development
o Model and Tool Development
o Initial Analysis Work
18. 18
Initial Analysis Capabilities
• Turbine component sizing tools allow for sub-optimization of various sub-
components
• Component outputs feed simplified component cost-models primarily based
on NREL Cost and Scaling model data
• New NREL Balance of Station Cost model and ECN O&M Offshore Cost model
used for plant costs
• Energy production determined by AWS Truepower OpenWind Community or
Enterprise versions
• Simpler NREL Cost and Scaling model able to substitute for any of the cost
model modules
20. 20
Initial Analysis: Sensitivity Study
• Any analysis requires both turbine and plant
design inputs:
o Study uses NREL 5-MW reference turbine model and
offshore reference site.
NREL 5 MW Reference Turbine Parameter Value
Rotor:
Rotor Diameter 126 m
Rated Wind Speed 12.1 m / s
Cut-In / Cut-Out Wind Speeds 3 m /s / 25 m / s
Maximum Allowable Tip Speed 80 m/s
Tower:
Hub Height 90 m
Tower Length / Monopile Length 60 m / 30 m
Tower Top / Base Diameters 3.87 m / 6.0 m
Tower
Drivetrain Configuration 3-stage Geared
(EEP)
Rated Power 5 MW
Gearbox Ratio 97:1
Drivetrain Efficiency at Rated
Power
94.4%
Offshore Site Conditions
Distance to Shore 46 m
Sea Depths <5% at ~20 m, 25%+ at
~30 m
Wind Speed at 90 m Mean = 9.78 m/s
Weibull shape = 2.15,
scale = 10.5
Significant Wave
Height
10-year Extreme = 7.5 m
50-year extreme = 8 to
8.5 m/s
Significant Wave
Period
10-year Extreme = 19.6 s
21. 21
• Baseline COE analysis CSM versus CSTs
– Overall cost of energy higher using new model set:
• Energy production of higher fidelity physics-based model slightly higher.
• Turbine capital costs are similar between two models
– slightly higher for the NREL CSM due to the inclusion of a 10%
“marine-ization” factor
• O&M costs roughly consistent from old to new model on first analysis
• Balance of Station costs of NREL CSM were known to be very low – new
model (still in develop) estimates BOS costs as much higher
– COE for new model closer to industry-reported offshore wind costs for
European projects.
Initial Analysis: Sensitivity Study
NREL CSM NREL SE w/ CSTs
COE $0.11 $0.18
AEP (MWh / turbine-yr) 18,800 19,900
Turbine Capital Costs ($ / kW) $1,200 $1,000
BOS Costs ($ / kW) $1,700 $3,600
O&M Costs ($ / kWh) $0.027 $0.026
22. 22
• Sensitivity of system cost to changes in key parameters (rotor
diameter, rated power, hub height, and maximum tip speed)
performed – change of +/- 10% for each.
• Example analysis: rotor diameter.
– Changes in COE was more pronounced using new set of models that
capture more system coupling:
• Rotor diameter influences balance of station model in latter case; overall BOS
impact on costs of energy higher for new model set.
• Impacts on energy production are much higher due to power curve changes
• Turbine capital cost are lower since costs are now associated with component
masses rather than directly to rotor diameter
Initial Analysis: Sensitivity Study
Percent Changes in Parameters NREL CSM NREL SE w/ CSTs
Rotor Diameter (m) -10.0% +10.0% -10.0% +10.0%
COE ↑ ↓ ↑ ↓
AEP (kWh / turbine-yr) ↓ ↑ ↓↓ ↑↑
TCCs ($ / kW) ↓↓ ↑↑ ↓ ↑
BOS Costs ($ / kW) -- -- ↓ ↑
O&M Costs ($ / kWh) ↓ ↑ ↓ ↑
23. 23
• Multidisciplinary Design Optimization Study # 1:
– Joint aerodynamic and structural optimization of a wind turbine
rotor using new NREL rotor aero and structural analysis tools
– Dual objectives for maximization AEP and minimization of weight
– Select design variables (i.e. chord and twist) and add appropriate
constraints (i.e. stress limits)
Initial Analysis: Multidisciplinary Optimization
24. 24
• Multidisciplinary Design Optimization Study # 2:
– Optimization of a full integrated design of wind turbine and plant under
uncertain site wind climate conditions
– Overall objective to minimize overall cost of energy for the wind plant
(under uncertainty of wind conditions)
– Choose design variables (i.e. rotor diameter), add constraints and sources
of uncertainty (i.e. site weibull distribution parameters)
Initial Analysis: Multidisciplinary Optimization
25. 25
1. A systems perspective to wind energy cost and
performance analysis is essential – extensive coupling
exists between physical assets over long periods of time.
2. NREL has developed initial capability for modeling
integrated wind plant systems for performance and cost.
3. Initial work shows improved representation of coupling in
analysis results; however, model improvement is needed
across all system models.
4. Research design becomes critical with a flexible tool
Conclusions and Summary