Saarblitz, Wind farm in HDR, flickr.com, creative commons by-nc-sa 2.0

IEA Wind Task 37 Systems Engineering

The purpose of IEA Wind Task 37 is to coordinate international research activities, towards the analysis of wind power plants as holistic systems.  To fully assess how a change, or an uncertainty, in a design parameter affects the myriad of objectives in system performance and cost, a holistic and integrated approach is needed. Integrated systems research, design and development (RD&D) can provide opportunities for improvements in overall system performance, and reduction in the levelized cost of energy. However,  there are significant challenges to developing such integrated approaches, both within and across organizations. There is a need to explore both the opportunities and the challenges for applying systems engineering to integrated wind energy RD&D across the entire wind energy system. This need surfaces both in the tools and methods used in wind plant RD&D.

The experience of the NREL 5 MW Reference Turbine, as well as previous IEA Tasks such as OC3 and OC4, shows that effective coordination can be achieved by providing frameworks such as reference designs and reference cases for analysis.  The proposed IEA Task goes one step further.  In addition to providing a forum for reference system (wind turbines and plants) development and benchmarking activities (in the area of multi-disciplinary design analysis and optimization (MDAO)), the proposed task will provide framework guidelines that will enable more seamless integration of analysis tools and reference models between organizations.  This proposal describes the current status of research in integrated RD&D for wind turbines and plants and a set of work packages to enable collaborative work in this growing research area.

Introduction

Over the last few decades, wind energy has evolved into a large international industry involving major players in the manufacturing, construction, and utility sectors. Coinciding with industry growth, significant innovation in the technology has resulted in larger turbines and wind plants with lower associated costs of energy. However, the increasing importance of wind energy’s role within the electricity sector imposes more requirements on the technology in terms of performance, reliability, and cost. To address these changing expectations, the industry has made efforts that focus on achieving a variety of goals including reducing installed capital costs for the turbine and plant, decreasing the downstream costs for operation and maintenance (O&M), increasing energy production, and minimizing negative external environmental impacts such as noise emission or habitat disruption. In many cases, these goals involve trade-offs. For example, up-front investment in a robust component design may avoid large downstream costs for component repair and replacement. In another case, the design of a machine with a higher tip speed can reduce required torque and loads through the drivetrain, but at the same time these higher tip speeds can lead to more aero-acoustic noise that adversely impacts surrounding communities. Trade-offs, and techno-economical conflicts such as these exist throughout the entire system.

Considering the plant as a collection of wind turbines, the operation of the wind turbines and design of a wind plant is linked via:

  • the atmosphere, where the operation of upwind turbines modifies the downstream flow locally, with a velocity deficit and added turbulence behind each rotor; and globally, acting as an additional surface drag on the atmospheric boundary layer;

  • the electric system, where the voltage and frequency at one wind turbine may be a function of the power output of other wind turbines; and,

  • plant control actions, where the logic used to distribute a power set-point to a given wind turbine may be a function of the operating conditions of other wind turbines.

Thus, understanding the design and dynamics of a wind power plant requires a system-level model. There are open research topics in all areas of wind power plant design and operation.  For instance, how does pitching the blades of a group of upstream turbines affect the atmospheric boundary layer flow at a group of downstream turbines, where the convection time at the nominal hub-height wind speed may be a half-hour or more?  How should the power dispatch function respond to rapidly varying flow conditions across the plant, for instance with the passage of weather fronts or thunderstorms?  How should the converters at each turbine -- and for plants with HVDC transmission, at each end of the HVDC line -- be operated to best support power-system stability?  How should the level of fatigue degradation of each turbine be monitored, and how should the turbines be operated, in order to most economically distribute the degradation among the turbines in the plant?  What influence does turbine placement have on energy production, effective turbulence intensity, fatigue degradation, maintenance requirements, and system costs?  

Finding answers to these and many more system-level questions is a multi-disciplinary effort, requiring the coordination of diverse research groups and analytical capabilities.  The sections that follow describe the state-of-the-art in MDAO activities for wind energy as well as the analysis tools, frameworks, and reference systems that support those activities.

Wind Turbine Modelling and Design with Emphasis on MDAO

Wind turbine design is an inherently multidisciplinary field, with strong couplings between e.g. the aerodynamics and the structural dynamics of the rotor. In the 1980s this led to the development of the first aeroelastic simulation codes for predicting the structural response to unsteady wind inflow, such as Flex [Øye]. This was followed by several other codes such as BLADED, FAST, HAWC and HAWC2, Cp-Lambda, and others. Design of wind turbines using optimization dates back to the mid 1990s, where e.g. Selig and Coverstone-Carroll [Selig et al] presented an aerodynamic design methodology combining a genetic algorithm with inverse design. Fuglsang et al [Fuglsang et al] presented one of the first works on multidisciplinary design of a wind turbine, where an aeroelastic model of the wind turbine blade was used to optimize  the cost of energy. Several other works followed, among them Bottasso et al, who combined a multi-body aeroelastic model with a detailed cross-sectional structural model to maximise AEP and minimise mass, also investigating design features such as bend-twist coupling. Merz et al presented a methodology based on a frequency domain aeroelastic code, combined with a simplified structural model for the design of stall regulated turbines. Also notable in this work is that the authors reproduce previously published optimization results as validation of their method. Ning et al systematically explored different types of objectives and constraints in turbine design in which the combined rotor, drivetrain and tower designs were included. MDAO is also used in industry with many of the current designs being based in MDAO, see e.g. Mayda et al. Ashuri, and Ashuri et al presented an integrated  multidisciplinary design optimization approach to minimize the levelized cost of energy. Haghi et al developed a computational framework to do integrated support structure optimization including the aerodynamics, soil and hydrodynamics. A method for optimizing wind turbine gross parameters, such as rotor diameter and hub height for a specific site was presented by Diveux et al. The system-level design of a wind turbine blade using a multi-level optimization approach was studied by Maki et al. Integrating the controller design of wind turbines with the structures and aerodynamics analysis to optimize the cost of energy was presented by Ashuri et. al. A system approach as a conceptual method was proposed by Petrini et al. to design  offshore wind turbine structures.

A common trait of work in MDAO of wind turbines is that there exists very little work that contains verification  of the design methods themselves, either based on comparison to previous works or to analytical results. This makes it difficult to carry out a systematic evaluation and comparison of the many methodologies and model types applied, and based on the literature, there is no decisive consensus of best practices in relation to e.g. gradient-based vs gradient-free optimization algorithms, model fidelity, cost functions etc.  One of the key aims in the present project will thus be to carry out more systematic benchmarks of the various methodologies and models used in the context of MDAO of wind turbines, to shed light on the current best practices and to direct the research community towards addressing major challenges in advancing MDAO of wind turbines.

Wind Plant Modeling and Design with Emphasis on MDAO

Wind plant modelling and design is a very broad multi-disciplinary field as the design of a sub-component, the control strategy and placement of the wind turbines can create complex interactions between each element of the plant system, and ultimately have significant impact on the plant economy or viability. The field is so broad that researchers typically only focus on a subset of the problem, e.g. wind plant energy production predictions, wind plant control or wind plant layout optimization. Furthermore, as the researchers typically come from mainly one discipline, they have a tendency to focus their effort on improving their own discipline,  reimplementing the simplest models in the other disciplines or even disregarding them. So, for instance, studies which approach wind power plant analysis from a "mechanical" perspective -- for example, trading off energy production with mechanical loads on the turbine structures -- tend to ignore the electrical grid [Schaak, Senjyu, Soleimanzadeh, Steinbuch, van Dam]; while studies which approach wind power plant analysis from the wider grid integration perspective tend to employ aggregated models, where many turbines are lumped into one equivalent unit, whose output is upscaled [Bozkho, Haileselassie, Liu, Wei].  Studies on the atmospheric dynamics of wind plants tend to ignore or greatly simplify the mechanical and electrical models [Barthlemie, Calaf, Frandsen, Rethore].

Notable exceptions are Akhmatov et al. [Akhmatov], who modelled the electric grid of an 80-turbine wind plant, including basic models for aerodynamics and drive train dynamics; the EU project Aeolus (2008-2011) (http://www.ict-aeolus.eu/), developing wind plant control algorithms, where some account was taken of aerodynamics, fatigue, and grid operator requirements; and TOPFARM [Rethore], which combines the aeroelastic code HAWC2 with the Dynamic Wake Meandering model,  a transient wind plant flow field model, and a rudimentary consideration of the electrical infrastructure and offshore foundation cost model.  Other studies have looked at innovation impacts from an entire system perspective [Dykes, Ning].  Yet, clearly, there is still much progress to be made in terms of multidisciplinary systems analysis of wind power plants.  As there are no reference open source workflows of models used within the community, this narrowed focus makes the comparison between individual research works quite difficult, and therefore the scientific impact of each contribution very low. The negative effect of the lack of international collaboration has been studied recently in a literature review on wind plant layout optimization [Herbert-Acero]. While the field of research has grown in interest in the recent years (25+ journal articles per year), the community has so far not succeeded in establishing a proper collaboration, quality check and common benchmarks around their scientific publications.  In order to accelerate the research in the area of wind plant MDAO, it is therefore important to support the development and dissemination of common frameworks allowing the development of reference workflows and to establish common benchmarks and reference plants, so that researchers can collaborate with each other by contributing to the fields within their areas of expertise.

Frameworks for wind turbine and plant MDAO

Common frameworks in modeling can enable accurate comparisons between different analytical modeling tools as well as enable more seamless collaboration among a variety of stakeholders. In the former case, the ability to define standardized input/output and interfaces among software packages can allow for more accurate comparison of analysis results among the very broad set of software tools used by the industry for system analysis. In the latter case, barriers to information exchange across and even within organizations can be addressed by the use of standardized frameworks where interfaces between historical silos are well-delineated or even support coupling of analysis tools across traditional boundaries.

To support MDAO for wind turbines and plants, various efforts have developed over the last several years to look at integrated modeling of wind energy systems.  As discussed, there are many traditional tools for modeling wind turbine components and overall system dynamics along with various tools for modeling wind plant performance, cost and supporting wind plant design.  The survey of tools shown in the table below focuses on tools that go beyond the system boundaries of traditional tools to look at wind turbines and plants in more integrated ways.



Table 1: Integrated Modeling Tools for Wind Energy Systems (* indicates OpenSource)

Name

Organization

Description

Cp-Lambda

TU Munich, Politechnico di Milano

Integrated design of wind turbine components with system dynamics analysis. Has also been integrated with cost models for cost optimization.

OpenFAST*

NREL

Newest version of NREL’s multi-physics wind turbine modeling tool which is designed for modularity and flexibility to integrate a large variety of dynamic models for aerodynamics, structural dynamics, hydrodynamics, controls, etc.

FOCUS6

ECN

Integrated design of wind turbine components with system dynamics analysis.

FUSED-Wind*

DTU / NREL

The Framework for Unified Systems Engineering and Design of Wind Plants (FUSED-Wind) does not contain models but prescribes interrelationships between models for wind turbine and plant design and analysis.

LMS Samtech Samcef Wind Turbines

Siemens LMS Samtech Subsidiary

Integrated design of wind turbine components with system dynamics analysis.

OneWind / OneWind Modelica Library

Fraunhofer IWES

Integrated design of wind turbine components with multi-physics system dynamics analysis.

QBlade*

TU Berlin

Combines airfoil design with performance analysis for HAWT and VAWT rotor design that can be coupled with the FAST aeroelastic code for dynamic simulation.

SOWFA*

NREL

Integrates dynamic wind turbine model with computational fluid dynamics (CFD) software to simulate complex flows within a wind plant.

TOPFARM*

DTU Wind Energy

An optimization platform for wind plant layout that accounts for wind turbine controls and operational costs for the plant over time by integrating the cost models with dynamic models for the turbine and wind flow in the plant.

WISDEM*

NREL

The Wind Plant Integrated System Design and Engineering Model (WISDEM) is an integrated set of turbine and plant design and cost tools implemented in FUSED-Wind to enable full system or sub-system analysis and MDAO.  

Reference Wind Energy Systems

Closely coupled to the need for frameworks from an integrated modeling perspective is the need for reference cases at both a wind turbine and plant level to  enable comparison across analyses, and  collaboration in the wind energy RD&D process. Several reference turbines have been developed in the past, and these past efforts have served as common start points for huge numbers of research analyses and even design efforts. Increasingly, however, the wind turbine analysis is not done in a vacuum, and there is a need to define reference plants for analysis of the whole system.

Reference Wind Turbines

Wind turbine data including the aerodynamics, structures, controls and costs for various components are needed for design studies, advanced technology assessments, and for verification and validation of computational models and codes  in both the onshore and offshore wind applications. However, turbine OEMs are reluctant to share any  turbine data  with the outside world.  For example, in offshore applications, this paradigm precludes support structure designers from taking advantage of fully coupled loads analysis. Sequential coupling, where turbine OEMs pass turbine loads to the support structure engineers, and these return stiffness parameters to the turbine OEMs at each design iteration, while acceptable does not lend itself to system optimization, substructure innovation, and cost reduction. Researchers, on the other hand, have very limited access to data for new code validation, and assessing the range of applicability of new aerodynamic, structural, and control design strategies is considerably delayed by the absence of reference turbine models.

NREL developed the so-called NREL 5-MW turbine in 2005 using a combination of commercial turbine data and parameters of other reference turbines available at the time.  Designers and researchers around the world have used that reference turbine, and while there are a number of turbine OEMs who have developed a 5-MW turbine (Repower, Areva, Bard, XEMC, Goldwind, and Gamesa), all of these turbines vary in rotor size (115-152 m) from the original reference.  Wind turbine technology has significantly advanced since 2005, making the 5-MW reference turbine less useful and not easily scalable to modern standards.  Siemens and Alstom are deploying 6-MW turbines, Samsung and Vestas have  7 and 8 MW models, respectively, and MHI (Mitsubishi and Vestas joint venture) is now selling an 8-MW turbine. A 10-MW turbine is likely to be conceived and manufactured in the near future.    The 5 MW reference turbine has a high speed drive train, while most offshore wind turbines have moved toward medium-speed or direct-drive architectures.  The 5-MW reference reaches its rated power at a higher wind speed than operating turbines, which will skew the predicted annual energy production (AEP) — leading to higher levelized cost of energy (COE) estimates when this turbine is used as the basis for economic assessments. Structurally, the reference turbine is also much heavier than modern turbine designs, having implications in tower and support structure design.

In 2013 DTU Wind Energy released the DTU 10 MW Reference Wind Turbine, which is a Class 1A offshore turbine. The turbine was not designed to represent the state-of-the art or incorporate innovative technologies, but was meant to be a baseline for benchmarking new technologies. The turbine is described with high detail, including blade geometrical data needed for both BEM based and CFD codes, as well as detailed structural data relevant for FEM codes. The turbine is currently being used widely in the research community and as of April 2015 has 374 users. It is used in a number of Danish national and European projects such as INNWIND, MareWint and IRPWind.

Reference Wind Plants

The idea of a reference wind plant is a relatively new concept, which came in focus during the past 3 years within the NOWITECH and NORCOWE centers in Norway and the EU project EERA-DTOC.  There are a myriad of ways to describe a wind plant, and different information and parameters are important to different user groups.  As described earlier, over the short-term, there is coupling between the atmosphere and the flow through the plant including the interaction of wind turbine wakes, which in turn impact the production of energy from each turbine and the overall plant.  The wake induced turbulence also results in additional fatigue loads on the turbine that are transferred from rotor to drivetrain to tower to ground.  Over the long term, these loads impact turbine and plant reliability and availability.

Modeling complete plant operation with knowledge of the behavior of the atmosphere, the flow through the plant, the energy production and loads and their translation into long-term energy production, reliability and associated operation and maintenance cost, the grid infrastructure, and the wind plant controller, involves a significant amount of data describing the wind plant. This type of data is typically confidential, sometime owned by different companies, and has so far never reached the research community as a whole. There are very few publicly available information sources for plant data.  The few that are available have thus become the focus of a number of wind plant discipline-oriented studies [Barthelmie 2009 and Churchfield 2012].  The Horns Rev 1 project in the eastern North Sea and the Nysted project near Lolland, Denmark provide data publicly from the SCADA systems which includes averaged power production for a row of turbines [Barthelmie 2009].  More detailed, per turbine data for the Lillgrund project off southern Sweden has been made available for select research studies [Churchfield 2012].

However, there is still significant work to be done to define a set of reference plants that can be used for not just investigating the physics but also as part of MDAO research activities. It is therefore important to create a reference wind plant that allow researchers and industry to collaborate to improve and benchmark wind plant design tools. Besides a description of the in-situ wind farm, this may require descriptions of e.g. installation procedures, operation & maintenance strategy and decommissioning activities.

 

Objectives and Expected Results

To fully assess how a change in a design parameter affects the myriad of objectives in system performance and cost, a holistic and integrated approach is needed. Integrated system research, design and development (RD&D) can provide opportunities for improvements in overall system performance and reductions in overall cost of energy.

The mission of this task is to improve the practice and application of systems engineering to wind energy RD&D.  This will be achieved through a set of coordinated international research activities that move the community towards the analysis of wind power plants as holistic systems.

Explicit goals of the task include:

  • Promoting general knowledge and understanding of systems engineering tools and methods and the overall value of these to wind energy RD&D

  • Improve quality of systems engineering by practitioners

  • Enable better communication across researchers and practitioners in different disciplines

  • Enable system-level analysis including technology evaluation, multidisciplinary design, analysis and optimization, multi-fidelity modeling, and uncertainty analysis and quantification

  • Promoting enhanced design of wind turbines and plants through the use of SE tools and methods

 

Expected results of the effort will include guidelines to support integration of analytical capabilities for modeling wind turbine and plants, reference wind turbine and plant models that may be used by the entire wind energy RD&D community, and reports on best practices in performing MDAO analysis of wind turbines and plants.  The specific results and deliverables, and their organization into work packages are described in the next section.




Approach and Methodologies

In this section the preliminary work plan of the IEA Wind Energy Systems Engineering and Integrated RD&D Task is given. The project structure is composed of four work packages:

  • WP0: Management and Coordination

  • WP1: Guidelines for a common framework for integrated RD&D at different fidelity levels (both turbines and plants)

  • WP2: Reference wind energy systems (both turbines and plants)

  • WP3: Benchmarking MDAO activities at different system levels (both turbines and plants)



WP0: Management and coordination

Period: Month 0-Month 36

Description

WP0 contains the management and coordination activities. The WP primarily focuses on the communication of essential information between the participants on the achievement of the technical objectives according to the time schedule of section 5. The WP also includes the activities which are required to inform the IEA Wind Executive Committee (i.e. the preparation of the progress reports and the attendance of the IEA Wind ExCo meetings). A dedicated project site will be developed for the follow-up of the project.

Tasks:

  • Task 0.1: Technical management

  • Task 0.2: Administrative management

Deliverables:

  • D0.1: First annual progress report (M12)

  • D0.2: Second annual progress report (M24)

  • D0.3: Final report (M36)




WP1: Guidelines on a common framework for wind energy systems

Period: M0-M24

Description

This work package will address the task goals by creating guidelines for a common conceptual architecture for wind turbines and plants so that practitioners can more easily:

  1. share descriptions of wind turbines and plants across multiple parties and reduce the effort for translating descriptions between models,

  2. integrate different models together and collaborate on model development

  3. translate models among different levels of fidelity in the system

 

This work package is a building block for work packages 2 and 3.

 

This work package acknowledges that there are many efforts within industry and research communities to integrate wind turbine and plant models into frameworks to support MDAO activities.  The effort here begins with the task of finding a common ontology (hierarchical framework of characteristics) for these types of models to enable more collaboration and integration across the different community stakeholders.  This will involve surveying current frameworks and then establishing a common ontology and set of guidelines.  This work will be closely coordinated with work package 2 on reference wind energy systems.  The initial framework will leverage work done to create a common data exchange format for the reference turbine and plant work.  The work will provide a basis further reference wind energy system and the benchmarking activities to follow.

Tasks:

  • Task 1.1 Survey of existing frameworks (wind and more general systems engineering frameworks) and summary

    • Task 1.1.1 Catalogue existing MDAO frameworks for wind turbines and plants

    • Task 1.1.2 Create matrix of relevant disciplines and fidelities for wind turbine and plant design

    • Task 1.1.3 Cross-walk catalogue with matrix (i.e. homepage.tudelft.nl/b84g6/rotor/ )

    • Task 1.1.4 Extended literature survey of MDAO activity in wind energy systems

    • Task 1.1.5 Develop requirements for ontology (selection of disciplines / fidelities for version 1, data format, etc)

    • Task 1.1.6 Develop task report on catalogue, matrix and requirements (publish selected report content as external IEA Wind report)

  • Task 1.2 Ontology development (Initial version)

    • Task 1.2.1 Development of turbine ontology and code agnostic data exchange formats

      • Task 1.2.1.1 Rotor

      • Task 1.2.1.2 Tower

      • Task 1.2.1.3 Hub/Drivetrain/Nacelle

      • Task 1.2.1.4 Complete report on turbine ontology

    • Task 1.2.2 Development of plant ontology and code agnostic data exchange formats

      • Task 1.2.2.1 Energy Production

      • Task 1.2.2.2 Balance of Station

      • Task 1.2.2.3 Operations

      • Task 1.2.2.4 Finance

    • Task 1.2.3 Harmonization of data exchange formats into final framework guidelines (publish ontology as external IEA Wind report and make available online)

  • (Tentative) Task 1.4 Extend ontology for additional disciplines and fidelities

    • Task 1.4.1 Select disciplines with second most common fidelity levels

    • Task 1.4.2 Establish ontology within each new discipline/fidelity level combination

    • Task 1.4.3 Develop mapping within discipline across the different fidelity levels

    • Task 1.4.4 Develop an open source toolkit for building data converter (between different fidelity levels)

 

Deliverables:

  • D1.1: Requirements document including survey results

  • D1.2: Document of framework guidelines for wind turbines and plants

  • D1.3: Publication of common data format

 

Participants:

NREL will lead this work package with expected participation from SINTEF/NTNU, DTU Wind Energy, Fraunhofer IWES, TUM, TU Delft and DNV GL.




WP2: Reference wind energy systems (turbines and plants)

Period: Month 0-36

Description

Innovations, improved methods, and research findings are often made at the level of a single component, or discipline.  One must then determine what the positive impact will be on the rest of the system if any.  If this evaluation can be made by modifying a reference system, whose behavior has been well characterized, then the results may be readily understood, and compared against other relevant results.  This helps to solidify the scientific basis, such as reproducibility and benchmarking, of the findings.  Also, from a purely practical standpoint, it is not possible that every publication related to wind power plants should include a complete description of the plant used for case studies, therefore reference configurations are needed.

Reference wind power plants and turbines were discussed at the IEA Wind Technical Experts Meeting #80 on Wind Energy Systems Engineering, held January 12th-13th, 2015, the IEA Wind Task 37 Kick-off Meeting, held Sept. 16th-18th, 2015, and the IEA Wind Task 37 Annual Meeting, October 3rd-4th, 2016.

 

For reference turbines, there was broad agreement on the typical application areas and top-level characteristics of two reference wind turbines:

  • Onshore locations with moderate winds: a 3.4 MW IEC Class III turbine, with a large rotor for a low specific rating/high capacity factor, and a geared drivetrain.

  • Next generation of offshore wind turbines: a 10 MW Class I turbine with a direct-drive configuration, where the rotor blades are stretched in comparison with the existing DTU 10 MW turbine.

 

The aeroelastic design of the 3.4 MW turbine is conducted by TU Munich, and the 10 MW turbine by DTU Wind Energy.  The reference wind turbines are developed using an iterative design process. The aeroelastic design of the rotor is conducted with preliminary definitions of the drivetrain, tower, electrical system, and controller.  After the aeroelastic design is considered acceptable, this is passed to other groups (NREL, Cener, TU Delft, Sandia, and others are among the participants) who design the systems, and verify compatibility with the aeroelastic definition.  This process is representative of good practices; it is not intended to be a global optimization.

 

The task participants will revisit the plan at the end of year 2 to assess the need for additional reference turbines.  In particular, a 5+ MW Class I turbine for land-based wind farms with land area constraints was identified as a potential market segment for which a reference design might be useful.

 

For reference wind plants, the short-term objectives will focus on 1) providing a reference plant to go with each of the reference turbines (a low-wind onshore plant and an offshore plant for the next generation offshore technology), and 2) compiling wind plant data from existing data collection campaigns at current and future wind plant sites to put into the common reference plant framework for use by the larger community.  

 

The definition of reference wind power plants will also be executed as an international cooperation between research institutes, industry, based on ongoing research projects working on reference plants (NOWITECH, NORCOWE, EERA-DTOC) as well as coordination with the reference turbine development process. The goal for each wind plant developed will be to adopt an approach using layers (similar to current wind plant modeling tools) so that the user may select which layers to use for a particular analysis.  Example layers will likely include flow conditions, terrain, soil conditions, water depth (for offshore), noise restrictions, set-backs, built environment, the number, placement, type and hub height of the wind turbines, balance of plant and operations & maintenance (IO&M) features including weather windows and grid interface.

 

From this effort, three or four reference wind plants are expected to be developed in years 1 and 2 of the task:

  • One or two small-scale reference wind plants based on existing sites with significant data collection efforts (i.e. Alpha Ventus, Middlegrunden)

  • One low-wind onshore site of around 50-100 MW with a complex terrain layer available that will use the low-wind 3 MW reference turbine.

  • One high-wind offshore site of 500 MW+ with a TBD support structure configuration that uses the offshore 10 MW reference turbine.

Tasks:

  • Task 2.0.0 Establish an online repository for the storage and exchange of data.

  • Task 2.1: Reference wind turbines

    • Task 2.1.0 Specify a common data format for exchanging aeroelastic/control/electrical descriptions of onshore and offshore wind turbines, suitable for building models in typical wind turbine simulation programs.

    • Task 2.1.1, 3.4 MW Low-wind Onshore Reference Turbine Development

      • Task 2.1.1.1 Define design specifications for a 3.4 MW reference wind turbine with a geared drivetrain, targeting the onshore/Class III market segment.  Design specifications include the conditions and load cases to be used in the design process; the rotor diameter (rated power to swept area ratio); the tip speed (related to noise constraints).  The specifications will also identify component mass targets, such as blade and rotor/nacelle assembly mass, where published data on industry trends is available.  Finally, the specifications will indicate a list of parameters required as a result of the initial design, suitable for building basic aeroelastic, control, and electrical power models in typical wind turbine simulation programs (lead: SINTEF with collaboration from DTU Wind Energy, NREL, TUM, ORE and industry oversight by Nordex, Vestas, Siemens, GE and DNV GL).

      • Task 2.1.1.2 Upscale an existing 2.4 MW direct-drive wind turbine design to the 3.4 MW range using established procedures (lead: TU Munich)

      • Task 2.1.1.3 Design the reference 3.x MW Class III geared wind turbine.  The design process will be collaborative and iterative.  The goal is a process which represents best practices, and integrates the component models in the design loop.  Responsible parties (coordinating the design) are

        • Rotor design: TUM with collaboration from DTU Wind Energy, NREL and other groups

        • Aeroelastic design and loads analysis: TUM with collaboration from DTU Wind Energy, NREL and other groups

        • Control: TUM with collaboration from DTU Wind Energy, NREL and other groups

        • Gearbox and nacelle: NREL lead

        • Electrical systems: SINTEF lead

        • Foundation: TUM lead with input from Nordex on typical stiffness matrix

        • Cost model: lead by ORE with input from TUM, SINTEF, NREL and others

      • Task 2.1.1.4 Design review and approval by OEM industry participants (Nordex, Vestas, Siemens, GE and DNV GL)

    • Task 2.1.2 10 MW offshore reference turbine

      • Task 2.1.2.1 Define design specifications for a 10 MW reference wind turbine with a direct-drive generator, targeting the offshore/Class I market segment; the scope of the design specifications is the same as Task 2.1.1 (lead: SINTEF with collaboration from DTU Wind Energy, NREL, TUM, ORE and industry oversight by Nordex, Vestas, Siemens, GE and DNV GL).

      • Task 2.1.2.2 Select the starting point for the 10 MW Class I direct-drive wind turbine, utilizing the results of the INNWIND and Avatar projects.

      • Task 2.1.2.3 Design the reference 10 MW Class I direct-drive wind turbine.  Responsible parties are:

        • Rotor design: DTU Wind Energy with collaboration from TUM, NREL and other groups

        • Aeroelastic design and loads analysis: DTU Wind Energy with collaboration from TUM, NREL and other groups

        • Control: DTU Wind Energy with collaboration from TUM, NREL and other groups

        • Direct-drive generator and nacelle: NREL lead

        • Electrical systems: SINTEF lead

        • Foundation: DTU lead with input from Nordex on typical stiffness matrix

        • Cost model: lead by ORE with input from TUM, SINTEF, NREL and others

      • Task 2.1.2.4 Design review and approval by OEM industry participants (Vestas, Siemens, GE and DNV GL)

    • Task 2.1.3 Develop and publish IEA Wind Task Report on the reference turbine survey and both reference turbine technical specifications

    • (Tentative) Task 2.1.4 Define high-fidelity models of relevant components, for instance 3D blade and nacelle models suitable for CFD, 3D geometry of the gearbox components, or high-frequency models of the electrical converter and transformer.  

    • (Tentative) Task 2.1.5 Plan for second phase of work involving higher fidelity models for turbine.

  • Task 2.2: Reference wind plants

    • Task 2.2.0 Catalogue offshore and onshore wind plants where we know we have data and identify what types of data are available for each

    • Task 2.2.1 Perform stakeholder survey of use cases for reference plants and associated data requirements

    • Task 2.2.2 Specify a common data format for exchanging wind plant designs

    • Task 2.2.3 Translate data from selected existing wind plants into common data format

    • Task 2.2.4 Select and establish plant design criteria for a series of reference wind plants

    • Task 2.2.5 Develop reference wind plant 1 (low-wind onshore site)

    Responsible parties are:

  • Determine Baseline (lead DTU)

  • Layout design: lead by DTU with collaboration by  NREL, DNV GL, Nordex

  • LCOE analysis and optimization (IO&M): led by ORE with collaboration Sintef/Norway, ORE Catapult, DNV GL, NREL, DTU, Nordex, Delft, Stuttgart, CRES

  • Control strategy: led by DNV GL with collaboration from SINTEF, NORCOWE, NREL, Stuttgart, DTU, Nordex, CRES

    • Task 2.2.6 Develop reference wind plant 2 (high-wind offshore site)

Responsible parties are:

  • Baseline: NORCOWE Reference Wind Farm

  • Layout: NORCOWE, Sintef, Delft, DTU, NREL, DNV GL, ORE Catapult, Stuttgart

  • LCOE analysis and optimization (IO&M): led by ORE with collaboration from  Sintef/Norway, ORE Catapult, DNV GL, NREL, DTU, Nordex, Delft, Stuttgart, CRES

  • Control strategy: led by DNV GL with collaboration from SINTEF, NORCOWE, NREL, Stuttgart, DTU, Nordex, CRES

  • Electrical: SINTEF, NORCOWE, DNV GL, (DTU)

  • Foundations:

 

Deliverables:

  • D2.1.1 Specifications document for the 3.4 and 10 MW reference wind turbines

  • D2.1.2 Publication of the refined 3.4 MW geared wind turbine design

  • D2.1.3 Publication of the refined 10 MW direct-drive wind turbine design

  • D2.2.1 Specifications document for the reference wind plants

  • D2.2.2 Publication of reference onshore plant 1

  • D2.2.3 Publication of reference offshore plant 2




WP3: MDAO case study activities at different system levels (turbines and plants)

Period: M1-M36

Description

The goal of this work package is to demonstrate the value of systems engineering / MDAO over traditional design approaches and to educate the industry and research community about best MDAO practices applied to wind energy. This will be achieved through a series of case study problems defined collaboratively by the project participants.

 

In this work package the case study activities relating to both turbine and plant optimization will be organized and carried out. The scope of these problems will be established with the help of a participant survey to provide an overview on state-of-the-art MDAO research for wind energy applications.  This information will be cataloged and stored in an online portal / information clearinghouse that is accessible to the public.  In addition, the survey will investigate what benchmarking has been done in MDAO in other fields (in aerospace, etc) and how they are conducted.

 

Also in year one, an overall process and evaluation criteria for the work package will be established and a plan for the first phase of turbine and plant case  studies will be completed.  The evaluation criteria used will be based on a common definition, established in the first part of the project. A common platform/programming agnostic file format that input data and results should be delivered in will also be established in WP1 on frameworks.

 

A series of case study problems.  The types of problems considered can be seen falling on a 3-axis graph:

 

Screen Shot 2015-10-06 at 19.20.37.png

Graph representing the three dimensions which the MDAO case studies will explore.

 

Due to the variety of possible approaches to MDAO for wind turbines and plants and the different possible fidelity levels that can be used, participants will be left to choose their approaches and models.  The case study exercise will only constrain the system scope and problem definitions. This way, discussion and comparison of approaches using different techniques and levels of fidelity will be possible.  The exact problems performed in the first 3 years of the task (2 each for the turbine and the plant) will be decided in the year 1 planning activity.

 

At the turbine level, problems can be sub-divided into single discipline and multi-disciplinary problems.  While multi-disciplinary problems are the focus of this work package and task, single disciplinary problems can be used as a precursor to MDAO problems and will enable the comparison of the MDAO to single disciplinary approaches.  A set of preliminary benchmarks will likely include: single-discipline, rotor only and full turbine MDAO problems. Below, is a list of proposed problem categories is given:

 

Single-discipline optimization problems:

  • Optimize wrt blade planform with a fixed absolute thickness distribution,

  • Optimize sizing of other components structure/design given load set,

  • Minimize fatigue wrt controller parameters with a fixed geometry.

 

Rotor MDAO problems:

  • Optimize wrt blade planform and internal structure,

  • Optimize wrt free form blade shape and internal structure.



Turbine MDAO problems:

  • Optimize wrt rotor and controller,

  • Optimize wrt rotor and drivetrain,

  • Optimize wrt rotor, drivetrain, control, substructure,

  • Optimize wrt turbine configuration.

 

The case studies will be planned to increase in complexity over time - starting from a fairly simple low-dimension problem and expanding to higher-dimension problems with more disciplines over time.  To the largest extent possible, the case studies will be defined to allow models of different fidelity to participate e.g. BEM-based or CFD-based aerodynamics, simple cross-sectional structural models or full FEM models etc.

At the plant level, the potential plan for activities are divided in three tracks:

  • Minimize the COE or maximize the AEP of the wind plant with respect to:

    • Wind farm layout

    • Wind plant control strategy

    • Wind turbine type

    • All combined

  • Uncertainty Quantification of wind plant COE

  • Minimize the COE or maximized the AEP of the wind plant taking into account the uncertainties



Tasks

  • Task 3.0.1: Survey MDAO research and catalogue on information web portal

  • Task 3.0.2: Establish benchmarking evaluation criteria, scope and overall process

  • Task 3.1: Benchmarking MDAO for wind turbines

    • Task 3.1.1: Phase 1 benchmarks: Rotor only

      • 3.1.1.1: Benchmarking of rotor aero only, responsible DTU. Based on a modified DTU 10 MW rotor, maximise power production at a single wind speed subject to constraints on blade root moment, rotor thrust, absolute thickness distribution, wrt blade chord, twist, relative thickness.

      • 3.1.1.2: Benchmarking of rotor aero and structure, responsible TUMDTU. Based on a modified DTU 10 MW rotor geometry and structure, minimise mass, subject to constraints on material ultimate and fatigue strength (with prescribed loads envelopes), tip deflections (prescribed loads distribution), material thicknesses and material thickness tapering, flapwise and edgewise frequencies, wrt distribution of thicknesses of spar cap, leading and trailing edge reinforcements, shell and shear webs.

      • 3.1.1.3: Benchmarking of rotor aero and structure, responsible DTU. This case will be based on the design study done for the 10 MW RWT  developed for Task  2.1.2.3, in which a 10 MW rotor was designed starting from a slightly modified DTU 10MW RWT, for which a simplified cost metric consisting of rotor AEP and blade mass was minimised, subject to constraints on a series of blade and turbine loads, material strength, tip deflection, as well as various blade geometry and structural constraints, wrt blade chord, twist, thickness, prebend, distribution of thicknesses of internal structural geometry: spar cap, leading and trailing edge reinforcements, and shell. The participants will as such be able to compare their designs with the Task 2.1.2.3 10 MW designed by DTU using the same problem formulation.

      • 3.1.1.4 Develop and publish report on Rotor optimization case studies

    • Task 3.1.2: Phase 2 benchmarks: full turbine

      • 3.1.2a Benchmarking of full turbine TBD

  • Task 3.2: Benchmarking MDAO for wind plants

    • Task 3.2.1 Layout optimization onshore - lead DTU (with NREL, DNV GL, Nordex, DTU)

    • Task 3.2.2 Layout optimization offshore - lead Norway (with Sintef/Norway, Delft, DTU, NREL, DNV GL, ORE Catapult, Stuttgart)

    • (Tentative) Controls optimization - lead  DNV GL (with Sintef/Norway, DNV GL, NREL, Stuttgart, DTU, Nordex, CRES)

    • (Tentative) Electrical analysis and optimization - lead Norway (with Sintef/Norway, DNV GL, (DTU))

    • (Tentative) LCOE analysis and optimization (O&M) - lead ORE / Catapult (with Sintef/Norway, ORE Catapult, DNV GL, NREL, DTU, Nordex, Delft, Stuttgart, CR)

 

Deliverables

  • D3.0.1: Online portal / information clearinghouse for MDAO research and software

  • D3.0.2: Report on benchmarking scope, process and evaluation criteria

  • D3.1.1: First turbine benchmark finalized and reported

  • D3.1.2: First plant benchmark finalized and reported

  • D3.2.1: Second turbine benchmark finalized and reported

  • D3.2.2: Second plant benchmark finalized and reported

Chronogram and Key Dates

Beginning on the date,  this Annex is formally initiated and shall in principle continue for a period of three years.    The proposed time schedule is presented in Figure 2. Note that the Work Package descriptions are given in section 4 and the description of the milestones and deliverables is given in sections 6 and 7.

 

IEA Wind Task 37: Wind Energy Systems Engineering

Work Plan

Project start: October 2015

Year 1

Year 2

Year 3

Q1

Q2

Q3

Q4

Q1

Q2

Q3

Q4

Q1

Q2

Q3

Q4

Work Package 1: Guidelines on a common framework for wind energy systems

Task 1.1 Survey of existing frameworks (wind and more general systems engineering frameworks) and summary

Task 1.2 Integration of reference system data

Task 1.2.1 Integration of reference turbine data exchange formats into framework guidelines

Task 1.2.2 Integration of reference plant data exchange formats into framework guidelines

Task 1.2.3 Harmonization of data exchange formats into final framework

Task 1.3 Develop common data exchange format (code agnostic)

Work Package 2: Reference wind energy systems

Task 2.0.0 Establish an online repository for the storage and exchange of data

WP 2.1 Reference Wind Turbines

Task 2.1.0 Common data format

Task 2.1.1.1 Specifications for a 3.4 MW land-based Class III turbine

Task 2.1.1.2 Upscale an existing 2.4 MW design to the 3.4 MW range

Task 2.1.1.3 Design the reference 3.4 MW Class III wind turbine

Task 2.1.1.4 Design review and approval of the 3.4 MW Class III wind turbine

Task 2.1.2.1 Specifications for a 10 MW offshore Class I turbine

Task 2.1.2.2 Select the starting point for the 10 MW Class I wind turbine

Task 2.1.2.3 Design the reference 10 MW Class I wind turbine

Task 2.1.1.4 Design review and approval of the 10 MW Class I wind turbine

WP 2.2 Reference Wind Plants

Task 2.2.0 Catalogue offshore and onshore wind plants where we know we have data and identify what types of data are available for each

Task 2.2.1 Perform stakeholder survey of use cases for reference plants and associated data requirements

Task 2.2.2 Specify a common data format for exchanging wind plant designs

Task 2.2.3 Translate data from selected existing wind plants into common data format

Task 2.2.4 Select and establish plant design criteria for a series of reference wind plants

Task 2.2.5 Develop reference wind plant 1 (low-wind onshore site)

Task 2.2.6 Develop reference wind plant 2 (high-wind offshore site)

Work Package 3: MDAO case studies for wind energy systems

3.0.1 Survey MDAO research and catalogue on information web portal

3.0.2 Establish benchmarking evaluation criteria, scope and overall process

WP 3.1 MDAO case studies for wind turbines

3.1.1 Phase 1 benchmarks: Rotor only

3.1.1a Benchmarking of rotor aero only

3.1.1a Case studies of rotor aero only

3.1.2 Phase 2 benchmarks: full turbine

3.1.2 Phase 2 case studies: full turbine

WP 3.2 MDAO case studies for wind plants

3.2.1 Layout optimization onshore

3.2.2 Layout optimization for an offshore plant

Figure 2: Part 1 of Gantt chart for Systems Engineering / Integrated RD&D Project




Reports and Deliverables

Annual progress reports will give an overview of the follow-up of the project. Within each Work Package a number of deliverables will be elaborated in order to summarize the most important results. These reports/deliverables will be composed by the Operating Agents based on the inputs and reviews from the Participants. The planned deliverables are given in Table 2.

The deliverables were described in the work packages in section 5 and are listed here as a complete set along with delivery date:

 

Table 2: List of deliverables and their duration

Deliverable

Month Due

WPO: Management

D0.1 First annual progress report

M12

D0.2 Second annual progress report

M24

D0.3 Final report

M36

WP1: Guidelines on a common framework for wind energy system modeling

D1.1 Requirements document including survey results

M12

D1.2 Document of framework guidelines for wind turbines and plants

M24

D1.3 Publication of common data format

M24

WP 2: Reference wind energy systems

D2.1.1 Specifications document for the 3.4 and 10 MW reference wind turbines

M6

D2.1.2 Publication of the refined 3.4 MW geared wind turbine design

M15

D2.1.3 Publication of the refined 10 MW direct-drive wind turbine design

M15

D2.2.1 Specifications document for the reference wind plants

M6

D2.2.2 Publication of reference onshore plant 1

M24

D2.2.3 Publication of reference offshore plant 2

M24

WP 3: Benchmarking MDAO for wind energy systems

D3.0.1 Online portal / information clearinghouse for MDAO research

M6

D3.0.2 Report on benchmarking scope, process and evaluation criteria

M12

D3.1.1 Report on turbine benchmarking activity 1 results

M24

D3.1.2 Report on turbine benchmarking activity 2 results

M36

D3.2.1 Report on plant benchmarking activity 1 results

M24

D3.2.2 Report on plant benchmarking activity 2 results

M36



Methods of Review and Evaluation of the Work Progress

The following key milestones are defined for the follow-up of the progress of the project:

M0.1 Kick-off meeting: Survey and reference system planning

M0.2 Progress Meeting 1. Workshop: Survey and reference system review; benchmarking kick-off

M0.3 Progress Meeting 2. Workshop: Framework, reference system and benchmarking review

Additional working meetings both in-person and via the Internet will be held as needed to support project collaboration.

References

Ashuri T, et al. Multidisciplinary Design Optimization of Offshore Wind Turbines for Minimum Levelized Cost of Energy. Renewable Energy 68 (2014)  893-905.  doi: 10.1016/j.renene.2014.02.045.

Ashuri T Beyond Classical Upscaling: Integrated Aeroservoelastic Design and Optimization of Large Offshore Wind Turbines. PhD thesis (2012). Delft University of Technology

Ashuri T, Zaaijer M, and et al.  Controller Design Automation for Aeroservoelastic Design Optimization of Wind Turbines. The Science of Making Torque from Wind (2010), Crete, Greece, 1-7.

Akhmatov V, et al.  Modelling and transient stability of large wind farms.  Electrical Power and Energy Systems 25 (2003) 123-144

Bak et al.  Description of the DTU 10 MW Reference Wind Turbine.  DTU Wind Energy Report-I-0092, Technical University of Denmark, 2013.

Barthlemie RJ, et al.  Modelling and measuring flow and wind turbine wakes in large wind farms offshore.  Wind Energy 12 (2009) 431-444.

Barthelmie, . et al. Modelling the impact of wakes on power output at Nysted and Horns Rev. In Proceedings of the European Wind Energy Conference, Marseille, France, 2009.

Bottasso, CL et al. "Optimization‐based study of bend–twist coupled rotor blades for passive and integrated passive/active load alleviation." Wind Energy 16.8 (2013): 1149-1166.

Bottasso, Carlo L, Filippo Campagnolo, and Alessandro Croce. "Multi-disciplinary constrained optimization of wind turbines." Multibody System Dynamics 27.1 (2012): 21-53.

Bozhko S, et al.  Control of offshore DFIG-based wind farm grid with line-commutated HVDC connection.  IEEE Transactions on Energy Conversion 22 (2007) 71-78.

Calaf M, et al.  Large eddy simulation study of fully developed wind-turbine array boundary layers.  Physics of Fluids 22 (2010) 015110-1 - 16.

M. Churchfield, et. al. A large-eddy simulation of wind-plant aerodynamics. In Proceedings of the AIAA Aerospace Sciences Meeting, Nashville, Tennessee, USA, 2012a.

Diveux T., Sebastian P. , and et al.  Horizontal axis wind turbine  systems: optimization using

   genetic algorithms. Wind Energy, 4 (4) (2001), pp. 151–171

Frandsen S, et al.  The making of a second-generation wind farm efficiency model complex.  Wind Energy 12 (2009) 445-458.

Fuglsang, Peter L, and Helge Aa Madsen. "Optimization of stall regulated rotors." Proc. 14. ASME Wind Energy Symp., 1995, pp. 151—158

Haghi R, Ashuri T, and et al, Integrated Multidisciplinary Constrained Optimization of Offshore      Support Structures, The science of Making torque from wind (2012). Journal of Physics, 555 012046. doi: 10.1088/1742-6596/555/1/012046

Haileselassie T, et al.  Main grid frequency support strategy for VSC-HVDC connected wind

farms with variable speed wind turbines.  IEEE PowerTech, Trondheim, Norway, 19-23 June, 2011, pp 1-6.

Herbert-Acero, José F et al. "A Review of Methodological Approaches for the Design and Optimization of Wind Farms." Energies 7.11 (2014): 6930-7016.

Larsen GC, et al.  TOPFARM - Next generation design tool for optimisation of wind farm topology and operation.  Report Risø-R-1805(EN), Risø DTU National Laboratory for Sustainable Energy, Denmark, 2011.

Liu B, et al.  Centralized power control strategy of offshore wind farm with permanent magnetic generators.  IEEE 6th International Power Electronics and Motion Control Conference, Wuhan, China, 17-20 May, 2009, pp 1075-1079.

Maki K., Sbragio R., and et al. System design of a wind turbine using a multi-level optimization

   approach. Renew Energy, 43 (12) (2012), pp. 101–110

Matevosyan J, Bolik SM, Ackermann T.  Technical regulations for the interconnection of wind power plants to the power system.  In: Ackerman T (editor).  Wind Power in Power Systems.  Second Edition, Wiley, 2012.

Mayda, E, et al, Wind turbine Rotor R&D an OEM perspective, International Conference on Future Technologies for Wind Energy, University of Wyoming, 2013

Merz, Karl O. "Rapid optimization of stall‐regulated wind turbine blades using a frequency‐domain method: Part 2, cost function selection and results." Wind Energy (2014).

Ning, S Andrew, Rick Damiani, and Patrick J Moriarty. "Objectives and constraints for wind turbine optimization." Journal of Solar Energy Engineering 136.4 (2014): 041010.

Øye, S. "FLEX4—Computer code for wind turbine load simulation." Technical University of Denmark (1992).

Petrini F, Manenti S, and et al. Structural design and analysis of offshore wind turbines from a       system point of view. Wind Eng, 34 (1) (2010), pp. 85–108

Réthore P-E, et al.  A CFD model of the wake of an offshore wind farm: using a prescribed wake inflow.  Journal of Physics: Conference Series 75 (2007) 012047-1 - 7.

Réthoré, Pierre‐Elouan et al. "TOPFARM: Multi‐fidelity optimization of wind farms." Wind Energy 17.12 (2014): 1797-1816.

Schaak P.  Heat & Flux – Enabling the wind turbine controller. Report ECN-C--06-017, Energy Research Centre of the Netherlands, 2006.

Selig, Michael S, and Victoria L Coverstone-Carroll. "Application of a genetic algorithm to wind turbine design." Journal of Energy Resources Technology 118.1 (1996): 22-28.

Senjyu T, et al.  Output power leveling of wind turbine generators using pitch angle control for all operating regions in wind farm.  Electrical Engineering in Japan 158 (2007) 31-41.

Soleimanzadeh M, Wisniewski R.  Controller design for a wind farm, considering both power and load aspects.  Mechatronics 21 (2011) 720-727.

Steinbuch M, et al.  Optimal control of wind power plants. Journal of Wind Engineering and Industrial Aerodynamics 27 (1988) 237-246.

van Dam FC, et al.  A maximum power point tracking approach for wind farm control. Proceedings of the Science of Making Torque from Wind, Oldenburg, Germany, October 9-11, 2012.

Wei M, Chen Z.  Intelligent control on wind farm.  IEEE Innovative Smart Grid Technologies Conference Europe, Gothenburg, Sweden, 11-13 October, 2010, pp 1-6.

Public

Files

Loading widget...
No files to show
{{node.name}}
({{node.children.length}})
{{node.date}}
{{node.modified}}
{{node.filesize}}
{{node.filename}}