The Advanced Research Projects Agency – Energy (ARPA-E) of the U.S. Department of Energy is challenging the research and industrial communities to discover, develop, and test innovative and disruptive Control Co-Design solutions for critical wind energy challenges. The ATLAS (Aerodynamic Turbines with Load Attenuation Systems) Competition is the first ARPA-E effort associated with this advanced design methodology.
The competition seeks to design individual and collective pitch control systems for large floating offshore wind turbines (Offshore Challenge) affected by asymmetric and unpredictable wind and wave fields. An additional land-based wind turbine case (Onshore Challenge) is also proposed as a simpler case and a potential training step in the process.
In both Challenges, the primary objective is to reduce the mechanical fatigue of the main system components without compromising the Annual Energy Production (AEP). Leveraging algorithms that are successful, future designers would be able to propose more reliable, more resilient and more optimal turbine designs, reducing the associated Levelized Cost Of Energy (LCOE) of the wind turbine.
Collective pitch control (CPC) regulates the rotor speed (Wr) of wind turbines when the turbine is experiencing wind conditions above the rated speeds –i.e., with wind velocities typically above 11 m/s. CPC changes the pitch angle of the three blades (b1, b2, b3) identically and simultaneously to regulate the rotor speed under variable wind –see Fig.1. CPC is the common solution applied in commercial multi-megawatt turbines. However, CPC does not consider the asymmetry of the wind velocity profile over the rotor area, which could introduce large unbalanced loads in the system.
Individual pitch control (IPC) is an innovative technique that considers these wind asymmetries. It has the potential to reduce the associated asymmetric mechanical loads on the blades, minimizing the mechanical fatigue of the structure (nacelle, tower, floating platform, mooring system, etc.) without significantly compromising the average of power generation. In addition to the CPC signal (bcpc), the IPC adds to each pitch motor an individual control signal (bipc1, bicp2, bipc3) to reduce the asymmetric loads. Calculations of these individual control signals can be based on the information provided by sensors that measure directly or indirectly the bending moments of each blade (M1, M2, M3), the rotor azimuth (f), and advanced control algorithms –see Fig.1.
Figure 2 exemplifies a general methodology to design wind turbines. The figure includes a representation of the sub-systems of a floating offshore wind turbine: rotor, drive-train, electrical generator, power electronics, substation, nacelle, tower, platform, mooring system, aerodynamics, hydrodynamics, and grid and control systems. It also shows the inputs: wind, waves, grid voltage and frequency, etc. The figure emphasizes the multiple sub-system interactions.
Control Co-Design is a powerful methodology to find optimal designs when significant sub-system interactions are present. It uses control principles, co-optimization techniques and co-simulation methods, starting at the very beginning of the design process. ARPA-E has recently released (1) a Request for Information, (2) a Teaming Partner List, and (3) plans to organize an Industry Day on January 15th, 2019, for a possible future program on Control Co-Design of Floating Offshore Wind Turbines.
The ATLAS Competition is designed to immerse participants in the earlier part of this iterative control co-design process, as shown in the red-shaded area of Fig. 2. It encompasses the first part of the iteration, with the design of an innovative control solution (IPC+CPC) and the simulation of the system dynamics (wind turbine) under a variety of input scenarios (wind profiles, waves and parametric uncertainty).
The scope of the ATLAS Competition covers the process until just before the re-design diamond block. It includes only the design of the IPC+CPC control solution to reduce the mechanical fatigue of the main components of the wind turbine under a variety of wind and wave conditions.
The land-based wind turbine (Onshore Challenge) and floating offshore wind turbine (Offshore Challenge) simulations of the ATLAS Competition will be performed using the OpenFAST computer program developed by the National Renewable Energy Laboratory (NREL) –see Fig.3. The participants will have access to OpenFAST, the wind turbines models and a set of inputs (Set A) that includes a variety of 3D wind velocity profiles and parameters. The Offshore Challenge also includes sea conditions (waves). The evaluation of the competition will run a super set of inputs (Set B) that includes Set A and other additional scenarios defined for the same wind class and sea state.
The competition participants will submit the control algorithms as a Simulink block with associated m-files or p-files –see IPC and CPC block in Fig.1 and Pitch Controller block in Fig.4 (note that all the units are in the metric system). Each participant must also submit a report explaining the uniqueness of the technological approach applied to develop the underlying algorithms. The report must not exceed four pages and can contain the level of detail chosen by the participant. The report is necessary to participate in the competition. However, evaluation and scoring will be based solely on the performance of the control solution across all the simulated cases (Set B).
In both Challenges, the control solutions will be applied to the well-known and well-characterized NREL 5MW reference wind turbine, and will be tested with the OpenFAST simulator under non-homogeneous 3D wind profiles and different wind turbine cases within the model parametric uncertainty, as well as additional sea scenarios in the Offshore Challenge.
Every practical design of a wind turbine uses a limited number of cases to understand the very large number of potential scenarios that the real-world system contains. In the same way, a subset (Set A) of the superset (Set B) in Fig.3 will be provided to participants. Boundary conditions for Set B will be shared. The competition will evaluate the reduction of tilt/pitch and yaw loads over the structures and flap-wise loads over the blades, all without compromising the average power generation.
 Jonkman, J. (2015). “FAST | NWTC Information Portal.” Retrieved from https://nwtc.nrel.gov/FAST.
 Jonkman, J. (2006). “NREL Offshore Baseline 5MW.” Golden, Colorado: National Renewable Energy Laboratory. Retrieved from https://www.ieawind.org/AnnexXXIIISecure/Subtask_ 2S_docs/OC3Files/BaselineTurbine/NRELOffshrBsline5MW.pdf
 Wheeler L. and Garcia-Sanz M. (2017). “Wind turbine collective and individual pitch control using quantitative feedback theory”. ASME 2017 Dynamic Systems and Control Conference, Tysons Corner, Virginia, USA.
This section describes the main rules and guidelines for the ATLAS competition. For additional information, please see the document named “ATLAS-modeling-control-simulation.pdf.”
R1. The ATLAS Competition seeks designs of advanced individual pitch control (IPC) and collective pitch control (CPC) systems for two wind turbines affected by asymmetric and unpredictable wind fields: (1) a land-based wind turbine –Onshore Challenge, and (2) a floating offshore wind turbine –Offshore Challenge.
Note: R2-R4 are excluded as they only apply to the Onshore Challenge.
R5. The turbine model for the Offshore Challenge is also the NREL 5MW baseline turbine, now attached to the OC3 Hywind Spar floating platform (floating offshore wind turbine).
R6. The Offshore Challenge is open to U.S. Entities. A U.S. Entity is U.S. Citizen, U.S. Permanent Resident, or U.S. Organization (i.e., a U.S.-based nonprofit or for-profit institution, college, or university). In this Offshore Challenge, a “Participant” is defined as the U.S. Entity. All Participant work related to the Offshore Challenge must be performed in the U.S. An individual, organization, or sponsor cannot participate in the ATLAS Competition if he, she, or it is on the Specially Designated Nationals list.
R7. Awards for the Offshore Challenge include prize money for those Participants placed first, second, or third. We anticipate $100,000 (one hundred thousand dollars) for first place, $50,000 (fifty thousand dollars) for second place and $25,000 (twenty-five thousand dollars) for third place.
R8. The participants must implement and submit their CPC and IPC solutions as Simulink models that include the original (un-modified) NREL 5MW baseline generator torque controller supplying generator torque commands to the OpenFAST S-Function in Simulink. The Participants are not allowed to modify or replace the NREL 5 MW baseline generator torque controller supplied (otherwise, the submission will be disqualified).
R9. The Participants can modify the blade pitch controller. They can incorporate both collective pitch control (CPC) modifications and individual blade pitch control (IPC), either by retaining the NREL baseline pitch controller and using their new control solution to add incremental pitch commands to the baseline, or completely replace the NREL 5MW baseline pitch controller (CPC) with their own controller (CPC and IPC). The participants are required to design and implement an IPC as part of their submitted control solution.
R10. The Participants should not modify the set of sensor signals already provided in the OpenFAST model downloads (otherwise, the submission will be disqualified). These sensor signals are defined through the OpenFAST module input files and must not be modified. For more information, please see the document named “ATLAS-modeling-control-simulation.pdf.”
R11. The pitch actuator model contained in the NREL 5MW Baseline Simulink model should be considered fixed, as it is part of the turbine model, and should not be modified by the competition participants. No changes should be made to the pitch actuator blocks or the values of parameters used in these blocks (otherwise, the submission will be disqualified). For more information on the pitch actuator, please see the document named “ATLAS-modeling-control-simulation.pdf.”
R12. From January 11th 2019 to April 19th 2019 and through the competition website, the Participants will have access to OpenFAST (executables and Matlab-mex functions), NREL 5MW models, wind and wave input files, and the OpenFAST Simulink model of NREL 5MW controlled by baseline controller. These executables are compiled for the Windows 64-bit operating system and for Matlab 2018. Participants may use the NREL downloads described in these rules, subject to any restrictions in the applicable license(s).
R13. Participants of both Challenges are allowed to submit their advanced individual/collective pitch control solutions from January 11th 2018 to April 19th 2019 (5:00 pm ET).
R14. Participants will submit the control algorithms as a Simulink block with associated m-files. For proprietary controllers, the participants will substitute the m-files with p-files. P-files are produced from Matlab m-files using the Matlab “pcode” command, which obfuscates the contents of the m-file. These Simulink models must be able to be run through the post-processing files by the Competition judges without errors.
R15. Participants are also required to submit a report – up to four pages in length -- along with the Simulink model and m-files (or p-files) explaining the uniqueness of the technological approach applied to develop the underlying algorithms. The report will not be a factor in the evaluation and scoring of a Participant’s control solution, but Participants must submit a report in order to be eligible for a prize.
R16. Each individual or team must submit its own unique technological (algorithmic) approach. ARPA-E and NREL reserve the right to perform technical due diligence to verify that each participant’s approach represents a unique algorithmic approach. This due diligence may include, but is not limited to, inspection and/or validation of the underlying source code.
R17. The maximum number of submissions per Participant per Challenge is two (see definition of “Participant” in Rules R3 and R6). Only the submission with the best score of these two in each Challenge will be considered for prizes and leaderboard purposes.
R18. The Competition judges (made up of ARPA-E and NREL staff) will simulate the Participants’ controllers under the Windows 64-bit operating system using Matlab 2018. Participants must submit their Simulink controllers so that they can be simulated under this operating system and this version of Matlab. The judges will not perform any Simulink model conversion to other version of Matlab or operating systems.
R19. Submissions will be scored based on a cost function value (single number). This controller performance cost function metric will be based on mechanical load results over a span of above rated operating conditions (Region 3), including turbulent wind inflow cases and a few deterministic cases. The Participants will be given a subset of this load suite (Set A) which can be used to evaluate the performance of their control solutions while under development.
R20. The cost function software (post-processing file) used to determine the scores for the submissions will be available on the ATLAS Competition website during the Competition. It will calculate the performance metrics, including mechanical fatigue and ultimate loads on blades, tower and other main turbine components, and power generation. The winning control solution will be the one that most reduces the mechanical fatigue while maintaining the average power generation across the Set B. For additional information on this cost function, please see also the document named “ATLAS-modeling-control-simulation.pdf.”
R21. In order to be eligible for a prize, a submission must place in the applicable spot on the leaderboard for each Challenge (see Rules R4 and R7 above) and must also score lower than the NREL 5MW baseline controller cost function metric (lower scores indicate better control performance relative to the baseline). The cost function has been normalized so that the NREL 5MW baseline controller gives a performance metric of one (=1) in both Challenges.
R22. The performance of each submission will be evaluated only once.
R23. Simulation and scoring of submitted controllers will be automatically carried out by a computer algorithm at NREL and periodically posted at the ATLAS Competition website. ARPA-E and NREL staff reserve the right to perform technical due diligence on each submission to verify performance and the uniqueness of the approach.
R24. In both Challenges, the names, affiliations and results of the participants (students, team of students or U.S. Entities) placing in the top-ten will be updated on the ARPA-E leaderboards and competition website every two weeks approximately, from January to April.
R25. Foreground Intellectual Property. Neither the Government, the Department of Energy, ARPA-E, NREL, nor anyone acting on their behalf will obtain any rights in intellectual property developed by a Participant in the Competition without the written consent of the Participant.
R26. Confidentiality of Participant Submissions. All technical information submitted to NREL by any Participant for the purpose of competing in the prize will remain confidential, if marked as confidential, e.g., such as “Do Not Publicly Release – Confidential Proprietary Business Information,” except for data that will be publicly releasable, as set forth in this paragraph and in Rule R27. Unmarked data delivered to NREL will be made publicly available as NREL and ARPA-E deem appropriate. The report explaining the uniqueness of the technological approach (described on page 3 and in paragraph R15) required in a Participant submission may be released publicly by ARPA-E unless it is marked as confidential by the Participant.
R27. Publications. ARPA-E intends to publish, i.e., make publicly available, the description, development and results of the Competition. ARPA-E has the right to publish the simulation results of each submission. ARPA-E will be the leader in any potential publication related to the Competition, including journals, conferences and other public platforms. Any proposed publication by a Participant relating to a submission or the competition must be approved in advance by ARPA-E.
ARPA-E may modify these rules at any time and for any reason. Please consult the ATLAS Competition website and the most up-to-date Rules Document.
Except where prohibited, participation in the ATLAS Competition constitutes each Participant’s consent to DOE’s and its agents’ use of each Participant’s name and/or hometown and state information for promotional purposes through any form of media, worldwide, without further permission, payment or consideration.
ARPA-E reserves the right to disqualify Participant whose actions are deemed to violate these rules or the spirit of the ATLAS Competition for any reason, including but not limited to, the violation of relevant laws or regulations in the course of participation in the ATLAS Competition. No appeal is allowed and the decision is at the sole discretion of ARPA-E.
By registering for and/or participating in the ATLAS Competition, Participants agree to follow these rules. Participants must agree to assume any and all risks and waive claims against the Federal Government, its related entities, and NREL, except in the case of willful misconduct, for any injury, death, damage, or loss of property, revenue, or profits, whether direct, indirect, or consequential, arising from participation in the ATLAS Competition, whether the injury death, damage, or loss arises through negligence or otherwise.
ALL DECISIONS BY ARPA-E ARE FINAL AND BINDING IN ALL MATTERS RELATED TO THE ATLAS COMPETITION.
ARPA-E reserves the right to cancel, suspend, and/or modify the ATLAS Competition, any part of the ATLAS Competition, or any associated events or funding opportunities at ARPA-E’s sole discretion. Considerations may include availability of funds and technical viability, or if any fraud, technical failures, or any other factor beyond ARPA-E’s reasonable control impairs the integrity or proper functioning of the ATLAS Competition. ARPA-E is not responsible for, nor is it required to accept, incomplete, late, misdirected, damaged, corrupted, unlawful, or illicit submissions.
ARPA-E does not authorize or consent to a Participant infringing on any U.S. patent or copyright while participating in the ATLAS Competition. No illegal activities may be undertaken for the purpose of participation in the ATLAS Competition.
Participant Representations and Warranties
Upon registration, Participant hereby represents and warrants on its behalf and on behalf of all team members, that:
Government Representations and Warranties
Any and all information provided by or obtained from the Federal Government is without any warranty or representation whatsoever including, but not limited to, its suitability for any particular purpose.
Personally Identifiable Information
Unless otherwise required, all information provided by the Participant must not include any Personally Identifiable Information (PII). PII refers to information which can be used to distinguish or trace an individual’s identity, such as the individual’s name, social security number, date and place of birth, personal telephone or cell phone number, or any other information this is linked or linkable to an individual such as employment information.
Relationship of the Parties
Nothing contained in this Rules Document is intended to create or constitute a relationship between ARPA-E or NREL with the Participants. Participation in the ATLAS Competition does not constitute or imply any form of sanction or support or endorsement of the Participant by ARPA-E or NREL, nor does it grant any party any authority to act as agent, nor assume or create any obligation, on behalf of another party. A Participant may not use the ARPA-E or DOE logos or official seals in their submission. APRA-E reserves the right to use Participant names and images for publicizing the ATLAS Competition and results.
ATLAS Competition and any associated nicknames and logos (“Competition Marks”) are trademarks owned by the U.S. Department of Energy. The trademark license granted to Participants is below. Non-Participants can request individualized trademark licenses (for the purpose of engaging with Participants and/or expressing interest in the technology); the decision to grant such licenses is under the sole discretion of ARPA-E.
Participants are granted, for the duration of the ATLAS Competition, a revocable, non-exclusive, royalty-free license to use the Competition Marks for the purposes of producing materials for the competition and other approved competition-related activities as long as the use does not suggest or imply endorsement of the Participant by ARPA-E or the U.S. Department of Energy, and the use of the Competition Marks by a Participant does not imply the endorsement, recommendation, or favoring of the Participant by ARPA-E or the Department of Energy.
Participants may not use the Competition Marks for any other purpose. Participants may not sublicense the Competition Marks.
All Participants can request individualized trademark licenses; the decision to grant such requests is under the sole discretion of ARPA-E.
 An individual or entity participating in the Offshore Challenge, may not be: 1) a Federal entity or employee acting within the scope of their employment or 2) a DOE employee, NREL employee, members of their immediate family (spouses, children, siblings, parents) or persons living in the same household as such persons, whether or not related.