Title: Design of Lidar, Data Processing, and Control Algorithms for Optimal Wind Farm Performance PI: Rod Frehlich, Cooperative Institute for Research in Environmental Sciences (CIRES) Co-Investigator: Lucy Pao, UCB Engineering Collaborators: Neil Kelley, Bonnie Jonkman, Patrick Moriarty, Alan Wright, NREL Bob Banta and Yelena Pichugina, NOAA/CIRES PhD Student: Jason Laks Rod Frehlich, Jason Laks, and Lucy Pao Abstract: Current wind energy farms are often not operated at full efficiency because of a lack of real-time measurements of critical atmospheric variables, especially the vertical profiles and spatial variations of velocity turbulence. Substantial improvements in wind energy production are possible if reliable atmospheric conditions or signals related to them can be input into optimal control algorithms to adjust the blades of each wind turbine. The expected benefit is a significant reduction in turbine maintenance with a commensurate increase in component lifetime and turbine reliability by the mitigation of the sources of the severe vibration and loading on the turbine blades and support structure. The main goal of the proposed research has been to demonstrate with computer simulations an optimal control system for wind farms based on input from advanced processing of lidar data using an autonomous commercially available Doppler lidar (see Figure 1). This has been a collaboration between three groups: CIRES (Frehlich and Meillier), NREL (Kelley, Moriarty, Wright, B. Jonkman) and ECEE (Pao). Figure 1: The seed project has been a multidisciplinary research project that uses atmospheric science and engineering methods to develop potentially transformational improvements in wind turbine performance. NREL has made considerable progress characterizing hazardous atmospheric conditions, generating realistic very high resolution numerical wind fields, calculating accurate wind turbine response based on these numerical wind fields, and implementing new control algorithms to mitigate the effects of organized turbulence, wind shear, and high winds. A key result of NREL's research has been the identification of the adverse effects of coherent turbulent structures on wind turbine operations, especially during conditions such as the nocturnal jet. Since these coherent structures are random and intermittent, it is difficult to identify their statistical and spatial structures with data from a single tower. Recent developments in the remote sensing of atmospheric turbulence by pulsed Doppler lidar will be adopted to produce the most reliable estimates of the critical profiles of atmospheric winds and spatial turbulence. The lidar approach is potentially superior to traditional tower or sodar measurements since it is derived from a three dimensional spatial average upstream of a wind farm and therefore produces better statistical accuracy with sufficient lead time. CIRES has investigated two critical components of the system: the optimal design of the lidar and advanced processing algorithms to extract the required turbulence profiles. NREL (Kelley, B. Jonkman) has produced simulated wind fields consistent with actual atmospheric conditions and calculated (Moriarty, B. Jonkman) the wind turbine response to evaluate optimal control algorithms. The control algorithms have been developed by Lucy Pao of ECEE in collaboration with Alan Wright of NREL. Active collaboration has helped us to begin to determine the most useful statistical descriptions of the turbulence that can be extracted in real-time from the various Doppler lidar scanning patterns and processing algorithms. Optimal performance of the control algorithms has been determined by maximizing the efficiency of power generation and minimizing the down time from damage resulting from high levels of vibration and loading based on the output of the advanced wind and turbine simulator at NREL. This is requiring advanced control theory to incorporate the random forcing by the turbulent structures. A roundtable of researchers interested in advancing wind energy formed since the first 2006 CU/NREL Energy Initiative Symposium, and we have continued to interact symbiotically with several additional (unfunded) collaborators. These interactions have further increased with the development of the Center for Research and Education in Wind (CREW), in which all of the investigators, collaborators, and consultants on this project have been active. A unique aspect of our research results is the beginnings of a roadmap to operational implementation at wind farms. The results from this seed effort is now being actively expanded and investigated in a follow-on project directly funded by NREL, and NREL is interested in pursuing implementation and demonstration of our results on their Controls Advanced Research Turbine (CART), where we will use real-time data from a Doppler lidar (provided by a company, such as Coherent Technologies, Inc. of Louisville) and the most promising lidar processing and wind turbine control algorithms that we have demonstrated under this seed project. If the experimental results on NREL’s CART are successful, we will seek an opportunity to demonstrate the novel algorithms developed at an operational wind farm such as Lamar, Colorado. This will provide valuable information to wind farm operators for future improvements in cost-effective wind energy production. Since Coherent Technologies is a local company, improvements in the design of the lidar can be easily tested for effective use of resources. Results of Seed Project: Professors Frehlich, a physicist and atmospheric scientist with CIRES, and Pao, a control systems engineer with the Department of Electrical, Computer, and Energy Engineering, met for the first time at the CU/NREL Research Symposium in the Fall of 2006. This seed grant has provided them with a better appreciation of the state of the art and capabilities of each other’s fields. This type of multidisciplinary effort is expected to lead to controllers that will significantly improve wind turbines and wind farm performance. This project gave them a much clearer idea of the steps needed to achieve a fully integrated lidar-based controller. Analysis of the scanning Doppler lidar data and the tower data at the NREL site produced good agreement for all types of atmospheric conditions. In addition, the ability of the lidar to provide accurate upstream profiles using short averaging times has verified the value of the three-dimensional averaging that is only possible with a scanning system. The PhD student, Jason Laks, has completed several simulations of synthetic data and performance of various turbine controllers. Initial results show the promise of combined feedforward/feedback control of wind turbines. The project was able to leverage the following funding: For Frehlich: From NREL: $22K ($7K for lidar data and $15K of salary) From NSF: $30K ($10K for lidar data and $20 K for salary) From ARO: $15K (mainly salary) For Pao: From NREL: $114,363 project (January 2009 to December 2009) on “Wind Turbine Advanced Feed-forward Controls” to continue to investigate combined feedforward and feedback control of wind turbines. A new PhD student (Fiona Dunne) was hired as a result of this contract. Publications and presentations: J. H. Laks, L. Y. Pao, and A. Wright, 2009, “Combined Feedforward/Feedback Control of Wind Turbines to Reduce Blade Flap Bending Moments,” Proc. AIAA/ASME Wind Energy Symposium, Orlando, FL. J. Laks, L. Y. Pao, and A. Wright, 2009, “Control of Wind Turbines: Past, Present, and Future,” Proc. American Control Conf., St. Louis, MO, pp. 2096-2103. L. Y. Pao and K. E. Johnson, 2009, “A Tutorial on the Dynamics and Control of Wind Turbines and Wind Farms,” Proc. American Control Conf., St. Louis, MO, pp. 2076- 2089. F. Dunne, L. Y. Pao, A. D. Wright, B. Jonkman, and N. Kelley, “Combining Standard Feedback Controllers with Feedforward Blade Pitch Control for Load Mitigation in Wind Turbines,” submitted in June 2009 for publication in the Proc. AIAA/ASME Wind Energy Symposium, Orlando, FL, Jan. 2010. J. Laks, L. Y. Pao, and A. D. Wright, “Blade Pitch Control with Preview Wind Measurements,” submitted in June 2009 for publication in the Proc. AIAA/ASME Wind Energy Symposium, Orlando, FL, Jan. 2010. R. Frehlich and N. Kelley, 2008, “Measurements of Wind and Turbulence Profiles with Scanning Doppler Lidar for Wind Energy Applications,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (J-STARS), 1, 42-47. R. Frehlich and N. Kelley, 2008, “Scanning Doppler Lidar Measurements for Wind Energy Applications,” AGU Meeting, Paper A21E-0225. R. Frehlich and N. Kelley, 2009, “Coherent Doppler Lidar for Wind Energy Research,” 15th Coherent Laser Radar Conference, Toulouse, France. Invited Scientific Presentations: Pao, L., 2007: Combined Feedforward/Feedback Control of Flexible Structures, University of Illinois at Urbana-Champaign Pao, L., 2008: Control of Flexible Structures: From Atomic Force Microscopes to Megawatt Wind Turbines, Iowa State University Pao, L., 2008: Control of Flexible Structures, Stanford University Pao, L., 2008: Combined Feedforward and Feedback Control of Flexible Structures: From Atomic Force Microscopes to Megawatt Wind Turbines, Department of Mechanical Engineering, University of California at Berkeley Pao, L., 2008: Interesting Connections in the Control of Systems Across Multiple Scales: From Megawatt Wind Turbines down to Atomic Force Microscopes, Miller Institute for Basic Research in Science, University of California at Berkeley Frehlich, R., 2008: Measurements of Turbulence Profiles with Scanning Doppler Lidar for Wind Energy Applications. Invited talk. Geophysical Turbulence Program Workshop, Observing the Turbulent Atmosphere: Sampling Strategies, Technology, and Applications, National Center for Atmospheric Research, Boulder, CO Frehlich, R., 2008: Doppler Lidar Measurements of Winds and Turbulence in the Boundary Layer. Invited talk, 14th International Symposium for the Advancement of Boundary Layer Remote Sensing, Riso National Laboratory, Denmark Frehlich, R., 2008: Fundamentals on Turbulence Observations using LIDAR. Invited lecture. PhD Summer School on Remote Sensing in Wind Energy, Riso National Laboratory, Denmark Frehlich, R., 2008: Three Dimensional Measurements of Boundary Layer Statistics using Scanning Doppler Lidar, AGU Meeting, paper A44A-01. Future work and funding: A major wind turbine manufacturer is planning to fund follow-on work to this seed grant project. A sponsored research project on “Lidar- Based Control of Wind Turbines: A Feasibility Study” was approved for funding in October 2008. The Project Agreement was negotiated during October 2008 to July 2009, and the project is expected to officially launch this month (July 2009). The study will look at a direct approach (disturbance feedforward), and an indirect approach (controller selection). In addition, Frehlich and Pao are investigators on: - Two ARPA-E concept papers that were recently submitted in June 2009 - An NSF Engineering Research Center pre-proposal being submitted by July 15, 2009, and - A DOE University / Industry Wind Energy Consortia proposal being submitted by July 29, 2009. Quote: “I met Rod and some of the NREL collaborators for the first time at the 2006 CU/NREL research symposium. This project has developed into a growing number of interesting research avenues for all of us.” Lucy Pao, co-investigator.
Pages to are hidden for
"Title Design of Lidar, Data Processing, and Control Algorithms"Please download to view full document