Page 30 - ASME DSCC 2015 Program

Technical Program

State estimation for the homogeneous 1-d euler equation by automatic Identification of closed-Loop Wind Turbine dynamics via
unscented Kalman filtering Genetic Programming
Contributed regular paper. DSCC2015-9603 Invited session paper. DSCC2015-9768
Sascha Wolff, rudibert king, Technische Universität Berlin, Berlin, William G. la Cava, Kourosh Danai, Matthew A. lackner, University
Germany, Phillip Berndt, Freie Universität Berlin, Berlin, Germany, of Massachusetts Amherst, Amherst, MA, United States, lee Spector,
Jan-Simon Schäpel, Berlin Institute of Technology, Berlin, Germany Hampshire College, Amherst, MA, United States, Paul fleming, Alan
Wright, National Renewable Energy Laboratory, Golden, CO, United States
In this article we present a dynamic state estimation method for a one
dimensional flow field which is described by the homogeneous Euler Wind turbines are nonlinear systems that operate in turbulent environments.
equation. The estimated quantities include the pressure, velocity, density As such, their behavior is difficult to characterize accurately across a wide
and temperature field, which are of interest, for instance, for some com- range of operating conditions by physically meaningful models. Customarily,
bustion concepts. The algorithm relies only on a small number of discrete data-based models of wind turbines are defined in `black box’ format, lack-
pressure measurements from the flow field. The influence of the number of ing in both conciseness and physical intelligibility. To address this deficiency,
used pressure measurements on the convergence speed of the algorithm is we identify models of a modern horizontal axis wind turbine in symbolic
investigated. For the state estimation, an Unscented Kalman Filter scheme form using a recently developed symbolic regression method. The method
is exploited. The proposed method is applied in numerical simulations to used relies on evolutionary multi-objective optimization to produce succinct
demonstrate its effectiveness. dynamic models from operational data without `a priori’ knowledge of the
system. We compare the produced models with models derived by other
InVITED SESSIon methods for their estimation capacity and evaluate the trade-off between
2-1-1 WM4 Dynamics and Control of Wind Energy Systems model intelligibility and accuracy. Several succinct models are found that
Geroge Bellows f 1:30pm–3:30pm
predict wind turbine behavior as well as or better than more complex
alternatives derived by other methods.
Session Organizer: Chris Vermillion, University of North Carolina
A Hardware-in-the-loop Strategy for Control of a Wind Turbine Test
Session Organizer: azad ghaffari, University of Michigan
Bench
Session Organizer: Y. li, University of Texas, Dallas
Invited session paper. DSCC2015-9930
Session Chair: Chris Vermillion, University of North Carolina
Session Co-Chair: Ryan Schkoda, Clemson University Ryan Schkoda, Amin Bibo, Clemson University, North Charleston, SC,
United States
Maximizing Average Power output of an Airborne Wind Energy
This paper presents a strategy for operating and simulating a full scale wind
Generator under Parametric uncertainties
turbine test bench in order to emulate the field conditions experienced by the
Invited session paper. DSCC2015-9764
mechanical portion of a nacelle’s drivetrain. The presented strategy makes
Michelle Kehs, Hosam K. fathy, Pennsylvania State University, University use of a real-time simulation of a wind turbine model to provide reference
Park, PA, United States, Chris Vermillion, University of North Carolina at signals for the test bench. These reference signals include the thrust force,
Charlotte, Charlotte, NC, United States lateral force, vertical force, pitch moment, yaw moment, and drivetrain speed
This paper presents a controller for maximizing the time-averaged power at the nacelle’s control point or hub point: a point along a nacelle’s drivetrain
output from an airborne wind energy generator in uncertain wind conditions. representing what would be the center of the nacelle hub. The effectiveness
This system’s optimal energy output often involves flying in periodic figure-8 of this approach is studied with the help of a multibody, dynamic model of the
trajectories, but the precise optimal figure-8 shape is sensitive to environ- 7.5~MW test bench located at Clemson University’s Wind Turbine Drivetrain
mental conditions, including wind speed. The literature presents controllers Testing Facility. Simulation results show that the proposed control strategy is
that are able to adapt to uncertainties, and this work expands on the current feasible for emulating field conditions on a wind turbine test bench.
literature by using an extremum seeking based method. Extremum seeking Torsional Vibrations in the Drivetrain of DfIG- and PMG-Based Wind
is particularly well-suited for this application because of its well under- Turbines: Comparison and Mitigation
stood stability properties. In this work, extremum seeking is used to search
Invited session paper. DSCC2015-9953
through a family of optimal trajectories (computed offline) that correspond
fariba fateh, Warren White, Don Gruenbacher, Kansas State University,
to discrete wind speeds. The controller is efficient in that it only searches
Manhattan, KS, United States
for the optimum trajectory over the uncertain parameter (in this paper, wind
speed). Results show that the controller converges to the optimal trajectory, In this paper, torsional vibrations in a five-mass drivetrain of a doubly fed
provided it is initialized to a stable figure-8. The speed of convergence is induction generator (DFIG)-based, as well as a permanent magnet gener-
dependent on the difference between the initial average power output and ator (PMG)-based wind turbine are investigated through simulations. The
the optimal average power output. simulated model includes aerodynamics of a 750kW wind turbine, as well
as dynamics of the generator, gearbox, and back-to-back power converters.
In this study, the effectiveness of a sliding mode based control scheme to
damp the drivetrain torsional vibrations for the events of a voltage dip occur-
ring on the power grid and a wind speed variation is presented. The simu-
lation results demonstrate mechanical disturbances have similar impacts on
30 the drivetrain of DFIG-based and PMG-based wind turbines. However, the
back-to-back converters in a PMG-based wind turbine effectively isolate the
effects of power oscillations on the drivetrain.