Utilizing Radar Measured Velocity

Fields to Forecast Turbine Wind Speeds

James B. Duncan Jr.1, Dr. Brian D. Hirth1, Dr. John L. Schroeder2

1National Wind Institute, Texas Tech University, Lubbock, TX, USA2Department of Geosciences, Texas Tech University, Lubbock, TX, USA

NAWEA 2015 SymposiumBlacksburg, Virginia, USA, June 10th, 2015

Introduction and Motivation

Wind energy today….• Multi-billion dollar industry with over 60+ MW (through

2014) deployed.

Ultimate goal of decreasing the cost of energy and mitigating risk.

AWEA/DOE/BNEF

Wind Turbine Control Systems and Remote SensingConventional wind turbine feedback controllers are entirely reactive rather than proactive.

Transients such as gusts, varying shears, and directional changes in the inflow all represent UNKNOWN disturbances.

Recent developments in remote sensing have led to much interest in the possibility of improving wind turbine controllers by providing preview information of the approaching wind field.

•Nacelle-mounted Look-Ahead Systems

– Benefits….

– Limitations…

?

Aim of Research

Although information of the near-upstream flow conditions has been examined through the use of nacelle mounted LIDAR

systems, employing three dimensional wind field maps derived from scanning instruments to provide an extended wind speed

forecast for individual turbine locations has never been investigated.

Radar Basics and Experimental Setup

Texas Tech University maintains two-research grade mobile Ka-band Doppler radar systems. A single radar system is limited to

deriving the radial component. A dual-Doppler scanning strategy allows

for the extraction of the full horizontal velocity vector and construction of horizontal wind field maps.

Radar Specifics

Half-Power Beamwidth

0.49°

Along-Beam Range Resolution

15m

Revisit Time ~45 sec

Dual-Doppler Horizontal Wind Speed Synthesis (~45 sec Revisit Time)

Preliminary Methods

In order to provide preview measurements of the upstream flow, the following steps must be taken:

1. Derive an upstream projection representative of the inflow.

• Look angle of the upstream projection (Rotor Sweep Area Average of WD located 1D Upstream).

• Spatial dimensions of the upstream projection (Point Measurement Along Path).

2. Forecast wind speeds measured upstream to the location of turbine.

• Analyze wind speeds at hub height within upstream projection at distances between 1-6 D upstream at intervals of ~10 m.

• Assuming the wind speed magnitude of the features remains constant with time, derive time offset for future feature arrival.

3. Generate a future-forecasted time series.

ABL Streaks and Advection of MomentumOrganized coherent structures are embedded within the turbulent flow-fields of the ABL.

Near-surface streaks, defined as elongated areas of enhanced/reduced wind speeds (Traumner et al. 2015).

Shown using full-scale radar measurements to be skewed left of the wind direction (Lorsolo et al. 2008; Marathe 2014).

Variation in Advection Between Actual and Projected Motion

Step 1 Step 2 Step 3 Step 4 Step 5 Avg

Feature 1 4.44° 8.09° 5.7° 8.3° 6.2° 6.55°

Feature 2 2.8° 3.9° 5.6° 5.5° 2.6° 4.07°

Variation in Upstream Projection Paths

Propagation of upstream conditions based upon the governing wind direction is not sufficient to provide an extended preview of the approaching wind field.

Variation between wind direction and actual momentum advection leads to inaccuracies in the future-forecasted time series.

1.03 ms-1

(49.5 sec) 0.82 ms-1

(19.5 sec)

0.77 ms-1

(34.5 sec)

Adapted Methods

Upstream projection offset counterclockwise of the governing wind direction by four degrees.

Adjust Upstream Path to Account for Momentum Advection

CC Degree Offset

RMSE Value

3-sec Avg. 5-sec Avg.

0 0.7471 0.7227

1 0.7210 0.6952

2 0.6978 0.6724

3 0.6803 0.6556

4 0.6740 0.6468

5 0.6768 0.6469

6 0.6849 0.6588

0.073 ms-1 & 0.076 ms-1 reduction in RMSE for 3 and 5 second averages respectively

Radar Derived Forecasted Wind Speeds

Radar Derived Forecasted Wind Speeds

Forecasts Across Observational Period (52 min 12 sec)

Averaging Time RMSE Value (m/s)

1-sec 0.71

3-sec 0.67

5-sec 0.65

10-sec 0.59

Concluding Remarks

Preliminary results demonstrate the ability to predict future conditions at the turbine leveraging radar-derived dual-Doppler horizontal wind field maps. Scanning instruments allow for the ability to accurately forecast

wind speeds 45+ seconds or more in advance from the turbine.

• Elevating the instruments to hub height could reduce time required to collect hub height information by an order of magnitude further enhancing this method.

Coordinated deployment of multiple specialized Doppler radar systems can provide the foundational information to construct proactively responding turbine control systems.

Next Steps…Plant Scale

Acknowledgements:The use of radar technology to document wind plant complex flows occurred with support from the United States Department of Energy Congressionally Directed Project: Great Plains Wind Power Test Facility (DE-FG-06-GO86092). This specific research is funded through support

from the National Science Foundation: Building the Foundation for Smart Wind Farms through First-Order Controls Opportunities based on Real-Time Observations of Complex Flows

(CBET-1336935).

ReferencesBossanyi, E. A., A. Kumar, and O. Hugues-Salas, 2014: Wind turbine control applications of

turbine-mounted LIDAR. J. Phys.: Conf. Ser., 555, 012011. Hirth, B. D. and J. L. Schroeder, 2013: Documenting wind speed and power deficits behind a

utility-scale wind turbine. J. Appl. Meteor. Climatol., 52, 39-46. Hirth, B. D., J. L. Schroeder, W. S. Gunter, and J. G. Guynes, 2015: Coupling Doppler radar-

derived wind maps with operational turbine data to document wind farm complex flows. Wind Energy, 18, 529-540.

Lorsolo, S., J. L. Schroeder, P. Dodge, and F. Marks, 2008: An observational study of hurricane boundary layer small-scale coherent structures. Mon. Wea. Rev., 136, 2871-2893.

Marathe, N. 2014: Investigation of power performances and wakes of wind turbines under yawed flow. Dissertation, Texas Tech University, 141 pp.

Mikkelsen, T., N. and Couauthors, 2013: A spinner-integrated wind lidar for enhanced wind turbine control. Wind Energy, 16, 625-643.

Schlipf, D., J. Anger, S. Kapp, O. Bischoff, M. Hofsaß, A. Rettenmeier, and M. Kuhn, 2011: Prospects of optimization of energy production by lidar assisted control of wind turbines. Proc. European Wind Energy Conference and Exhibition, EWEA, Brussels, 14-17 2011.

Traumner, K., T. Damian, C. Stawiarski, and A. Wieser, 2015: Turbulent structures and coherence in the atmospheric surface layer. Bound.-Layer Meteor., 154, 1-25.

Atmosphere to Electrons (A2e) Initiative Overview, DOE Wind and Water Power Technologies Office, 2014.