Output Power Correlation Between Nearby Wind Power Plants

The National Renewable Energy Laboratory (NREL) started a project in 2000 to record long-term, high-frequency (1-Hz) wind power output data from large commercial wind power plants. Outputs from about 330 MW of wind generating capacity from wind power plants in Buffalo Ridge, Minnesota, and Storm Lake, Iowa, are being recorded. Analysis of the collected data shows that although very short-term wind power fluctuations are stochastic, the persistent nature of wind and the large number of turbines in a wind power plant tend to limit the magnitudes and rates of changes in the levels of wind power. Analyses of power data confirm that spatial separation greatly reduces variations in the combined wind power output relative to output from a single wind power plant. Data show that high frequency variations of wind power from two wind power plants 200 km apart are independent of each other, but low frequency power changes can be highly correlated. This fact suggests that time-synchronized power data and meteorological data can aid in the development of statistical models for wind power forecasting.

Optimization of Wind Turbines Using Helicoidal Vortex Model

The problem of the design of a wind turbine for maximum output is addressed from an aerodynamical point of view. It is shown that the optimum inviscid design, based on the Goldstein model, satifies the minimum energy condition of Betz only for light loading. The more general equation governing the optimum is derived and an integral relation is obtained, stating that the optimum solution satisfies the minimum energy condition of Betz in the Trefftz plane “in the average”. The discretization of the problem is detailed, including the viscous correction based on the 2-D viscous profile data. A constraint is added to account for the force on the tower. The minimization problem is solved very efficiently by relaxation. Several optimized solutions are calculated and compared with the NREL rotor, using the same profile, but different chord and twist distributions. In all cases, the optimization produces a more efficient design.

A Straight-Bladed Variable-Pitch VAWT Concept for Improved Power Generation

The paper presents three modifications for an improved performance in terms of increased power output of a straight-bladed VAWT by varying its pitch. Modification I examines the performance of a VAWT when the local angle of attack is kept just below the stall value throughout its rotation cycle. Although this modification results in a very significant increase in the power output for higher wind speeds, it requires abrupt changes in the local angle of attack making it physically and mechanically impossible to realize. Modification II improves upon the first by replacing the local angle of attack by the blade static-stall angle only when the former exceeds the latter. This step eliminates the two jumps in the local effective angle of attack curve but at the cost of a slight decrease in the power output. Moreover, it requires a discontinuous angle of attack correction function which may still be practically difficult to implement and also result in an early fatigue. Modification III overcomes the limitation of the second by ensuring a continuous variation in the local angle of attack correction during the rotation cycle through the use of a sinusoidal function. Although the power output obtained by using this modification is less than the two preceding ones, it has the inherent advantage of being practically feasible.

Navier-Stokes and Comprehensive Analysis Performance Predictions of the NREL Phase VI Experiment

A vortex lattice code, CAMRAD II, and a Reynolds-Averaged Navier-Stoke code, OVERFLOW-D2, were used to predict the aerodynamic performance of a two-bladed horizontal axis wind turbine. All computations were compared with experimental data that was collected at the NASA Ames Research Center 80-by 120-Foot Wind Tunnel. Computations were performed for both axial as well as yawed operating conditions. Various stall delay models and dynamics stall models were used by the CAMRAD II code. Comparisons between the experimental data and computed aerodynamic loads show that the OVERFLOW-D2 code can accurately predict the power and spanwise loading of a wind turbine rotor.

Analysis of Blunt Trailing Edge Airfoils

The adoption of blunt trailing edge airfoils for the inner regions of large wind turbine blades has been proposed. Blunt trailing edge airfoils would not only provide increased structural volume, but have also been found to improve the lift characteristics of airfoils and therefore allow for section shapes with a greater maximum thickness. Limited experimental data makes it difficult for wind turbine designers to consider and conduct tradeoff studies using these section shapes. This lack of experimental data precipitated the present analysis of blunt trailing edge airfoils using computational fluid dynamics. Several computational techniques are applied including a viscous/inviscid interaction method and several Reynolds-averaged Navier-Stokes methods.

Wind Shear and Turbulence Effects on Rotor Fatigue and Loads Control

The effects of wind shear and turbulence on rotor fatigue and loads control are explored for a large horizontal axis wind turbine in variable speed operation from 4 to 20 m/s. Two and three blade rigid rotors are considered over a range of wind shear exponents up to 1.25 and a range of turbulence intensities up to 17%. RMS blade root flatwise moments are predicted to be very substantially increased at higher wind shear, and resultant fatigue damage is increased by many orders of magnitude. Smaller but similar trends occur with increasing turbulence levels. In-plane fatigue damage is driven by 1P gravity loads and exacerbated by turbulence level at higher wind speeds. This damage is higher by one to two orders of magnitude at the roots of the three blade rotor. Individual blade pitch control of fluctuating flatwise moments markedly reduces flatwise fatigue damage due to this source, and to a lesser degree the in-plane damage due to turbulence. The same is true of fluctuating rotor torque moments driven by turbulence and transmitted to the drive train. Blade root moments out of the plane of rotation aggregate to create rotor pitching and yawing moments transmitted to the turbine structure through the drive train to the yaw drive system and the tower. These moments are predicted to be relatively insensitive to turbulence level and essentially proportional to the wind shear exponent for the two blade rotor. Fluctuating moments are substantially reduced with individual blade pitch control, and addition of a teeter degree of freedom should further contribute to this end. Fluctuating pitching and yawing moments of the three blade rotor are substantially less sensitive to wind shear, more sensitive to turbulence level, and substantially lower than those for the two blade rotor. Mean rotor torque and hence power are essentially the same for both rotors, independent of wind shear, and somewhat reduced with individual blade pitch control of fluctuating flatwise moments. The same is true of mean rotor thrust, however fluctuations in rotor thrust are substantially reduced with individual blade pitch control. It appears, on balance, that higher wind shear coupled with turbulence effects should be accounted for in the fatigue design of large, long life turbines. Much more work is required on this problem.

Dynamic Simulation of a Wind Farm With Variable Speed Wind Turbines

Wind power generation has increased very rapidly in the past few years. The total U.S. wind power capacity by the end of 2001 was 4,260 megawatts. As wind power capacity increases, it becomes increasingly important to study the impact of wind farm output on the surrounding power networks. In this paper, we attempt to simulate a wind farm by including the properties of the wind turbine, the wind speed time series, the characteristics of surrounding power network, and reactive power compensation. Mechanical stress and fatigue load of the wind turbine components are beyond the scope this paper. The paper emphasizes the impact of the wind farms on the electrical side of the power network. A typical wind farm with variable speed wind turbines connected to an existing power grid is investigated. Different control strategies for feeding wind energy into the power network are investigated, and the advantages and disadvantages are presented.

Probabilistic Analysis of List Data for the Estimation of Extreme Design Loads for Wind Turbine Components

Robust estimation of wind turbine design loads for service lifetimes of 30 to 50 years that are based on field measurements of a few days is a challenging problem. Estimating the long-term load distribution involves the integration of conditional distributions of extreme loads over the mean wind speed and turbulence intensity distributions. However, the accuracy of the statistical extrapolation is fairly sensitive to both model and sampling errors. Using measured inflow and structural data from the LIST program, this paper presents a comparative assessment of extreme loads using three distributions: namely, the Gumbel, Weibull and Generalized Extreme Value distributions. The paper uses L-moments, in place of traditional product moments, to reduce the sampling error. The paper discusses the application of extreme value theory and highlights its practical limitations. The proposed technique has the potential of improving estimates of the design loads for wind turbines.

A Parallel h-Adaptive Finite Element Model for Wind Energy Assessment in Nevada

A Petrov-Galerkin finite element model that employs local mesh adaptation is being developed to determine potential wind energy sites within the state of Nevada. Meteorological data collected from various private, county, city, and government agencies are used to generate diagnostic flow fields, which subsequently provide initial conditions for the prognostic solution of the time-dependent equations of motion and species transport. The model runs on a multiprocessor SGI Onyx 3800. Results of the data collection, including wind energy site forecasts, will be made available on the web when the assessment for the entire state is completed.

The Edge Singularity of an Actuator Disc With a Constant Normal Load

All rotor and propeller design methods using momentum theory are based on the concept of the actuator disc, formulated by Froude. In this concept, the rotor load is represented by a uniform pressure jump. This pressure jump generates infinite pressure gradients at the edge of the disc, leading to a velocity singularity. The subject of this paper is the characterization of this velocity singularity assuming inviscid flow. The edge singularity is also the singular leading edge of the vortex sheet emanating from the edge. The singularity is determined as a simple bound vortex of order O(1), carrying an edge force Fedge = −ρ Vedge × Γ. The order of Fedge equals the order of Vedge. This order is determined by a radial momentum analysis. The classical momentum theory for actuators with a constant, normal load Δp appears to be inconsistent: the axial balance provides a value for the velocity at the actuator, with which the radial balance cannot be satisfied. The only way to obtain consistency is to allow the radial component of Fedge to enter the radial balance. The analysis does not resolve on the axial component of Fedge. A quantitative analysis by a full flow field calculation has to assess the value of Fedge for the various actuator disc flow states. Two other solutions for the edge singularity have been published. It is shown that both solutions do not comply with the governing boundary conditions.