Harvesting wind energy from sites with lower annual average wind speeds at costs comparable to what can now be achieved in higher wind sites is an international challenge. It requires pushing the technology to optimize the use of materials and machinery and tuning the structure to withstand the loads unique to the lower wind speed sites. The U.S. Department of Energy’s laboratories—The National Renewable Energy Laboratory (NREL) and Sandia National Laboratories—are engaged in both basic science and applied research to overcome current technology limitations and enhance the individual components necessary to achieve cost-effective low wind speed turbines.
Technology improvements are necessary in three principal areas to minimize the cost of electricity production at lower wind speed sites.
1 Turbine rotor diameters need to be larger, harvesting the lower energy winds from a larger inflow area. This must be accomplished without increasing the cost of the rotor beyond what a smaller rotor might cost in an energetic site.
2 Towers must be taller to take advantage of the increasing wind speed at greater heights. Again, the approach taken for these towers needs to be less costly than traditional approaches if the benefit from increased energy capture is to exceed increased tower costs.
3 Generation equipment and power electronics must be more efficient to accommodate sustained light wind operation at lower power levels without increasing electrical system costs.
Achieving equivalent or lower COE production for low wind speed sites is difficult. Increased energy capture cannot come at the expense of increased machine cost and fundamental technology advances are required if the cost goals are to be achieved. For example, stretching the rotor to improve capture area, a length squared effect, will have its benefits swamped by increased material usage, a length cubed effect, unless some changes in design approach or materials are made. The following discussion explains the research issues created by the need to meet the above goals.
As both the rotor diameter and turbine hub height grow to extract more energy from the available wind resource, a better understanding of the inflow is necessary. Until now, most commercial turbines have operated at hub heights below 50 m in the sub-viscous region of the Planetary Boundary Layer (PBL). Here, the near surface heating effects produce a well-mixed and relatively homogeneous inflow boundary layer. As the hub height increases, the characteristics of the PBL change significantly as both surface heating and friction effects are reduced. At 100+ m, larger, more coherent flow phenomena, such as Kelvin-Helmholtz waves and the nocturnal jet, are formed. In addition to the more homogeneous small-scale turbulence experienced at lower-hub heights, large-scale coherent flow structures with characteristic length scales on the order of the turbine diameter will also be encountered. Thus, the potential for single-event turbine/coherent inflow structure interactions must be considered in addition to the more typical stochastic inflow events currently modeled.
A new understanding of this flow environment is being gleaned from both field measurements and numerical simulation. From these results, new inflow models are being developed to assess the performance of turbine designs incorporating larger diameters and taller towers. In the near term, field data collected on tall towers is being used to identify coherent inflow structures that produce the largest single event load histories. These structures are simulated analytically and used as dynamic inflow inputs for current aerodynamic and structural response models.
In a parallel research effort, computational models based on Reynolds Averaged Navier-Stokes (RANS) and Large Eddy Simulation (LES) methods are being used to model the development of large coherent vortical flows under conditions of neutral inflow boundary layer stability. These results provide additional inflow simulation input and will help establish similarity between computationally derived and measured inflow data. In addition, these models should provide some insight as to the frequency and extent of the coherent vortex/turbine rotor interactions that can be expected. Finally, more detailed inflow models based on better physical models must be developed. In the longer term, computational fluid dynamic (CFD) models will provide detailed, high fidelity inflow data based upon the underlying flow physics and local topography.
Rotor Loads and Controls
The new Low Wind Speed Turbines (LWST) will have larger diameters, taller towers, and proportionally lower tower top mass if cost targets are to be achieved. Thus, these new designs will be much more dynamically active and will need to respond to a more complex inflow environment with load mitigation strategies not currently employed. Significant improvements in the ability of the machine to sense the loads and reduce them at the rotor, before they begin to feed through the system, are critical.
There are several efforts currently in progress aimed at achieving this end. Control systems, including sensors, actuators, and software to connect the two, are an important part of the research effort. In the past, control systems have been collective in nature (operating on all the blades in unison) and have sought primarily to affect smooth power generation and maintain the desired rotor speed. Such systems are already seeing limited use to mitigate tower loads by controlling the rotor thrust and eliminating some large-scale turbulence loads. Future control systems will be able to sense the loads (or winds) locally and respond with each blade individually. Pitch control is the current option of choice of many turbine manufacturers, but activation devices that use local flow control to influence and mitigate turbulence driven loads may also be built into the blade. For example, twist-coupled blades that naturally change their angle of attack and reduce the load peaks in response to gusts are being investigated as a passive way to accomplish at least a portion of the desired load reduction.
As the complexity of the controller increases, there is a further need to improve the reliability of sensors, actuators, software, and all the moveable parts of the rotor control system. Control optimization and advanced multi-input, multi-output designs will be required. Substantial research is being devoted both to control development and to the ability to model the complete aeroelastic system in order to facilitate the transfer of advanced controls concepts from the laboratory to the commercial world. All of these control efforts are working toward the goal of enabling designers to stretch the rotors while holding down the loads that would increase structure costs.
Active turbine control will only be achieved if the aerodynamic loads from inflow/rotor interactions are well understood and properly modeled. “On average” knowledge and prediction of rotor loads is insufficient for a rapid response LWST control system that must literally “fly” a dynamically active soft turbine system in a complex inflow. The NREL full-scale wind tunnel test at the NASA Ames 80 × 120-ft. wind tunnel indicated that current design and prediction tools based on blade element momentum (BEM) theory perform well in conditions for which they have been tuned. They do not perform in novel and unusual environments like those anticipated for the LWST.
Aeroelastic and CFD tools validated with field and wind tunnel data offer the best approach for understanding the underlying flow structure and rotor loads. However, typical CFD models contain millions of nodes and take days of computation time using typical PC and workstation resources available to the wind industry. This time constraint neither meets the near real-time loads prediction needed for active control nor is attractive as a design tool when numerous design iterations are required.
Current research efforts to improve aerodynamic codes are split between furthering our basic understanding of the rotating, 3-D, unsteady separated flows experienced by turbines operating in the field and incorporating this knowledge into advanced predictive models for loads. A concerted effort is underway to enhance existing predictive codes with empirically derived modules to maintain computational speed and increase accuracy until we are able to bring CFD tools into the realm of everyday design usage.
Materials and Manufacturing
Even if improved controls are able to lower the dynamic blade loading, increased material stiffness will be necessary to prevent blade tower strikes and lower overall rotor costs. There has been great interest in the use of carbon fiber materials in blades to achieve these results; however, significant uncertainties remain with respect to how best to incorporate higher cost fibers into the blade design. Because the traditionally expensive carbon fibers drop in cost as the tow size (numbers of fibers in each bundle) increases, it is important to determine if these cheaper but stiffer fibers have a role to play in wind turbine blades. Material usage will also remain inefficient if the uncertainties in strength remain high, thereby mandating larger design margins. Extensive testing programs are required to establish strength and fatigue properties of these new materials, given the unique stochastic loading environments in which they must reliably perform.
With the development of much larger rotors and the associated incorporation of new materials, the manufacturing of blades becomes significantly more complicated. The wind industry is moving from traditional hand-lay-up fabrication methods plagued with high labor and inconsistent quality to resin infusion methods (VARTM, RTM) and pre-preg fabrication techniques. As carbon and carbon-hybrid materials are incorporated and unique laminate structures are proposed (e.g., off-axis for twist-coupled blades), new questions arise as to the fabrication efficiency and acceptance of this new technology by industry.
Many combinations of fiberglass materials, including braided, woven, stitched, and pre-preg constructs, are likely candidates for use in composite blade structures. Different forms of the same materials can have significantly different fatigue, strength, and handling properties. Trade-offs between materials, structure, form, manufacturing process, and design requirements must be made to obtain large, stiff, relatively light, yet economic blade designs. Research with industry participation will be required to validate these technologies in order to reduce risk, lower costs, and increase commercial viability.
Towers and Logistics
The connection between tower size and logistics is undeniable. With traditional tapered tower design, the taller the tower, the larger the base diameter becomes and the more troublesome manufacturing and transporting the tower become. Attention has, therefore, begun to shift away from the traditional concept of a tapered steel tower that is transported to the site as whole sections and bolted together to form the final structure. Some attention has been given to alternate tower materials with reinforced concrete because of its low cost. Guy-cables are getting a new look as a means of spreading the load at the tower base. Multiple sections welded in place on the site may be required to eliminate the need to move the enormous diameters of multimegawatt machines. Various other on site manufacturing innovations are also being investigated. Manufacturing towers on site provides an opportunity to merge the manufacturing and erection functions into a single operation, eliminating the need for the massive cranes that would otherwise be required to lift the heavy sections to great heights.
Generators and Power Electronics
The use of variable speed (VS) technology has significantly improved both energy capture and power quality; however, there are still significant advances that can be achieved. Much of the increased energy capture from the improved aerodynamics of VS operation is dissipated in the associated power conversion losses. Power electronics are most efficient at their maximum power rating, dropping off rapidly at lower power levels. Wind turbines often operate below 50% of their maximum power rating where the power conversion efficiencies are far from ideal. This operating characteristic represents a significant opportunity for the development of unique conversion architectures that maximize energy capture over the operating wind spectrum.
Tower-top weight constraints represent another significant opportunity. Optimum combinations of generator, gearbox, and drive train designs are required to achieve maximum energy capture and performance. With larger turbines, direct drive, permanent magnet machines offer the promise of lower maintenance costs as a result of the elimination of the gearbox. In contrast, single-stage gearbox designs can significantly reduce the generator size and reduce the tower top weight. Multiple gearbox designs offer both component redundancy and component assembly/replacement, minimizing the time required to make repairs. No clear cost or performance advantage has been gleaned from any of the current design alternatives. Ultimately, it will be the vertical integration and the optimization of the entire turbine system that produces the lowest overall machine cost for LWST applications.
The cost of electricity produced by wind turbines is already competitive at locations with excellent wind resources. While it may be several years before the U.S. has exhausted these prime wind development sites, more marginal wind resources at locations where transmission access and load proximity increase electricity market value are even now under consideration. Research efforts at the DOE National Laboratories are focused on making the economics of wind-generated electricity at marginal sites comparable with the economics at the best wind sites available today. There is no silver bullet that will drop the cost of energy with a single change in approach or technology innovation. Efforts across the gamut of technical specialties are being integrated and optimized in order make wind power the least expensive and most accessible power generation on the grid.
This work was performed at the National Renewable Energy Laboratory in support of the U.S. Department of Energy under Contract No. DE-AC36-99GO10337. This material is a declared work of the U.S. Government and is not subject to copyright protection in the United States.
Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin company, for the U.S. Department of Energy. This work is supported by the U.S. Department of Energy under Contract DE-AC04-04AL85000 and DE-AC36-83CH10093.