Pioneering work by EOLOS researchers at the University of Minnesota (UMN) using snow during a Minnesota blizzard is giving researchers new insight into the airflow around large wind turbines. A winter storm and giant spotlights allowed the team to visualize wind currents around a full-scale turbine for the first time, a breakthrough that could hold important implications for wind farm design to maximize efficiency. The complete findings were published in Nature Communications in June of 2014. “Natural snowfall reveals large-scale flow structures in the wake of a 2.5-MW wind turbine” is available here, the press release from the University of Minnesota is available here, and a website principle investigator Dr. Jiarong Hong created to explain the project concepts in greater depth is available here. These findings were also covered in a number of mainstream news publications, including National Geographic and Live Science, and these stories can be found under News.   

University leaders hosted U.S. Deputy Secretary of Energy Daniel B. Poneman and Senator Al Franken on January 27 for a special visit to the University’s Eolos Wind Energy Research Field Station at UMore Park in Rosemount, Minnesota. This visit is among a handful of national stops by senior U.S. Department of Energy (DOE) administrators to highlight federal investments in clean energy research and technology development.

The University of Minnesota was one of three university consortia to receive a DOE wind energy research grant in 2009. The University’s $7.9 million award, funded through the American Recovery and Reinvestment Act, supports an academy-industry consortium focused on wind energy research and education to advance technology development, create economic opportunities and help to further the national goal to increase wind power to 20% by 2030.

A key outcome of the award is construction of the Wind Energy Research Field Station, featuring a U.S.-made, 2.5 megawatt Clipper Liberty wind turbine and a 426-foot meteorological research tower. Commissioned in October 2011, the facility is equipped with state-of-the-art instruments and sensors to measure weather conditions, wind speed and turbulence and the impact they have on the turbine structural reliability and wind energy capturing ability. Eolos researchers are currently focused on describiing and characterizing the turbine in its baseline, un-modified state.

The visit included a technical briefing by Fotis Sotiropoulos, director of the Eolos Wind Research Consortium, professor of civil engineering and director of the St. Anthony Falls Laboratory in the U’s College of Science and Engineering, on current and planned industry-academy research collaboration for development of novel technologies and enhancements that can improve the functionality and efficiency of wind turbines. In addition, Sotiropoulos described ongoing educational initiatives to produce the next generation of wind industry leaders.

Following a tour of the Eolos field station, Senator Franken and Deputy Secretary Poneman shared brief remarks on the visit and reflected on goals and opportunities for wind and other renewable energy.

Eolos Researchers in the University of Minnesota’s Department of Civil Engineering are developing a novel computational model for describing the structural behavior of wind turbine blades by applying behaviors typical of shell structures to the more straightforward beam theory. Wind turbine blades are thin-walled structures and, as such, are properly classified as shell structures. However, the ratio of their span length (approximately 45 meters) to chord length (1 to 3 meters) is more typical of a structural beam than a shell. In practice, both shell and beam models are used to analyze wind turbine blades.

With the goal of capturing ever more power from the wind, wind turbine manufacturers are extending blades lengths to 45 meters and longer.  However, with increased blade length comes with an increase in mass, blade flexibility, structural forces, and a ¬¬potential for decrease in operating lifetime due to fatigue damage.  In order to account for all these factors, turbine manufacturers will need to better understand how blades react to static and dynamic forces.
 
When a blade is subjected aerodynamic forces, the blade will undergo a deformation as the structure is inherently flexible.  With the complexity of the geometry and material used in wind turbine blades today, it is difficult to predict how much the blade will deflect.  It is crucial in the design process, therefore, to develop tools capable of predicting how a turbine blade will react to these forces.

The objective of this project is to develop a computational model that can account for the structural dynamics of a typical wind turbine blade. In addition, the parameters of this model can be easily modified for experimentation with various design options of blades.

Beam versus Shell
Approaches for blade modeling utilize both beam models and shell models. The choice between these two models depends on the particular objective of the analysis. The shell model is necessary for investigating details of the blade’s local response, such as buckling or delamination (the separation of layers of the composite materials). However, in a typical operating environment, the overall response of a structurally-sound blade is much like that of a beam. Using a beam model is significantly simpler than the simplest of shell models and is adequate for many investigations. For example, a beam model may be used to explore the blade-fluid interaction, where the objective is to identify forces acting on the blade or details of fluid (air or water) flow around the blade. The results of such analysis can then be used to optimize the parameters of the blade design in order to improve performance and increase power output.

Beam theories can be devised or modified to account for various features of beam composition. The geometric and material compositions of turbine blades have many features that are different from those of beams found in structures such as bridges or buildings. In addition to being thin-walled, the blades have variable and multi-cellular cross-sections that can change abruptly, the wall of the blades typically has variable thickness, and the material of the wall may be different in different parts of the blade (even within a single cross-section).
 

The model developed through this research is a combination of the standard beam and the finite element shell models. The beam component is described by a simple theory (called Euler-Bernoulli beam theory). The model is enriched by including the deformation pattern called warping, described using the finite element methodology. The finite element technique is based on the idea that an accurate representation of a quantity, such as displacement, temperature, or force, over the entire surface area of the blade can be achieved by the knowledge of that quantity at a large number of points on that surface. The finite element technique depends on having knowledge of a sufficient number of points. While thin-walled beam models that account for warping exist, they are typically subjected to various simplifying assumptions and restrictions which are incompatible with the design process of turbine blades. In addition, they are often too cumbersome to be effective for the purpose of exploring various design options. In the model developed in this project, such restrictions are not present and experimentation with various designs is made easy by describing warping using the finite element points. An additional advantage of this approach is that the points used in the finite element analysis are compatible with the data required by blade fabricators in the process of construction.
 

Researchers at the St. Anthony Falls Laboratory have since converted the code that was initially developed in Matlab into C++ so that it can be applied to models of fluid motion around the turbine. The ultimate goal is to create a program that considers the fluid-structural interaction between the air and the blade, so that users can input information about the wind and have the program predict the blade response. At present, the model assumes a stationary blade with wind flowing around it.

Next steps for this research include accounting for complexities such as the effect of the turbine tower and the motion of the blade in the model, as well as customizing the program to a particular blade shape.
 

Ariel Dahl
Professor Henryk Stolarski

As wind turbines grow in size and power output, so do the complications in maintenance and repair work that are necessary to prevent failure. Structural failure of a wind turbine can quickly turn catastrophic as the tall, rotating tower collapses. The goal of one research group at Eolos is to prevent such disasters through structural health monitoring.

Structures like bridges and aircraft bodies that are exposed to high wind loads are already using telemetry for remote monitoring. Telemetry allows workers to discover small structural failures and fractures without requiring human inspectors to be constantly on site. The sensors detect vibrations, stresses, and material deformations using a variety of different signals, and will alert maintenance workers when a change in material behavior indicates a crack or other problem. Wind turbines present some extra challenges to structural monitoring. For example, the rotation of the blades makes powering the sensors difficult. A wired connection to a power source is almost impossible, whereas changing a short-lived battery would cause turbine down time and an interruption to power generation. The schematic below shows how data collected by a sensor node emits a signal that is picked up and recorded by a computer at a remote location.

Harvesting Strain Energy with Zinc Oxide Nanowires

Co-PIs Rusen Yang and Susan Mantell of the University of Minnesota’s Mechanical Engineering department hope to answer the challenges of wind turbine monitoring by using zinc oxide nanowires to harvest the strain energy of the wind turbine blades, powering a sensor indefinitely without wires. Zinc oxide is a semiconductor capable of recovering strain energy, the mechanical energy created when a material deforms under stress, and turning it back into electricity. This technology was first demonstrated only a few years ago.With a grant from the Initiative for Renewable Energy and the Environment, Mantell, Seiler, and Yang and graduate student Dongwon Lim conducted simulation studies of several turbine designs in order to determine whether the amount of power generated from strain energy would be sufficient to enable a sensor to collect and transmit its data.

Nonlinear simulations were performed on 600kW, 1.5MW, and 5.0MW wind turbine designs using FAST (Fatigue, Aerodynamics, Structures, and Turbulence) aeroelastic design code, developed by the National Renewable Energy Laboratory. FAST models the wind turbine as a system of interconnected rigid bodies (nacelle and hub) and flexible bodies (blades, tower and drive shaft), all subjected to dynamic wind loads. The wind loading, blade geometry, and blade material parameters are input to FAST; the output data include blade displacement, forces, and bending moments as functions of rotation. The moments are then related by Hooke’s law to the strain.

The study found that the maximum strain occurs at about 20% to 33% of the blade’s length from the hub, and that larger turbine blades create more strain energy because they are longer and more flexible. The flapwise and edgewise strain were both measured, with more mixed results as to which was larger: on the smallest turbine, the flapwise strain is higher, but on the two larger designs, the edgewise strain is higher. The figure shows the points on a cross-section of the blade where the strain is at a maximum.

The next step for developing a strain energy harvester is to take advantage of the full-scale, 2.5MW turbine at UMore park to obtain more data. Strain measurements from both the simulations and the field will inform the design of an efficient energy harvester. As sensor technology improves, structural health monitoring will be able to detect damage and prevent disaster more effectively.

Interesting Links for the Interested:
Read the news story following the structural failure of a wind turbine in Wasco, Oregon
Read about the first demonstration of strain energy harvesting

Feature photo credit: Patrick O'Leary

In an effort to better understand how wind turbines interact with the atmospheric boundary layer and with each other, Eolos researchers have developed a novel technique to use natural snowfall to visualize the turbulent structures generated down-wind of a utility-scale wind turbine.  These turbulent structures in the wake of a turbine can potentially impact both power production capability and mechanical strain or load on turbines by changing the flows throughout a wind farm. Unfortunately, present understanding of these turbulent structures at field scale is limited by a lack of available tools to visualize and study the turbulence and flow variations produced by wind moving past the tower and rotating blades.

Particle image velocimetry (PIV) is a technique typically used in wind tunnels at small scales to obtain measurements of velocity and turbulence. In the wind tunnel, the flow is seeded with small particles which can be observed with a 2D pulsed laser sheet. A camera fixed perpendicular to the laser sheet captures images of the particles as they flow past the laser sheet. The images can then be processed using computational tools to obtain measurements of velocity and turbulence. 

Given the unpredictable nature of field measurements at a utility-scale wind turbine, PIV has often been thought to be impossible for full-scale measurements of wind turbine turbulence. The tracer particles must be seeded uniformly throughout the flow and the light source and camera must be placed in precise locations. However, as flow characteristics can vary with scale, there is a need to perform actual measurements of turbulence on a field-scale wind turbine. 

To address these problems, an Eolos research team, led by Jiarong Hong, assistant professor of mechanical engineering at the University of Minnesota, explored the use of natural snow during a winter snow storm as tracer particles in the wake of the wind turbine at the University of Minnesota’s Eolos Wind Energy Research Field Station. PIV measurements must be taken at night so that the light sheet illuminates only the particles in a 2D plane behind the wind turbine. The light sheet in this particular pilot experiment was provided by a 5 kilowatt search light and a convex reflector that converts the beam from the searchlight into a light sheet that extends up to nearly 40 meters. 

Starting around 2:00 a.m. one morning in February 2013, a team of University of Minnesota Researchers collected nearly an hour and a half of PIV data in the form of pictures and videos. The results of this data collection demonstrate the feasibility of using natural snowfall to visualize the full-scale turbulent structures in the wake of a 2.5 megawatt wind turbine.

Future work of the research team will focus on using the pictures and video gathered using this technique to obtain actual measurements of the turbulence generated by a utility scale wind turbine. With advanced understanding of the turbulence produced by wind turbines, researchers hope to enhance wind farm siting and control to improve the efficiency and reliability of wind turbines.