Blog

How to Optimize Wind Farms with CFD and Wind Turbines Placement

How to Optimize Wind Farms with CFD

Most renewable energy sources come from the sun. Overall, hourly, the sun is estimated to provide the earth with over 175 Trillion kilowatt hour of energy and, of this, about 1-2% is converted to wind energy. The places on the equator are more directly heated compared to the poles leading to the global wind systems. In addition, the local environment like the presence of cities, water bodies, forests alter the local wind profiles in any area. In the last decade, an emphasis on renewable energy sources and rapid growth of wind farms has led to significant decrease in associated costs.

The wind turbines developed in the 1980s became the breakthrough for the modern turbines. The “wind rush” that started in the 80’s suddenly ended overnight when the California government stopped its support schemes. In the last decade starting 1998, the MegaWatt-sized turbines and large wind farms have revolutionized the wind energy industry and making it a potent source of renewable energy. Spearheading this revolution has been Denmark and Germany, who are also the largest wind energy markets.

In this article, we explore ideas for usage of the SimScale 3D simulation software and the exploitation of its features towards design, simulation, and optimization of the power output of wind farms.

Betz Law and Power Efficiency

The simplest way to assess the efficiency of a wind turbine has been through the Betz Law.

BetzTubeFig 01: Schematic of fluid flow across a wind turbine (Image source: Wikipedia)

Consider the flow of air through an imaginary tube as shown in Fig 01. Initially, the velocity of undisturbed air, downstream, is v1 m/s and upstream, after passing through the turbine, is v2 m/s. The average mass of air (m) flowing through an area (A), can be written in terms of density and velocities as

Mass transferThe kinetic energy of the originally undisturbed air can be written as

Kinetic energy of air

This change in velocity, as the wind passes through the rotating turbine blades, signifies the transfer of kinetic energy from the wind to the turbine. Using the above relation for mass, the change in kinetic energy of the wind can thus be written as

Change in kinetic energy

A simple measure of efficiency of the wind turbine can be given, by the ratio of kinetic energy of change of energy to the undisturbed energy of the wind, asEnergy efficiency

Optimization Parameters: Wind Farms

There are several aspects that are of importance in designing wind farms. Some of the important aspects include:

  • Roughness & roughness length: As one moves higher in altitude, the wind speeds are much less influenced by the ground while at lower altitudes, the winds are affected by friction with the ground. Roughness length is that length, above the ground, up to which the wind speed is considered to be zero. For example, the presence of cities and forests can slow down the wind considerably. In contrast,  grassy or concrete land offers much less friction. Finally, water bodies offer almost nearly zero friction to the flow of winds. Now, that’s a clue why there are several offshore wind farms! Overall, the terrains are characterized in terms of roughness classes: Water is roughness class 0; grasslands / concrete areas are 0.5; cities & large forests have a high roughness number of 3-4.
  • Wind speed variability: The wind speeds are not constant but vary through the day and also show seasonal changes. It is important to consider the magnitude of these variations. As shown in Fig. 02, the power output of the turbine depends on the wind speed and are only operational between preset cut-in and cut-out speeds only. Optimality of wind conditions, as expected, is a critical parameter in optimization.

power vs. wind speed

Fig 02: Power output vs. wind speed of wind turbines

  • Wind obstacles: Obstacles are caused due to various reasons like presence of cities, forests, houses etc. In the presence of such obstacles, wind speeds are decreased to a large extent. In addition, they cause turbulence which can decrease the quality of wind energy. A rule of thumb considered is that the hub is at least 30 ft higher than the nearest obstacle. In addition, several other effects commonly observed include the park, tunnel and hill effects as shown in Fig. 03. Park effect refers to the distance between the towers. Each turbine produces a wake and placing a turbine in the wake can lead to significant inefficiency. The later part of the article demonstrates the park effect in more detail.

Effects observed

Fig 03: Effects commonly observed in design of wind farms

  • Tunnel and hill effects: As we have generally observed, the wind speeds are much higher when compressed into smaller places like between mountains, buildings or at high altitudes etc. Strategically, placing the wind turbines in valleys or above hills (and elevated locations) can improve the wind speeds and the overall efficiency of the system. However, if the mountain terrain is significantly rough, it can also increase the surface roughness and needs to be accounted for in the design.
  • Tower heights: It is generally expected that the hub height (or height at which the rotor is attached), is at least 30ft higher than any obstacle in the nearby vicinity of 100 m. The rotor diameter and hub height largely determines the power that can be generated by a wind turbine. The area of the rotor determines how much wind energy can be collected and transformed into electrical energy. The wind speed increases with height. Given that the power generation is proportional to the third power of the wind speed, the hub height has a major effect on the overall power output of a wind turbine. Fig 04 shows the variation of nominal power with rotor diameter (left) and hub height (right) while Fig. 05 shows the evolution of these over the recent decade. Over the last two decades, possible rotor diameters and hub heights have increased making wind energy a potent source for the coming decade.

Rotor Dia & Hub height vs. Power

Fig 04: Variation of nominal power as a variation of rotor diameter and hub height in wind farms (Image source: Keiler, J., and Häuser, H. Betreiberdatenbasis. IWET Datenbank)

Evolution of rotor diameter / hub heights

Fig 05: Evolution of rotor diameter and hub height in wind farms over the years (Image source: Keiler, J., and Häuser, H. Betreiberdatenbasis. IWET Datenbank)

  • Tower arrangement: Finally arrangement of towers is one of the most important aspects. This is also more commonly known as the park effect. The general rule of thumb is that the distance between turbines is about 7 times the rotor diameter along the head wind. Along the cross-wind, the distance is expected to be at least 4 times the rotor diameter. However, recent studies have shown that it should be at least double of this. There have also been recent studies that are using the ideas from nature (like fish schooling, the flight of long distance birds) to better understand how wake and turbulence could be used to boost overall efficiency. This article specifically discusses how SimScale could be used to simulate the tower arrangement and optimize for best possible solutions.

Simulation of Wind Farm

In this article, wind turbines with a hub height of around 82m and the rotor diameter of 80m are considered. Rotor speed is considered to be constant with an angular velocity of 1.8325 rad/s. Such a system is generally rated around 2MW. The international standard for measurement for wind speeds is at a height of 10 m from the surface and the velocity varies as

velocity variationwhere Z0 refers to the roughness length, Z to the distance from the ground, Zref is the reference height at which the measurements are made (here 10m) and vref is the wind velocity at the reference height (here 5 m/s). We assume that the roughness is significantly small and of the order of 0.1m. In the simulations discussed below, in order to avoid the singularities of a log function, the above relation is simplified to an affine function as

linear velocity

The fluid flow can be considered as an incompressible turbulent flow in a steady state. Here the k-Omega SST turbulence model is used. The entire project can be accessed through the public projects database and can be imported to your account. Different arrangements are considered to analyze their effect on the wind velocity profiles and related turbulence.

Wind turbines placed one behind the other

Starting with the worst case scenario, where three wind turbines are aligned one behind the other as shown in Fig. 06.

Vertical configuration plus wakes

Fig 06: Configuration of the wind turbines (left); Wakes observed upstream and its interference with the subsequent turbines (right)

Fig. 06 also shows a cut slice of the simulation box and the influence of the first turbine on the subsequent turbines upstream is visually evident. A simple estimate shows that the input velocity for the first turbine is around 20 m/s and the output is around 10 m/s. On the contrary, the velocities downstream and upstream of the second and third turbines are almost same and leading to nearly zero power efficiency in them.

Turbulence in turbine

Fig 07: Velocity profile downstream and upstream (top-left); Velocity (u-y) depicting turbulence (top-right); Velocity (u-x) upstream from first turbine (bottom-left) and third turbine (bottom-right) shows changes in turbulence profiles

The above velocity profiles in Fig. 07 confirms the conventional knowledge that such a placement is highly undesirable. There is a significant influence of wakes and turbulence from the turbines in front. However, if the turbines are placed reasonably far away, the effects of such wakes can still be negotiated.

Wind turbines placed one beside the other

Alternatively, it is possible to place the turbines one beside the other, to the direction of the head wind, as shown in Fig. 08.

Horizontal placement

Fig 08: Configuration of turbines (left); Velocity (uz) profile downstream and upstream including depiction of wakes (right)

Alternative to earlier, now if the wind turbines are placed beside each other as in Fig. 08,  all the turbines are faced with the same wind speeds. Thus, as the air flows across the turbines, the turbines can be equally efficient.

Horizontal arrangement

Fig 09: Downstream wind velocity profile u-x (top-left) and u-y (top-right) demonstrate the influence of turbulence caused by each turbine; velocity profile (uz) upstream (bottom-left) and downstream (bottom-right) of each turbine.

Fig. 09 depict the velocity profiles. The downstream velocity profiles for u-x and u-y shows a little interaction between the turbines. The turbines are just placed about three rotor diameters away in these simulations and such an interaction should be expected. At double these distances, such interactions tend to zero. Such simulations can be easily performed with SimScale tools with much larger distances to optimize the wind farms.

Wind turbines placed along a diagonal

The last configuration is to place the turbines one beside the other, along a diagonal, to the direction of the head wind, as shown in Fig. 10. Most commercial wind farms use such a configuration with rows of wind turbines.

Placement of turbines in diagonal wind farms

Fig 10: Downstream wind velocity profile u-x (top-left) and u-y (top-right) demonstrate the influence of turbulence caused by each turbine; velocity profile (uz) upstream (bottom-left) and downstream (bottom-right) of each turbine.

Wind turbines need to always face the direction of the wind since any wind from the side will not cause rotation of the turbines. Thus, the turbines are allowed to “yaw” or rotate towards the direction of the wind. Placing the turbines along a diagonal can also result in much smaller areas required. As shown in Fig. 11, as long as sufficient distances are maintained along the two directions, it can provide a compact way of packing more turbines in the same area.

Velocity profiles - wind farm simulation

Fig 11: Downstream wind velocity profile u-x (top-left) and u-y (top-right) demonstrate the influence of turbulence caused by each turbine; velocity profile (uz) upstream (bottom-left) and downstream (bottom-right) of each turbine.

As shown in Fig. 11, the diagonal distance between the turbines is same as the horizontal distance in the previous configuration. Yet, just assessing pictorially, the interaction of wakes generated by the turbines are much smaller.

Additionally, one does not necessarily need to have the same number of turbines as in the real farms. Using periodic boundary conditions that couple two opposite faces will be equivalent to have a very big system in that direction. Such a configuration demonstrates significant potential and larger distances can be simulated using SimScale to overall optimize the wind farms. The simulation with periodic boundary condition is left as an exercise to the reader.

Overall, this is just a glimpse of how SimScale could be used to simulate practical problems of significance interest. Much larger wind farms can easily be considered to test the fluid flow patterns and optimize their placements. In addition to the park effect discussed above, hill and tunnel effects, the effect of rotor diameter and / or hub height could also be explored using SimScale. So next time, you plan to build a wind farm, pop-in to take advantage of the SimScale simulation environment.


Download the company’s booklet to learn more about SimScale.

SimScale is the world's first cloud-based simulation platform, enabling you to perform CFD, FEA, or thermal analyses. Sign up for the 14-day free trial and join the community of 70 000 engineers and designers. No payment data required.