Noise from wind farms falls into two categories: aerodynamic noise is created by the blades of a turbine swishing through the air, while mechanical noise is associated with the machinery housed in the nacelle of a turbine. As mechanical noise tends to be tonal, it is this that is most often a nuisance factor for residents living nearby. As a result, there are strict regulatory standards throughout Europe and North America, and when operators do not meet these requirements, they face potentially heavy penalties.

Figure 1. Acoustic-Structural Interaction model of a wind turbine surrounded by air showing the vibration acceleration amplitude of the turbine, and slices through the air showing the resultant sound pressure level. The figure shows vibration with localized amplification of the noise at the top of the tower (red area).
Xi Engineering Consultants specializes in complex vibration issues. One of Xi’s recent projects was carried out for the manufacturer of a set of megawatt-scale wind turbines emitting excessive

noise in the 800-830 Hz frequency band.

Xi has extensive experience with simulation and used COMSOL Multiphysics to develop a model of the tower, the rotor blades, and the nacelle housing the gearbox (Figure 1). The model incorporated all of the air inside and outside the turbine so that engineers could identify the modal shape of each vibrating element and see exactly how vibration was travelling out of the gearbox. This led them to the tower wall where resonances around 820 Hz were clearly visible. The next step was to make an eigenfrequency model of the tower structure and skin. Dr. Brett Marmo is Senior Consultant for Xi. He noted that within the complex, there were certain hotspots near the top of the tower where the tower skin was rippling at resonant frequencies of 800 and 830 Hz. These resonances were amplifying the gearbox vibration and producing the annoying tonal noise.

Figure 2. The Sound Pressure Level measured at a virtual microphone outside that tower for the wind turbine in its native state, when the top 20 m2 of the inside wall is covered with the anti-vibration material, and when the top half of the tower (300 m2) is covered in the anti-vibration material.
In this case, the design solution was to break the vibration pathway between the gearbox and the tower wall. The simplest method would have been to modify the rubber buffers below the gearbox. However, engineering constraints meant that it was impractical either to stiffen or soften these. The most effective alternative solution was to coat the inside of the tower with a specialist material that reduces the amplitude of the vibration. The big question was exactly how much material would be required to solve the noise problem. The only option was to create a third model in COMSOL Multiphysics to simulate the effects of the material inside the tower (Figure 2).

The high costs of the material would have meant that full coverage of the 600 m2 surface inside the tower would have cost tens of thousands of pounds. Even if material had only been applied to the top part of each tower, 100 m2 per unit would have been needed. Given the number of turbines, the bill would have been substantial. Instead, Xi was able to advise the client that only 20 m2 of material was required for each tower. One prototype installation confirmed the findings exactly: the noise was sufficiently attenuated.

This work was done by Xi Engineering Consultants Ltd., using COMSOL Multiphysics. For more information, Click Here .