EsBo is a research project funded by the German Federal Ministry for Economic Affairs and Energy. The project specifically focuses on the noise emissions of modern wind turbines.

In Germany, strict noise protection regulations are in place to protect the population. Especially during nighttime operation, these regulations often require wind turbines to operate at reduced power. Due to the limited accuracy of current prediction models, these restrictions are usually very conservative. As a result, the actual noise emissions are statistically well below the permitted limits. This significantly reduces the potential energy yield of wind turbines.

The overall goal of the EsBo joint research project is to improve noise prediction methods and to use them for the design of low-noise components and acoustically optimized control methods. In this way, the energy yield of wind turbines can be increased.

Within the subproject at the University of Rostock, modulation methods for the generator-side converter are being investigated. The aim is to avoid mechanical excitation of the generator in frequency ranges that result in high acoustic emissions.

For this purpose, a simulation toolchain is first being developed in close cooperation with WRD and the University of Hannover. This toolchain represents the complete chain of effects: from harmonics in the converter output voltage, to the resulting harmonics in the generator current, the corresponding harmonic electromagnetic forces in the generator, and finally the resulting noise emissions.

Using this toolchain, different modulation methods are investigated at the University of Rostock. In the first step, pulse-width modulation (PWM) with a variable switching frequency is considered. In the second step, optimized pulse patterns are investigated. For this purpose, both a method using pre-optimized switching angles and a method based on online model predictive control (MPC) will be evaluated regarding their suitability.

The control and modulation methods are first analyzed using simulations. Afterwards, they will be experimentally tested on a 50 kW test bench.