Wind turbine noise code benchmark: A comparison and verification exercise

In a number of institutions and companies, researchers and engineers are developing numerical models and frameworks that are used to predict the aerodynamic noise emissions from wind turbine rotors. The simulation codes range from empirically tuned engineering models to high-fidelity computational ones. Their common feature is the fact that they all specifically model the main aerodynamic noise mechanisms occurring at the rotating blades (namely, the turbulent boundary layer): trailing-edge and turbulent inflow noise. Nevertheless, different modelling techniques and implementations may generate different results, even when assessed on the same rotor design and operating conditions, which raises the question of the actual fidelity and reliability of these models. Trailing-edge noise is put at the forefront of the present study, as it is recognized to be the main source of audible noise from modern wind turbines. The present benchmark aims at comparing the results from different modelling approaches and drawing some conclusions from these comparisons. This effort, denoted as Wind Turbine Noise Code benchmark, was initiated in 2019 as a joint activity between the IEA Wind Task 39 (Quiet Wind Turbine Technology) and Task 29 (Detailed Aerodynamics of Wind Turbines, now Task 47). In addition to the investigation of the noise emissions themselves, the rotor aerodynamic charac teristics are investigated, as they are the source of the noise generation mechanisms discussed herein. A number of test cases are defined, and the aerodynamic and aeroacoustic predictions from the various models are compared. A fair agreement between the aerodynamic predictions is observed. There exist some discrepancies between the different noise prediction methods, but it is difficult to conclude if one methodology is better than another in order to design a wind turbine with noise as a constraint.