The rapid growth of wind energy production in recent years has reached a stage where the aerodynamic noise emission from wind turbines is a critical issue to overcome in order to successfully continue increasing the scale of the turbines and reducing the cost of energy (CoE). The fundamental principle to tackle the aerodynamic noise issue is to design the turbine blades in such a way that the source of noise is alleviated without changing the aerodynamic efficiency of the blades that is critical for CoE. One of the most effective ways to achieve such a noise-noise blade design is to use "serrations" on the trailing-edge of the blades from which the noise emission is strongest (at an operating condition). Vestas (one of the largest wind-turbine manufacturers in the globe) has been successful in developing blades with serrated trailing-edges (STEs) in recent years and they are in service now. However, the research on STEs is still underdeveloped and there are various areas where STEs should be better understood and improved upon what they are at present. Vestas aims to make a major breakthrough in the development of STEs within the next a few years in order to successfully implement the technology in their next-generation wind turbines.

The proposed PhD project at the University of Southampton is part of the Vestas multidisciplinary programme for the development of next-generation STEs. This particular PhD project aims to achieve detailed understandings of the physical mechanisms of the noise generation and its reduction due to the STEs, and to derive a semi-empirical engineering model that provides predictions of the noise reduction through the STEs for various geometries and flow conditions. The project will be carried out mainly based on numerical simulations (large-eddy simulations) and some mathematical derivations for the prediction model. The large-eddy simulations will be performed by using an in-house code CANARD (Compressible Aerodynamic & Aeroacoustic Research coDe) developed at the University of Southampton. The code is based on high-order finite-difference methods and is fully parallelised on an MPI platform (running on the national supercomputer ARCHER as well as the local IRIDIS-4 cluster with a supra-linear scalability with up to 10,000+ processor cores). Some more relevant information about the computational work can be found in https://doi.org/10.1017/jfm.2016.841.