Development of loss and noise models for fan design based on high fidelity simulations

点击翻译按钮获取中文摘要

The main objective of this research proposal is to advance knowledge in the design and installation of turbomachines, to increase their efficiency and lower their acoustic emissions. Widely used in the fields of energy production, aircraft propulsion, ventilation and processes, turbomachines are the key components of many engineering solutions developed to adapt to future challenges involving sustainability and mobility. Canada is undergoing major changes with digital technologies driving innovative concepts in the Industry 4.0 framework. A rapid evolution towards more efficient and versatile rotating machines is needed to equip future light flying vehicles and to respond to on-demand operation and various loads in ventilation or energy production. Social acceptability is also a challenge for new concepts and in particular noise emissions in urban and commercial areas are to be considered at early stages of their development. The flow in a turbomachine is intrinsically unsteady, characterized by strong interactions of secondary flow features with the main stream flow and with the fixed and rotating components of the machine. These interaction mechanisms yield substantial performance losses and noise emissions, that must be taken into account in advanced design tools. The planned research activities will focus on the mechanisms related to (i) boundary layer development on the rotor blades and (ii) the tip leakage vortical structures in the tip clearance. Several configurations with increasing complexity from isolated fixed airfoil to rotating turbofan blade will be simulated using a high-fidelity computational approach. This approach will make it possible to capture the complex turbulent flow structures both in time and space with a minimal use of turbulent closure modelling. The projected results will provide unique and large databases for the development of wall-pressure fluctuation and tip clearance entropic loss models. They will also provide detailed insights as to the interaction mechanisms in the tip gap responsible for both instability and the noise generation. To build improved models, modern machine learning techniques will be used. In the long term, the knowledge acquired will improve classical pre-design tools for turbomachines to extend their range of validity, provide quantitative and parametric performance predictions and reduce noise emissions.