Development of numerical methods for low pressure turbine tone noise

One of the major concerns in the urban areas surrounding airports is the noise pollution caused by the acoustic emissions coming from aircraft operations. It has been demonstrated that a continuous and prolonged exposure to noise pollution, not only results in a lower quality of life, but has harmful effects on human health as well. Definitely, noise may be the cause of psychological disorders, such as insomnia, anxiety, aggressiveness, and physiological disturbances, such as hearing impairment, hypertension and ischemic heart diseases. For these reasons, since the 1960s, when turbojet engines spread in the aviation industry, regulations on aircraft noise emissions have been introduced. Afterwards, due to the continuous growth in air traffic, these legislations have become more and more restrictive. This led to the introduction of the noise certification process for commercial aircrafts at the operating points critical in terms of noise perceived on the ground. Due to the introduction of noise regulations, the goal of reducing the acoustic emissions produced by aircrafts has been a constant challenge for aeronautical industry and engine designers continuously strives hard to to develop quieter components. In this context experimental and numerical investigations on noise generation and propagation mechanisms have become a main research field for aeronautical manufacturers. Such studies are aimed at gaining a deeper insight on the physical phenomena in order to develop effective low noise strategy. In particular, well-calibrated and reliable numerical tools for noise prediction can be used within the engine design loop allowing the evaluation of the acoustic emissions caused by each component before it is tested in a demonstrator. As a consequence, these tools are of great importance in the reduction of aircraft noise. Aeroacoustic simulations within an engine environment can be performed by means of different techniques, ranging from analytical models to 3D numerical solvers: fast and robust methods have primary importance at the beginning of the design loop, when it is essential to gain a general understanding of the effect of primary design parameters on the noise emissions. On the other hand, three-dimensional aeroacoustic solvers, more accurate and reliable, are able to to provide detailed knowledge of the acoustic phenomena during the advanced design verifications. Aircrafts acoustic emissions can be divided into either external and internal noise. The former contribution arises from the interaction between the airflow and the aircraft itself (fuselage, wings, control surfaces, etc...), the latter instead is generated inside the engines as a result of various mechanisms and then radiates outside. At the dawn of civil aviation, the most critical noise source was the exhaust jet, but with the introduction of modern high-bypass ratio turbofan engines jet noise has been strongly reduced. Also fan noise has been drastically lowered over the years taking advantage of low-noise design criteria and acoustic treatments in the bypass duct. Hence, the other engine noise sources, such as the low-pressure turbine (LPT), have become significant in the overall aircraft emissions, first and foremost at some critical operating conditions. For this reason, nowadays a wide range of components need to be investigated during the engine design in order to apply low-noise criteria required to meet the increasingly restrictive noise regulations. In this context, the main topic of this PhD thesis is the development of two different methods to be used when evaluating the tone noise emissions of low pressure turbines. The first one is a two-dimensional model that computes the transmission of an acoustic wave across a blade row. This method has been included in a preliminary noise prediction procedure capable of quickly estimating the noise emissions of a multistage turbine. The second method allow the extraction of the acoustic waves from a 3D unsteady CFD solution obtained by Traf code. Traf code is commonly used to perform URANS simulations of a pair of neighbour blade rows. During this PhD activity, a post processing technique of the unsteady solution, based on the discrete Fourier transform, has been implemented in order to extract the acoustic components from the time depending solution at each BPF. Finally, after the DFT in time domain, tone noise levels can be calculated in terms of SPL and PWL values by means of radial mode analysis. The key aspects of this procedure are the capability to provide detailed results in terms of acoustic emissions for rotor/stator interactions within the pair of rows with a single simulation and to account for non-linear aeroacoustic effects, which can be relevant in modern low-pressure turbine environments (high pressure ratio, low axial spacing). The validation of these methods has been carried out by comparing their results both with the results of simulations coming from a previously validated approach based on Lars code and acoustic experimental data measured at a cold-flow rig. Finally, although these methods have been developed specifically for low-pressure turbines, they can be actually used in any axial turbo-machinery environment where rotor-stator interactions are relevant sources in the overall emissions.