Broadband flow-induced noise for airfoils: large inflow turbulence distortion effect on leading-edge noise generation and prediction
Fernanda Leticia dos Santos is a PhD student in the department Engineering Fluid Dynamics. (Co)Promotors are prof.dr.ir. C.H. Venner and dr. L.D. De Santana from the faculty Engineering Technology.
Many marine animals have developed an excellent sensory perception of sound, strongly relying on acoustic communication to navigate, locate prey, avoid predators, and reproduce. Unwanted sound, i.e., noise, dramatically affects marine animals. The propeller of marine vessels significantly contributes to underwater noise. When propeller cavitation occurs, cavitation noise is the dominant noise source. Thus, avoiding cavitation inception is important, as reflected in the guidelines of the International Maritime Organization. In addition, cavitation must be avoided due to strict noise requirements for certain vessels, such as fishery research ships, oceanography research ships, and naval vessels. In this thesis, noise sources for noncavitating propellers are studied, specifically flow-induced broadband noise sources.
The primary sources of broadband noise for a propeller blade section are leading- and trailing-edge noise. Leading-edge (LE) noise is generated by the impingement of the hull wake in the propeller LE. It is usually dominant in the low- and mid-frequency ranges, whereas trailing-edge noise is dominant for high frequencies. Low-frequency sound waves are of great interest for marine applications because the noise generated by vessels is dominant in this frequency range and these waves propagate for long distances due to the high speed of sound in water and the low dissipation. Therefore, LE noise is one of the most relevant broadband sources for noncavitating marine vessels. Accurate LE noise prediction is paramount to designing silent propellers. One of the most used prediction methods is Amiet's theory. This theory neglects the blade section geometrical effect on the near-field and radiated noise. This results in inaccurate noise predictions for foils in the mid- and high-frequency ranges. This inaccuracy is attributed to the distortion of the turbulent inflow as it approaches the foil LE. Hence, to improve LE noise prediction, the turbulence distortion phenomenon should be understood and quantified so that it can be accounted for in Amiet's LE noise prediction model.
Trailing-edge (TE) noise is generated by the interaction of the TE region of the blade with the boundary layer and the near wake flow and strongly depends on the boundary layer at the foil TE. In scaled tests, tripping devices are installed on the model surface to hasten the laminar-turbulent transition due to the relatively low Reynolds numbers in these tests. As these devices affect the boundary layer development and thereby the source of TE noise, a better understanding of the influence of tripping devices on the TE noise source and radiated noise is needed.
This thesis aims to investigate the noise generation mechanisms of broadband noise sources relevant for a noncavitating marine propeller profile section, with the objective of enhancing noise predictions for the dominant noise source. Aerodynamic and aeroacoustic experiments were performed using various profiles and conditions of inflow turbulence in the Aeroacoustic Wind Tunnel of the University of Twente. Even though the motivation of this thesis originates in marine applications, all measurements were performed in air. This approach is suitable because the current research focuses on the noise generation mechanism, i.e., the turbulence interaction with the foil surface, which will be similar when cavitation is absent. Also, the relevant nondimenional numbers (Mach number, Helmholtz number) are comparable for the propeller operating underwater and the airfoil used.
This thesis is divided into three parts. In the first part, the remote microphone probe (RMP) technique and the noise sources of an airfoil are investigated. The RMP technique is shown to be well suited for measuring the pressure fluctuations for a turbulent with the designed experimental setup. An airfoil with pressure ports that can be used as RMPs was designed and manufactured. A NACA~0008 airfoil was chosen because this geometry is representative of a marine propeller blade section. The wind tunnel results show that LE noise is dominant for this geometry in the entire frequency range measured (nondimensional frequency based on the free-stream velocity and the airfoil chord length up to 31.5).
In the second part, the distortion of the turbulent inflow in the vicinity of the LE of various airfoil geometries is investigated, eventually leading to improvement of LE noise prediction. First, the dissipation range of the von Kármán turbulence spectrum is modeled based on the experimental data acquired. This modified model agrees well with literature results and is expected to influence the LE far-field noise for high frequencies. The turbulence distortion is experimentally and numerically investigated for different airfoil geometries. The turbulence distortion affects the near field of the LE noise, i.e., the turbulent inflow near the LE, the spanwise correlation length, and the wall-pressure fluctuations (WPFs) in the LE region, consequently also impacting the far-field radiated LE noise. Lower LE noise levels are observed for mid and high frequencies at most directivity angles for thicker airfoils. Empirical formulations are proposed to model the integral length scale and the streamwise velocity fluctuations at the stagnation line of an airfoil as a function of the relevant airfoil geometrical parameter for the turbulence distortion, i.e., airfoil maximum thickness. These formulations are essential because more accurate LE noise predictions are obtained when the turbulence parameters used as input to Amiet's model are extracted near the airfoil LE as shown in this research. Furthermore, the turbulence spectrum near the LE has a high-frequency decay different from the prediction of the rapid distortion theory (RDT). Based on the experiments, a new formulation is developed for the turbulence spectrum. A much better agreement between experimental and predicted LE noise is observed when this new turbulence spectrum formulation is used with as input the turbulence parameters obtained near the airfoil LE. This leads to a new approach to account for turbulence distortion in Amiet's LE noise model. As the turbulence distortion also impacts the WPFs in the airfoil LE region, poor estimations of the WPF spectrum using Amiet's model are observed. Also here a better agreement between experimental and predicted WPF spectra is observed when the turbulence parameters used as input to Amiet's model are extracted near the airfoil LE. This shows the importance of accounting for the turbulence distortion in Amiet's model for LE noise and WPF spectrum in order to obtain accurate predictions.
In this thesis, it is also shown that the turbulent flow at the stagnation line of a cylinder is similar to the turbulent flow at the stagnation line of an airfoil as long as the cylinder diameter is representative of the airfoil average thickness in the LE region (average thickness from the LE position up to the position of maximum thickness). This result suggests that the RDT for a cylinder can be used to model the turbulence distortion for an airfoil. Finally, the implications of incorporating the turbulence distortion in Amiet's LE noise prediction are analyzed for propeller blade sections. The predicted LE noise levels are lower when turbulence distortion is accounted for. However, this noise mechanism is dominant in the low- and mid-frequency ranges. The frequency at which the dominant noise source shifts from the LE to the TE reduces when turbulence distortion is considered. The analysis also shows that the observed frequency range in which the turbulence distortion affects the LE noise is within the range where noise generated by a noncavitating marine application is relevant. This demonstrates the relevance of the results for the design of silent noncavitating propellers.
In the last part of this thesis, it is shown that tripping devices affect the transition process from a laminar to a turbulent boundary layer and the characteristics of the boundary layer developed at the TE, such as the WPFs. When the trip height increases, the spectral level of the WPFs increases in the low-frequency range, resulting in an increase of TE far-field noise, reaching up to 4~dB.
In conclusion, the results discussed in this thesis contribute to a better understanding of the turbulence distortion mechanism for airfoils and its effect on LE noise. The proposed approach to account for the turbulence distortion effect on Amiet's LE noise prediction can be used to obtain more accurate noise estimations for airfoils and to consider the airfoil shape effect on LE noise during the design phase of a propeller blade, for example. This improves the ability to design more silent noncavitating propellers and other lift-generating surfaces, thus positively contributing to the well-being of both humans and animals affected by noise pollution. In addition, the discussion on the influence of tripping devices on the turbulent boundary layer developed at the TE and on the radiated TE noise can be used as a first guideline to determine the appropriate trip height for aeroacoustic measurements and to understand the uncertainty introduced by the tripping device on the determination of the boundary layer parameters and on the measured noise. Finally, the RMP technique is documented in detail, which can help other researchers interested in starting working with this technique and in knowing the comparability of the results obtained with RMPs to a standard WPF measurement technique.