Inflow Turbulence Noise Research Articles

Noise of a shrouded propeller due to ingestion of grid-generated turbulence

An experimental investigation of the noise produced by a shrouded propeller interacting with grid-generated turbulence in an anechoic wind tunnel is presented. Inflow turbulence statistics and the far-field noise directivity were measured. Measured inflow turbulence statistics and propeller operating parameters were used in a theoretical model to predict the turbulent inflow noise spectra and directivity. The theoretical model uses rapid distortion theory to calculate the distorted turbulent field at the propeller plane. The unsteady loading on the rotating blades is calculated using isolated blade response functions. The noise radiated to the far-field is then evaluated using a tailored Green's function approach which accounts for the acoustic scattering effect of the shroud. The theoretical model is able to capture the shape of the directivity pattern of the turbulent inflow noise produced by the shrouded propeller. The agreement between the measured and predicted spectra is reasonable, however the high frequency broadband levels tend to be underpredicted.

Sensitivity Analysis of Wind Turbine Broadband Noise Estimation to Semi-Empirical Models Parameters

The continuous increase of energy demand and the rising concerns on climate change, are pushing the European Union decarbonization strategies and transition toward renewable based energy systems, with wind energy playing a leading role. It is therefore necessary to have a better understanding of how wind turbines (WTs) impact on their surroundings, including their noise emissions. Among the different methods to compute noise emissions of WTs, semi-empirical models are a valid choice to have a-priori estimations of noise spectra and sound pressure levels. These models are based on correlation laws for different physical mechanisms that contribute to noise generation. Popular models for dominant noise sources include the Amiet approach for inflow turbulence noise and the Lowson model for turbulent boundary layer-trailing edge noise. Determining the parameters involved in these models can be challenging, potentially leading to significant errors in noise prediction. In this study, we conducted a novel sensitivity analysis of the models by varying different parameters such as turbulent intensity and dissipation, boundary layer thickness, and temperature. The selected test case is the reference multi-MW horizontal axis wind turbine Neg-Micon 80. The results of the multilevel-multivariate analysis, involving 63,360 combinations of the input parameters, clearly demonstrate a significant dependence of these models on atmospheric turbulence parameters. Furthermore, these models exhibit an higher sensitivity to input parameters at lower frequencies of the noise spectrum, which are generally associated with higher values of sound pressure level.

Propagation effects in the synthesis of wind turbine aerodynamic noise

The sound field radiated by a wind turbine changes significantly with propagation distance, depending on the meteorological conditions and on the type of ground. In this article, we present a wind turbine noise synthesis model which is based on theoretical source and propagation models. The source model is based on Amietâ’s theory for the prediction of the trailing edge noise and the turbulent inflow noise. The trailing edge noise uses the wall pressure spectrum calculated with Leeâ’s model for the suction side and Goodyâ’s model for the pressure side. The Kolmogorov spectrum is used for the prediction of the turbulent inflow noise. To account for the propagation effects associated with atmospheric refraction and ground reflection, a wide angle parabolic equation in inhomogeneous moving medium is considered. The scattering due to the turbulence in the atmosphere is accounted for using the Harmonoise model. The synthesis method is based on the moving monopole model to accurately predict the amplitude modulations at the receiver, and uses cross-fading between overlapping grains to obtain the time signals from the frequency-domain prediction model. Finally, audio signals are provided for a few test cases to emphasize various propagation phenomena associated with wind turbine noise.

Turbulent inflow noise produced by a shrouded propeller

A theoretical model is presented for predicting the noise produced by the interaction of turbulence with a shrouded propeller in an axisymmetric mean-flow. For cases where the free-stream velocity is low, the turbulent eddies well upstream of the propeller are significantly elongated as they pass through the induced streamtube contraction. These elongated turbulent eddies interact multiple times with the rotating propeller blades causing partially coherent loading on the blades. The resulting radiated acoustic pressure spectrum contains peaks which occur at harmonics of the blade passing frequency. The theory is developed in the frequency domain and incorporates rapid distortion theory to obtain the distorted turbulent field at the propeller face, accounting for the potential mean-flow effect of the propeller and the shroud. The unsteady loading on the rotating blades is calculated using isolated blade response functions. The noise radiated to the far-field is then evaluated using a tailored Green's function approach which accounts for the scattering effect of the shroud and is solved using a boundary integral approach. Results are presented for shrouded and unshrouded propellers operating in flows with a range of different free-stream velocities. It is shown that as the free-stream velocity increases, the peak in the noise spectrum at the blade passing frequency reduces in level and the ‘haystack’ surrounding this peak widens. Results showing the effect of the shroud on the radiated noise levels at different frequencies and observer locations are also presented.

Impact of turbulent inflow and orography on the low-frequency noise sources of a wind turbine

A high-fidelity process chain is used to numerically investigate the low-frequency emissions from a generic wind turbine under turbulent inflow conditions. It consists of the Computational Fluid Dynamics (CFD) solver FLOWer, which is coupled to the multi-body simulation software SIMPACK to take unsteady aeroelastic effects into account. A realistic flow field for the complex terrain of Perdigão, Portugal is obtained by including the orography and vegetation in the simulation and using a precursor simulation with E-Wind to generate a site- and situation-specific inflow. The acoustic emissions are calculated with the Fflowcs-Williams-Hawkings (FW-H) acoustic solver ACCO. The simulations show that the tower emits noise caused by the blade-tower interaction (BTI) equally strong in all directions. To the sides of the turbine, this contribution is dominant, regardless of the turbulent inflow. The blades, on the other hand, emit significantly more noise under turbulent inflow, especially in streamwise direction, where they become the dominant source. The main noise mechanism here is the low-frequency part of the inflow turbulence (IT), followed by the BTI. The flow field in Perdigão causes additional large-scale variations in IT noise over time.

Synthesis of wind turbine trailing edge noise in free field.

The proposed sound synthesis tool converts a physics-based frequency-domain model of wind turbine trailing edge noise to a time-domain signal while accounting for the appropriate time shift due to the propagation between the moving blades and the fixed observer. A window function that implements cross-fading between consecutive signal grains is proposed and a method to objectively estimate the influence of the synthesis parameters is described. As the synthesis tool is independent of the aerodynamic noise model, it can be readily adapted to auralize other noise sources such as turbulent inflow noise or stall noise.

Low-Noise Synthetic Turbulence Tailored to Lateral Periodic Boundary Conditions

The present work is dedicated to turbulence synthesis tailored to lateral periodic boundary conditions for direct noise computations through compressible large eddy simulations. Synthetic turbulence can be essential for aeroacoustic applications when computing airfoil turbulent inflow noise or for accurately capturing the behavior of boundary layers. This behavior determines both trailing edge noise and complex flow structures such as laminar separation bubbles. For airfoil simulation purposes, spanwise periodic boundary conditions are usually considered. If synthetic perturbations are injected without observing the periodicity rule, strong spurious pressure waves are emitted and pollute the entire computational domain. In this work, the random Fourier modes method for turbulence generation is adapted in order to respect the spanwise periodicity constraint right at the computational domain inlet. This approach does not affect the turbulence properties such as the spectral shape and the turbulent kinetic energy decay. Since the emphasis is put on the generation and convection of the turbulence, only the turbulence convection region between the inlet and the airfoil is considered in this paper, without the airfoil. Two geometrical configurations are tested: the first one is a simple box with a constant mesh size, and the second one concentrates the fine cells on the area in front of the airfoil. In the second configuration, the computational cost is reduced by up to 25%, but more spurious noise is present because of interpolation areas between different grids using the Chimera method. Finally, the results’ reproducibility is assessed using different turbulence realizations.

A rapid distortion theory-based airfoil turbulent inflow noise prediction method

Turbulent inflow (TI) noise is reported as an important source of wind turbine (WT) broadband aeroacoustic noise, coexistent with many other sources (e.g., the airfoil self-noise). This manuscript intends to elucidate the discussion on TI noise by providing an alternative prediction method based on the rapidly distorted anisotropic turbulent velocity energy spectrum. A review on turbulent inflow noise prediction is presented, followed by the discussion on turbulence modeling and the derivation of a modified TI noise prediction expression, which allows a more detailed assessment of wind turbine noise sources. This can contribute for WT blade optimization, once it is implemented in design codes that consider noise as a constraint. Criteria for the turbulent velocity spectrum to be characterized by the rapid distortion theory (RDT) are established, and the validation methodology is carried through comparisons between the novel semi-empirical prediction method and data gathered from recent literature for a flat-plate airfoil with $$3\%$$ relative thickness and a NACA 0008 airfoil. The RDT-modified turbulent inflow noise prediction method shows better agreement with the measured data, when compared to predictions that adopt an isotropic model for the turbulent velocity energy spectrum, a condition only achieved in the absence of the airfoil in the test section of aeroacoustic facilities.

Comparison of inflow-turbulence and trailing-edge noise models with measurements of a 200-kW vertical axis wind turbine

Models of inflow-turbulence noise and turbulent-boundary-layer trailing-edge noise are compared to earlier measurements of a 200-kW vertical axis wind turbine so that conclusions regarding the origin of the aerodynamic noise can be drawn. The measurement campaigns, which aimed at establishing the noise emission value and locating the aerodynamic noise sources with a microphone array, are here both compared to further modified versions of the trailing-edge and inflow-turbulence models respectively. Unlike the case for horizontal axis wind turbine, inflow-turbulence noise is deemed as the prevailing noise mechanism. Reducing the self-induced turbulence could then be an effective way of lowering the noise levels for vertical axis wind turbines. Also, looking at the directivity of the inflow-turbulence noise model which indicate most noise in the cross-wind directions, a deviation from the standard downwind measurement position for measuring noise emission is suggested for the VAWT case.

Noise emission from wind turbines in wake - Measurement and modeling

The influence of the wake of an upstream turbine impinging another one located further downstream is studied focusing on the latter’s noise emission. Measurement data are investigated in the form of surface pressure fluctuations acquired using microphones flush-mounted in a wind turbine blade near its tip, characterizing the noise sources. Numerical results from a wind turbine noise model are also included in the analysis. The wind speed deficit and increased turbulence levels of the wake flow are clearly observed. Surface pressure measurements strongly support the fact that turbulent inflow noise is increased. However, numerical results show that the wake velocity deficit reduces noise in certain circumtances. This can compensate, or even sometime more than compensate, the additional noise emission expected as a result of the wake turbulence. Furthermore, noise amplitude modulation appears to increase when the turbine is impacted by the wake flow.

Comparison of some wind turbine noise emission models coupled to BEM aerodynamics

Noise is an important design driver for onshore wind turbines. Therefore, there is a need to include validated noise prediction models within wind turbine and wind plant design procedures. However, the literature presents a multitude of different models and formulations, combined with a severe lack of publicly available experimental data. The present work presents a comparison between semi-empirical frequency-domain methods coupled to blade element momentum (BEM) aerodynamics, which is the typical approach used for the transient aeroelastic analysis of wind turbines. Among the various noise sources, only turbulent boundary layer – trailing edge and turbulent inflow noise models have been considered in the present analysis. The noise models are benchmarked in terms of emission spectra, highlighting differences and discrepancies. A time-domain Ffowcs Williams-Hawkings formulation is also coupled to the same BEM model, investigating its ability in predicting low frequency emissions. Thanks to this study, a few inconsistencies among published noise models are identified and reported.

Coupling of an aeroacoustic model and a parabolic equation code for long range wind turbine noise propagation

This study proposes to couple a source model based on Amiet's theory and a parabolic equation code in order to model wind turbine noise emission and propagation in an inhomogeneous atmosphere. Two broadband noise generation mechanisms are considered, namely trailing edge noise and turbulent inflow noise. The effects of wind shear and atmospheric turbulence are taken into account using the Monin-Obukhov similarity theory. The coupling approach, based on the backpropagation method to preserve the directivity of the aeroacoustic sources, is validated by comparison with an analytical solution for the propagation over a finite impedance ground in a homogeneous atmosphere. The influence of refraction effects is then analyzed for different directions of propagation. The spectrum modification related to the ground effect and the presence of a shadow zone for upwind receivers are emphasized. The validity of the point source approximation that is often used in wind turbine noise propagation models is finally assessed. This approximation exaggerates the interference dips in the spectra, and is not able to correctly predict the amplitude modulation.

Experimental Validation of a Wind Turbine Turbulent Inflow Noise Prediction Code

This paper describes the comparison of a semi-empirical wind turbine noise prediction code to a large volume of acoustic data gathered on a full-scale 2.3 MW wind turbine. Particular focus is placed on the turbulent inflow noise model due to the relatively small amount of validation that has been performed previously. Two novel techniques are developed in order to effectively validate the model. The first is that the turbulence is characterized in the acoustic model using a single quantity, the turbulence dissipation rate, instead of the combination of turbulence intensity and integral length scale as has been used in previous studies. The second development is a method of measuring the turbulence dissipation rate using blade-mounted accelerometers such that the turbulence dissipation rate in the vicinity of the blade is accurately determined. Using these methods, it is shown that the flat-plate airfoil analytical turbulence noise model underpredicts the turbulent inflow noise levels by 3–5 dB, whereas semi-empirical corrections for finite-thickness airfoils reduce prediction errors to less than 3 dB. Comparisons are also made on the basis of scaling with blade tip-speed, scaling with turbulence dissipation rate, and noise directivity.

An implementation of an aeroacoustic prediction model for broadband noise from a vertical axis wind turbine using a CFD informed methodology

This paper presents an enhanced method for predicting aerodynamically generated broadband noise produced by a Vertical Axis Wind Turbine (VAWT). The method improves on existing work for VAWT noise prediction and incorporates recently developed airfoil noise prediction models. Inflow-turbulence and airfoil self-noise mechanisms are both considered. Airfoil noise predictions are dependent on aerodynamic input data and time dependent Computational Fluid Dynamics (CFD) calculations are carried out to solve for the aerodynamic solution. Analytical flow methods are also benchmarked against the CFD informed noise prediction results to quantify errors in the former approach. Comparisons to experimental noise measurements for an existing turbine are encouraging. A parameter study is performed and shows the sensitivity of overall noise levels to changes in inflow velocity and inflow turbulence. Noise sources are characterised and the location and mechanism of the primary sources is determined, inflow-turbulence noise is seen to be the dominant source. The use of CFD calculations is seen to improve the accuracy of noise predictions when compared to the analytic flow solution as well as showing that, for inflow-turbulence noise sources, blade generated turbulence dominates the atmospheric inflow turbulence.

Location of aerodynamic noise sources from a 200 kW vertical-axis wind turbine

Noise levels emitted from a 200kW H-rotor vertical-axis wind turbine have been measured using a microphone array at four different positions, each at a hub-height distance from the tower. The microphone array, comprising 48 microphones in a spiral pattern, allows for directional mapping of the noise sources in the range of 500Hz to 4kHz. The produced images indicate that most of the noise is generated in a narrow azimuth-angle range, compatible with the location where increased turbulence is known to be present in the flow, as a result of the previous passage of a blade and its support arms. It is also shown that a semi-empirical model for inflow-turbulence noise seems to produce noise levels of the correct order of magnitude, based on the amount of turbulence that could be expected from power extraction considerations.

Experimental characterization of turbulent inflow noise on a full-scale wind turbine

An extensive experimental campaign was conducted on a 108-m diameter 2.3-MW wind turbine in order to assess the effect of inflow turbulence conditions on wind turbine acoustics. Over 50h of continuous acoustic data was acquired at power-generating wind speeds. Twelve precision microphones were used, arranged in a one rotor radius ring about the turbine tower in order to assess the directivity of the noise emission. Turbine operational and atmospheric conditions were gathered simultaneously with acoustics measurements. The testing and analysis constitute perhaps the most thorough experimental characterization of turbulent inflow noise from a wind turbine to date. Turbulence intensities typically varied between 10 percent and 35 percent, and wind speeds covered most of the operational range of the wind turbine, from cut-on to well above its rated power. A method was developed for using blade-mounted accelerometers for determining the turbulence conditions in the immediate vicinity of the blades, which are the primary turbulence noise generating bodies. The method uses the blades׳ vibrational energy within a specified frequency range to estimate the overall turbulence conditions by assuming a von Kármán turbulence spectrum. Using this method, a clear positive correlation is shown between turbulence intensity and noise levels. The turbulence noise is dominant at low frequencies and is primarily observed in the upwind and downwind directions. Low frequency noise increases by as much as 6dB for the range of turbulence conditions measured. Comparisons are made between the measured turbine noise directivity and theory using a simple acoustic model of the turbine as three point-sources. Strong agreement is found between the theoretical leading edge noise directivity model and the measured low frequency noise directivity.

Aerodynamic Noise Characterization of a Full-Scale Wind Turbine through High-Frequency Surface Pressure Measurements

The aim of this work is to investigate and characterize the high-frequency surface pressure fluctuations on a full-scale wind turbine blade and in particular the influence of the atmospheric turbulence. As these fluctuations are highly correlated to the sources of both turbulent inflow noise and trailing edge noise, recognized to be the two main sources of noise from wind turbines, this work contributes to a more detailed insight into noise from wind turbines. The study comprises analysis and interpretation of measurement data that were acquired during an experimental campaign involving a 2 MW wind turbine with a 80 m diameter rotor as well as measurements of an airfoil section tested in a wind tunnel. The turbine was extensively equipped in order to monitor the local inflow onto the rotating blades. Further a section of the 38 m long blade was instrumented with 50 microphones flush-mounted relative to the blade surface. The measurements of surface pressure spectra are compared with the results of two engineering models for trailing edge noise and for turbulent inflow noise. The measured pressure fluctuations are related to the local inflow angle and are also compared to measurements in a wind tunnel on a copy of the blade section of the full scale blade. Computational Fluid Dynamics calculations were conducted to investigate the influence of the inflow conditions on the airfoil and blade sections aerodynamics and aeroacoustics. Comparisons between measurement data and model results show the influence of atmospheric turbulence. The different noise generation mechanisms can be identified and the influence of various parameters can be consistently reproduced by the models.

Adaptive neuro-fuzzy methodology for noise assessment of wind turbine.

Wind turbine noise is one of the major obstacles for the widespread use of wind energy. Noise tone can greatly increase the annoyance factor and the negative impact on human health. Noise annoyance caused by wind turbines has become an emerging problem in recent years, due to the rapid increase in number of wind turbines, triggered by sustainable energy goals set forward at the national and international level. Up to now, not all aspects of the generation, propagation and perception of wind turbine noise are well understood. For a modern large wind turbine, aerodynamic noise from the blades is generally considered to be the dominant noise source, provided that mechanical noise is adequately eliminated. The sources of aerodynamic noise can be divided into tonal noise, inflow turbulence noise, and airfoil self-noise. Many analytical and experimental acoustical studies performed the wind turbines. Since the wind turbine noise level analyzing by numerical methods or computational fluid dynamics (CFD) could be very challenging and time consuming, soft computing techniques are preferred. To estimate noise level of wind turbine, this paper constructed a process which simulates the wind turbine noise levels in regard to wind speed and sound frequency with adaptive neuro-fuzzy inference system (ANFIS). This intelligent estimator is implemented using Matlab/Simulink and the performances are investigated. The simulation results presented in this paper show the effectiveness of the developed method.

A discussion of wind turbine interaction and stall contributions to wind farm noise

Wind farms have recently been reported to produce a noise signature that is described as possessing a “thumping” quality. Measurements of these signatures are limited and their effects are debated but their effect on public opinion and complaints make them a concern for researchers in this field. Proposed reasons for these noise signatures include amplitude modulation, interference patterns and wake–rotor interaction. This paper discusses these effects and concludes that wake–rotor interaction plays a role by causing variations in turbulent-inflow noise and dynamic stall. The current state of research into stall noise and wind turbine wake structure is also reviewed and it is concluded that the available information and collected data on wind turbine wake are insufficient to determine how strong this role is. More information on the velocity and turbulence fields in the wake of horizontal-axis wind turbines as well as a characterisation of the noise produced by an airfoil experiencing dynamic stall is required in order to make a full assessment of rotor–wake contributions to wind farm noise.

Time- and frequency-domain computations of broadband noise due to interaction between incident turbulence and rectilinear cascade of flat plates

Time-domain computational aeroacoustic (CAA) techniques are developed to investigate the broadband noise resulting from the interaction of a rectilinear cascade of flat plates with incident homogeneous, isotropic turbulence. The investigation is carried out by comparing the prediction results obtained by employing the time-domain CAA method with those using existing frequency-domain methods. A semi-analytic model (Wei & Cheong, 2010) and a full three-dimensional rectilinear cascade model (Lloyd & Peake, 2008; Lloyd, 2009) are adopted for the frequency-domain computations. By comparing these computation results, the three-dimensional characteristics of inflow turbulence noise are investigated; in particular, the effects of the wavenumber components of ingested turbulence in the spanwise direction are taken into consideration in the investigation. First, CAA results are compared with those from the semi-analytic model. The results for the acoustic modes of relatively low spanwise wavenumbers obtained using both methods show good agreement, but as the spanwise wavenumber increases, the results obtained by the two methods become increasingly different. To investigate in detail the reason for these differences, mode-decomposition analysis is performed by adopting a hybrid method as well as by employing the CAA and the semi-analytic method. The hybrid method involves the following two sequential computations: (i) the upwash velocities on the flat plate airfoils of the rectilinear cascade are first predicted using the frequency-domain method, and (ii) the acoustic wave propagation is subsequently analyzed using time-domain CAA techniques, with these upwash velocities applied as the boundary conditions on the flat plate. It is seen that the results of the time-domain CAA technique and the hybrid method show good agreement, irrespective of the wavenumber and frequency. However, comparisons of the acoustic solutions from three computations reveal that the prediction results of the semi-analytic model deviate more from the other two predictions as the spanwise wavenumber of the acoustic wave increases and the frequency decreases. On a basis of this observation, a formulation is derived for the error in the pressure jump across the flat-plate predicted by using the semi-analytic method. This formulation shows that the error is approximately inversely proportional to the sound speed in the spanwise direction of the concerned acoustic modes. This result quantitatively clarifies the limitations of applying the frequency-domain method of Wei & Cheong (2010) to the three-dimensional turbulence-cascade interaction problems. Secondly, the prediction results using the time-domain CAA method are compared with those from the full three-dimensional rectilinear model that is believed to be exact model for the cascade geometry considered in this paper. This comparison shows the good agreements between two predictions, which support the above arguments for the error and the successful application of the time-domain CAA methods. It is expected that these methods can be extended to the broadband noise problem in an annular cascade, including the nonlinear interaction of the real-airfoil cascade with the incident nonhomogeneous gust.