Research on flow-sound separation algorithm of aerodynamic noise based on immersed boundary method

Zhi-jun Meng, Wei He, Dao-chun LiJournal of Vibroengineering 2015

TL;DR

This paper presents a novel hybrid numerical method for predicting aerodynamic noise from complex shapes at low Mach numbers. By separating the flow and sound fields and employing an immersed boundary method, the approach effectively handles intricate geometries and demonstrates good agreement with experimental data for tandem cylinders and a rudimentary landing gear.

What Problem Does This Paper Address?

Predicting aerodynamic noise, especially from complex structures like aircraft landing gear, is a significant challenge. Traditional computational aero-acoustics (CAA) methods struggle with these intricate geometries, often requiring computationally expensive grid generation or facing limitations with standard grid-based techniques. Existing methods for handling complex shapes, such as block-structured or overlapping grids, have drawbacks, and while the immersed boundary method (IBM) offers a universal solution, its application to aeroacoustics, particularly for low Mach number flows, requires careful integration with sound propagation models. This paper aims to bridge this gap by developing and validating a flow-sound separation algorithm based on IBM for low Mach number aeroacoustic problems with complex geometries.

Methods

The core of this research is a hybrid INS/LPCE (Incompressible Navier-Stokes/Linearized Compressible Perturbation Equation) method. This approach decomposes the flow variables into incompressible and compressible perturbation parts.

  • Flow Field Calculation: The incompressible Navier-Stokes (INS) equations are solved to capture the flow field dynamics. Crucially, an Immersed Boundary Method (IBM) is employed here. This allows the use of a Cartesian grid for the entire computational domain, simplifying grid generation for complex geometries. The IBM handles the solid boundaries by introducing "image points" and "body intercept points" to enforce boundary conditions, effectively treating the solid object as "immersed" within the fluid domain. A higher-order polynomial interpolation based on weighted Least Squares Error is used to determine values at these boundary points.
  • Sound Field Calculation: Once the incompressible flow field is obtained, the Linearized Compressible Perturbation Equation (LPCE) is solved to simulate the generation and propagation of acoustic waves. This separation of flow and sound calculations is a key aspect of the "flow-sound separation" algorithm.
  • Numerical Discretization:
    • Spatial derivatives for the INS equations are handled by a second-order central difference scheme.
  • LPCE spatial discretization uses a sixth-order compact central difference.
  • Time integration for both is performed using a fourth-order Runge-Kutta method.
  • Specialized boundary treatments (three-order and four-order) are used near the immersed solid boundary.
  • Implicit spatial filtering is employed to suppress high-frequency errors.

Datasets

The paper does not introduce new datasets. Instead, it relies on experimental data for validation:

  • Tandem Circular Cylinders: Results are compared with experimental data from NASA and other research groups for a similar configuration.
  • Rudimentary Landing Gear: The numerical results are compared with experimental data obtained from a wind tunnel test specifically conducted for this study. The experimental setup involved a rudimentary landing gear model, microphones, and a pulse test system.

Key Insights and Ideas

  • IBM for Aeroacoustics: The successful application of the Immersed Boundary Method to aeroacoustic problems with complex geometries is a significant contribution. It bypasses the need for complex body-fitted grids, simplifying the pre-processing stage.
  • Flow-Sound Separation Efficiency: The hybrid INS/LPCE approach effectively separates the flow and acoustic phenomena, allowing for efficient computation of aerodynamic noise at low Mach numbers. This is particularly beneficial for airframe noise where Mach numbers are typically low.
  • Handling Complex Geometries: The method's ability to accurately predict noise from intricate shapes like a landing gear, which would be challenging for traditional grid-based methods, highlights its practical applicability.
  • Experimental Validation: The consistent agreement between numerical simulations and experimental results for both the tandem cylinders and the landing gear provides strong evidence for the method's reliability and feasibility.
  • Turbulence and Noise Interaction: The study demonstrates the method's capability to capture the interaction between turbulent flow structures and sound generation, as seen in the landing gear case where chaotic flow fields correlate with higher sound pressure levels.

Key Results

  • Tandem Cylinders: The simulation of turbulent flow noise around two tandem circular cylinders showed good agreement with experimental data. The main vortex shedding frequency was found to be St = 0.196, which was close to other numerical and experimental results. The aerodynamic force fluctuations were more pronounced on the downstream cylinder due to vortex interference.
  • Rudimentary Landing Gear: The method was applied to a rudimentary landing gear model at ReD = 2000 and M = 0.3. The computed sound pressure levels at three observation points were compared with experimental data. The results showed that the sound pressure levels of observation point 2 were significantly higher, correlating with a more chaotic flow field at that location. The overall trends between the numerical simulation and experimental data were found to be consistent, with sound pressure levels generally decreasing with increasing frequency.

Limitations and Future Work

  • Span-wise Length Modification: For the tandem cylinder case, the initial span-wise length (Lz = 3D) was insufficient for a full three-dimensional sound field. The results were modified using Oberai's formula to account for a larger experimental span-wise length (16D). This suggests that while the method can handle complex geometries, achieving accurate 3D sound fields might still require careful consideration of domain size or advanced modification techniques.
  • Computational Accuracy: The paper acknowledges that the "further problem" is the computational accuracy of the method. While verified against experiments, further investigation into higher-order accuracy and error analysis might be beneficial.
  • Low Mach Number Focus: The method is specifically designed for low Mach number flows. Its performance at higher Mach numbers, where compressibility effects are more dominant, is not discussed.
  • Future Work: The authors suggest that the method is useful for predicting noise from various low subsonic airframes, fan noise, and electronic equipment. Further research could explore its application to a wider range of complex aerodynamic noise problems and potentially investigate higher Mach number regimes.

Summary updated

Abstract

With the development of theoretical models, numerical algorithms and available computational power, numerical aero-acoustics presents a huge prospect for solving the actual problems of aerodynamic noises. Under low Mach number, structures with a complicated geometric shapes were selected as the research objects to analyze their aerodynamic noises by means of flow-sound separation algorithm based on immersed boundary method. Firstly, the incompressible complicated flow field was solved on the basis of immersed boundary method, and the flow field parameters were obtained as the input values. Then, the linear compressible perturbation equation was solved to simulate generation and travel of acoustic waves. This method was used to predict the noise of flow past two circular cylinders in tandem arrangements and the aerodynamic noise of the rudimentary landing gear. The numerical simulation was then compared with the corresponding experiments, and results were consistent. That showed the algorithm proposed was feasible.
BibTeX
@article{meng2015research,
  author = {Zhi-jun Meng and Wei He and Dao-chun Li},
  title = {{Research on flow-sound separation algorithm of aerodynamic noise based on immersed boundary method}},
  journal = {Journal of Vibroengineering},
  year = {2015},
}
Figure