Case Study No: FTLCFD

Introduction

Low-pressure axial fans are widely utilized in various commercial applications, with design optimizations targeting both aerodynamic performance and acoustic characteristics. Aeroacoustic noise arises from several significant sources, such as turbulent inflow, self-generated secondary flows at the blade edges (resulting from turbulent or laminar boundary layers, boundary layer separation), trailing edge noise, and noise originating from the interaction between the rotor and stator. Aerodynamic noise, identified as the major source of noise in axial fans[1], is a prevalent issue, and its accurate prediction and understanding are crucial for optimizing fan performance while reducing acoustic emissions.

Figure 1: Axial Fan used for aeroacoustic modelling

Both broadband and discrete noise exist in axial fans, though the former makes a more significant contribution to the overall weighted sound power level in low-speed axial fans. Though numerical prediction of acoustics using Computational Fluid Dynamics (CFD) is computationally expensive, there is an emerging interest in this approach. Numerical approaches for noise prediction are categorized as direct (Computational Aero Acoustics(CAA)) and hybrid methods (Ffowcs Williams and Hawkings (FW-H) and other analogous methods), with the direct method being more accurate in noise prediction, but prohibited by high computational cost, mostly limiting its applications to idealized problems. A more detailed comparison of the aeroacoustics modelling approaches is discussed in the full-length paper here.

Computational Domain and Mesh

We used benchmark data from the European Acoustics Association for the study.[2]. The computational model used for CFD modelling of aeroacoustics is illustrated in Figure 2. The duct nozzle is incorporated into the stationary inlet domain, sized at 2.32 m × 2.4 m × 2.4 m, comparable to the test chamber size. The rotating domain, inclusive of the fan and the straight duct segment, is connected by non-conforming interfaces on the pressure and suction sides to the inlet and outlet domains. The exit domain has a length of 4𝐷 (where 𝐷 is the duct diameter) and is scaled at 2.0 m×2.4 m×2.4 m, matching the inlet cross-section.

Figure 1. The modelled computational domain with the axial fan and microphone positions for noise prediction

Mesh sizing plays a crucial role in accurately predicting aeroacoustic phenomena. The mesh size is selected according to the guidelines for aeroacoustic modeling, considering the required frequency resolution and range. A sliding mesh approach, with non-conformal interfaces, was employed to capture the unsteady aerodynamic phenomena in the rotating domain, crucial for acoustic noise prediction. The computational domain is composed of both rotating and stationary regions, as shown in Table 2. Figure 2 illustrates the mesh of the full computational domain for acoustic analysis. To ensure high-fidelity simulations, fifteen prism layers were generated as near-wall cells of the axial fan solid surfaces, maintaining an average 𝑦+ value of approximately 1 on the blades. A refined unstructured poly-hexcore mesh was applied in the downstream region to accurately capture the tip vortex and other small-scale flow structures.

Figure 2 : The computational mesh for the domain, consisting of 22 million poly-hexcore cells.

Numerical Model

The 3D steady-state, incompressible, single-phase analysis was performed using ANSYS Fluent. We first simulated the rotating propeller using the MRF approach as a frozen rotor, followed by the sliding mesh approach for improved accuracy.

For the numerical setup, second-order discretization schemes were applied to the continuity and momentum equations. A convergence criterion of 1 × 10⁻⁶ was set for continuity, momentum, and turbulence equations. The k-ω SST model with a wall function approach was used to obtain the initial steady-state solution. Subsequently, a high-fidelity Detached Eddy Simulation (DES) was conducted, using the steady-state results for initialization.

Parameter	Value
Fluid properties	Air (Ideal Gas), non – buoyant
Solution type	Transient
Turbulence model	k-ω SST and DES
Timestep size, Δ𝑡	1 × 10-5 s

Table 1: Simulation settings

Boundary conditions
Fan rotational speed	1486 rpm
Inlet	Velocity inlet
Outlet	Pressure outlet
Walls	No Slip Wall
Reference pressure	0 atm

Table 2: Boundary conditions

Results

Figure 3 shows the Q-criterion, which depicts the turbulent structures predicted by the URANS and DES models, since these turbulence structures are a major source of aeroacoustic noise. The resolved velocity field on the midplane (DES turbulence model) is shown in Figure 4, with the acoustic response at the mic M7 using both the acoustic modelling approaches – CAA and FW-H. 

Figure 3: Q-Criterion showing the turbulent structures of the tip flow: (A) URANS (B) DES

Figure 5: (A) Midplane view of the resolved velocity field using DES model. (B)Power Spectral Density at microphone M₇ using CAA and FWH approaches.

Conclusion

• Aeroacoustic modelling of an axial fan in a short duct using CFD was investigated.
• The aeroacoustic modelling approach was benchmarked with the axial fan benchmark data by the European Acoustics Association (EAA).
• A comparison of the effectiveness of the URANS
• The DES approach provided insights into unsteady flow characteristics, enhancing the accuracy of the aerodynamic analysis.

* A more extended version of this work was presented by us at the ASME Turbo Expo 2025 in Memphis, TN. For more details, please refer to the publication here: GT2025-153849, V011T33A008.

References

1. Sharland, I Jr. “Sources of noise in axial flow fans.” Journal of Sound and Vibration Vol. 1 No. 3 (1964): pp. 302–322.
2. Schoder, Stefan and Czwielong, Felix. “Dataset fan-01: Revisiting the eaa benchmark for a low-pressure axial fan.” arXiv preprint arXiv:2211.12014 (2022).
3. Schoder, Stefan, Junger, Clemens and Kaltenbacher, Manfred. “Computational aeroacoustics of the EAA benchmark case of an axial fan.” Acta Acustica Vol. 4 No. 5 (2020):p. 22.
   

Explore our Case Studies and Publications

Gain insights from real-world CFD case studies demonstrating our technical expertise, problem-solving approach, and industry-driven solutions. Explore the full portfolio now.

Case Studies Our Publications