Load-oriented aerodynamic design approach of low-speed axial fans in the preliminary design stage

Proceedings of the Institution of Mechanical Engineers Part G Journal of Aerospace Engineering, Год журнала: 2025, Номер unknown

Опубликована: Янв. 21, 2025

For low-speed axial-flow fans, the arbitrary vortex design method is widely used in high total pressure rise requirement or small hub-to-tip ratio applications. In conventional method, of fan blade commonly increases from hub to tip due greater work potential near tip. However, aerodynamic load region will inevitably result an intensification leakage flow which limits improvement efficiency, especially for fans with a requirement. To solve this problem, paper has developed load-oriented approach take into account spanwise distribution preliminary stage so as reduce loss caused by overloading airfoil sections local fractions and enhance performances. specific, firstly parameterized control along direction. Then, diffusion factor correlated coefficient achieve rapid evaluation performance. meanwhile, can be automatically optimized optimal distribution. Finally, two are designed based on respectively. The computational results show that achieved lower compared fan. At point, approximately 1.5% increment adiabatic efficiency also obtained. Moreover, over entire operating range, shows higher than fan, without sacrificing rise. motivation offer useful guidelines high-performance axial under requirements

Язык: Английский

---

Finely integrated design of end wall region for a compressor linear cascade through adjoint optimization

Physics of Fluids, Год журнала: 2025, Номер 37(1)

Опубликована: Янв. 1, 2025

With the advancement of aerodynamic design techniques for turbomachinery, high-loss flow in end wall region has emerged as a bottleneck, limiting further improvements performance fans and compressors. In this study, linear cascade is flexibly parameterized using advanced extended free-form deformation method subsequently optimized through adjoint optimization, enabling an integrated entire region. A refined featuring flexible hub smoothly corner achieved. The results demonstrated that finely significantly reduced loss coefficient by 16.81% effectively controlled separation.

Язык: Английский

---

Surrogate Modeling of Hydrogen-Enriched Combustion Using Autoencoder-Based Dimensionality Reduction

Processes, Год журнала: 2025, Номер 13(4), С. 1093 - 1093

Опубликована: Апрель 5, 2025

Deep learning-based surrogate models have received wide attention for efficient and cost-effective predictions of fluid flows combustion, while their hyperparameter settings often lack generalizable guidelines. This study examines two different types models, convolutional autoencoder (CAE)-based reduced order (ROMs) fully connected (FCAE)-based ROMs, emulating hydrogen-enriched combustion from a triple-coaxial nozzle jet. The performances these ROMs are discussed in detail, with an emphasis on key hyperparameters, including the number network layers encoder (l), latent vector dimensionality (dim), stride (s). results indicate that larger l is essential capturing features strongly nonlinear flowfields, whereas smaller more effective less distributions, as additional may cause overfitting. Specifically, when employing CAE-based to predict spatial distribution H2 (XH2) weak nonlinearity, reconstruction absolute average relative deviation (AARD) two-layer model was marginally higher than three- four-layer prediction AARD approximately 5% lower. A dim yields better performance weakly flowfields but increase local errors some cases due excessive feature compression. ROM = 10 achieved notably lower 4.01% XH2 prediction. s enhance resolution yet raise computational costs. Under identical outperformed FCAE-based both cost-effectiveness accuracy, achieving 35 times faster training speed These findings provide important guidelines selection (AE)-based other similar engineering design problems.

Язык: Английский

---

Numerical Prediction of Fluid Flow Noise with a Hybrid CFD-Acoustics Approach

SAE technical papers on CD-ROM/SAE technical paper series, Год журнала: 2025, Номер 1

Опубликована: Май 5, 2025

<div class="section abstract"><div class="htmlview paragraph">This paper presents a fully parallelized Computational Acoustics (CA) module, integrated within the Simerics-MP+ platform, developed for prediction of noise source power and far-field propagation across range Fluid Dynamics (CFD) applications. Utilizing Ffowcs Williams-Hawkings (FWH) acoustic analogy, CA module seamlessly integrates with existing CFD workflows, offering minimal computational overhead less than 5% increase in runtime. Extensive validation has been conducted against analytical, numerical, experimental data various scenarios, including monopole dipole emissions, flow around slender bodies, circular cylinders aero-propellers. These studies underscore reliability framework accurately identifying sources assessing impact design modifications, significantly reducing need expensive physical prototyping industries such as automotive aerospace.</div><div paragraph">Building upon these foundational capabilities, study extends application to challenging task simulation axial fans, commonly employed HVAC systems, components, industrial processes. Given proximity fans human operators, control is critical, necessitating compliance stringent regulatory standards. In low-Mach-number aerodynamic fluctuations complex interactions are key contributors emissions. By combining Wall-Modeled Large Eddy Simulation (LES) FWH approach second-order finite volume simulates European Association (EAA) benchmark case ducted fan rates. Validation shows good agreement, demonstrating effectiveness this facilitating virtual testing. This methodology enables engineers implement mitigation strategies early process, development costs enhancing overall system performance.</div></div>

Язык: Английский

---

An Efficient Cooling Design Strategy for the Embedded Driving Motor under the Hub by Using the High-Pressure Airflow after the Rotor in an Electrical Ducted Fan System

Aerospace Science and Technology, Год журнала: 2024, Номер unknown, С. 109893 - 109893

Опубликована: Дек. 1, 2024

Язык: Английский

---