3D-CFD Simulation methodologies for the assessment of HVAC aeroacoustics performance

With the advent of electrification as a continuously growing trend in automotive industry, unprecedented changes are shaping the vehicle evolution. The new concepts at the basis of vehicle design not only follow the transition towards a sustainable mobility, but lead also to major changes in user experience and comfort inside the vehicle. The electrification changed completely the spectrum of noise emissions from vehicles and tractors, cutting-off all the engine-related noises. The noise reduction presents a unique opportunity to delve deeper into the study of aeroacoustics phenomena, to better understand their impact on the cabin occupants and the driving or working experience. With a major noise source switched off, other noises unveil themselves, as for example the ones coming from the Heating, Ventilation, and Air Conditioning (HVAC) system, and could cause significant disturbances inside the cabin. Especially in tractors and industrial vehicles, it is crucial to provide a high comfort level. Computer Aided Engineering (CAE) becomes crucial in the investigation of aeroacoustics phenomena during vehicle development, when physical prototypes are not yet available but possible noise sources and countermeasures can be identified through 3D-CFD analysis, which in this project is performed with Simcenter Star-CCM+. The aim of this thesis, which was carried out in GammaTech Engineering in collaboration with CNH-Industrial, is to investigate proper methodologies to predict noise generation and propagation in HVAC ducts, exploring a wide range of possibilities, from direct noise computation to hybrid methods that are capable of decoupling aerodynamics and acoustics. More in detail, a simplified HVAC duct geometry, for which experimental data were available from literature, was analyzed with different simulation strategies. At first, the Direct Noise Computation (DNC) method was adopted to assess the capability of the 3D-CFD methodology to correctly reproduce experimental data, using the most detailed numerical approach available. In this context, the DNC led to results in good agreement with experimental evidences. Then, the project shifted the focus on the adoption of hybrid methods, with the aim to describe the aeroacoustics behaviour in the far-field, and obtaining a significant reduction of simulation time and computational effort. Both the hybrid methods, based on Lighthill model and Perturbed Convective Wave model, provided an important speed up during simulations, and led to results showing good agreement with experimental ones in terms of Sound Pressure Level (dB) in the far-field.