Optimization of the structure-borne noise analysis methods with the help of virtual sensing with numerical and experimental results integration

Vibro-acoustic comfort has become an increasingly integral part of vehicle development over the past 20 years. The perception inside the cabin has a high impact on customer satisfaction and expresses the quality of premium vehicles. The phenomenon of interior noise has always been very complex to study. The problem is complicated by the fact that there is a well-known variability in products. This is due both to variability in the assembly line and in the properties of individual components. The goal is for as many vehicles as possible to meet the design requirements. Among the various categories of noise, low-frequency noise has deserved special attention from development engineers over the years. It is, in fact, particularly annoying, and can also cause loss of attention and drowsiness while driving. In the category of low-frequency noise, booming is one of those phenomena that one surely wants to avoid while developing a vehicle. This phenomenon is used in this work as a target for the analysis. Almost all components, which directly, or via isolating elements, are mounted to the vehicle chassis, can generate booming under certain circumstances. This is the case, for example, of the powertrain. In conventional vehicles powered by an internal combustion engine, the powertrain is bound to the chassis with a system of elastomeric mounts. The forces generated by the engine, during operation, are transferred through the mounts to the body, and here generate structural noise, which is perceived with discomfort by passengers. Using isolating elements to decouple components is a very common practice. Yet the dynamic properties of the mounts are hardly known with sufficient accuracy. Therefore there is always the need to make measurements and to check if the used mounts meet the requirements. It is also well known that the characteristics of these components depend on various parameters. The most important of these are static preload, frequency and amplitude of excitation. Hardly all three dependencies are measured and simulated simultaneously. Mounts present, moreover, a marked variability of their characteristics, from which it follows a variability of vehicle's acoustic quality. In the first part of this thesis, a methodology is presented for the characterization of powertrain mounts, under real operating conditions, which allows to experimentally measure and model the three dependencies described above. Starting from measurements on a dedicated test bench, using response surfaces, and a virtual point transformation methodology, parametric models of the dynamic stiffnesses of the mounts are generated. With these, considering a representative pool of vehicles, a robust optimization of the powertrain mounts is proposed, aiming at reducing the booming noise in a population of vehicles. However, not only conventional vehicles are affected by this phenomenon. In electric vehicles, the endothermic engine is replaced by a quieter electric motor. In doing so, other components, whose noise was previously masked by the powertrain, become paramount. One such component is the air-conditioning compressor, which has acquired a new function in electric vehicles, that of cooling the battery module during charging. When in full load operation, it can generate low frequency booming noise. In the second part of this thesis, this effect is analyzed, using inverse methodologies, coupled with virtual point transformation techniques. The analysis includes, in this case, the identification of the characteristics of the component, identifying its internal forces and moments acting during operation. An analytical modeling of the component mounted in the vehicle is proposed and validated through experimental measurements. The dynamic stiffnesses of the mounts are modeled through parametric functions, whose parameters are optimized through "in-situ" experimental measurements.