Numerical methodologies for noise and vibration sources assessment in Powertrain applications

The Aim of this project is to define numerical methodologies for noise sources assessment, validate them and analyse their limits of validity. In this project, dynamic substructuring for Numerical Transfer Path Analysis (NTPA) and Surface Contribution Analysis (SCA) within Wave Based Method (WBM) were developed. A validation on a particular case was also performed for the NTPA method. In addition, the robustness of these methodologies was investigated. Moreover, an newly energy based method was developed: Energy Flow Contribution Analysis (EFCA). EFCA is a methodology which can be also investigated in an experimental setup. Therefore, in the last part of the project a series of experimental measurements will be held in Milano. NTPA is based on dynamic substructuring of multi-body dynamic systems. The body under investigation is isolated and the body joint contacts replaced with contact forces. Each excitation is then treated as a separate source and its response is computed. SCA is a methodology thought to operate within a wave based method framework. It is based on the decomposition of the boundary conditions of the acoustic radiation problem, under the assumption of the validity of the superposition principle on the boundary conditions (structural vibrations can be treated with linear models). Different surfaces of the radiating object consider different boundary conditions and with this procedure the noise radiated from each different surface is separately investigated. EFCA is similar to NTPA analysis but it directly operates with power quantities. At the interfaces of the investigated body, the transmitted power is evaluated giving direct information of the interaction of the different noise sources. Concerning the problem of sources assessment and ranking, no experimental methodologies work with power quantities. Hence, similar analysis can be done also on real systems via direct measurements of interface transmitted power. Finally, this projects deals also with the estimation of the robustness of the proposed methodologies. Therefore, the influence of system uncertainties is investigated. In particular the effect of cyclic cylinder variability (uncertainties on excitation input) and the uncertainties on model parameters (applying random matrix theory) are considered. In these investigations it was shown the frequency range of validity and how to correctly perform multibody dynamic analysis or acoustic radiation analysis in order to derive useful information on the source assessment and ranking.