S. Donders, Computer-aided engineering methodologies for robust automotive NVH design, 2008

Abstract

This dissertation deals with Computer-Aided Engineering (CAE) simulation methods to predict the vibro-acoustic behavior of vehicles, more specifically in automotive industry. The Finite Element Method (FEM) is often used for numerical design validations in the low-frequency range. For accurate predictions up to higher frequencies, the element mesh must be further refined, which requires a prohibitively large computational effort. Moreover, the effect of uncertainty and variability on the vehicle behavior increases with the frequency. To predict these effects, typically a high number of variants of the nominal vehicle must be analyzed. Finally, there is an industrial need to have simulation results already in the early concept phase, before the availability of the initial Computer-Aided Design (CAD) model. Conventional FEM models can only be created on the basis of this initial CAD model.

The above challenges are the framework of the research in this dissertation. Among others, the effect of uncertainty on the frequency response characteristics of structures has been investigated, which led to a novel “Short Transformation Method” (STM) for efficient response predictions for the class of structures with a monotonic influence of parameter uncertainty. In addition, the robustness of the dynamic system behavior to spot weld failure has been investigated. The strive for more efficient design variant analysis led to a novel “Wave-Based Substructuring” (WBS) technique. A smaller wave-based description of the coupling interface enables a more efficient reduction of substructures using modal superposition. The smaller-sized system model can then serve as a basis for local design variant analysis. Regarding CAE in an early design stage, a Reduced Beam and Joint Modeling approach has been applied to efficiently analyze and modify the beam-like structures and joints in a vehicle body. The added value of the contributions in this dissertation in a vehicle engineering design practice has been demonstrated by means of applications to realistic models taken from automotive industry.

Order Code

Code: 08D02