Development of a Multibody Model for NVH Analysis of a High-Performance Internal Combustion Engine

In the development of high-performance motorcycle engines, managing mechanical noise represents a sophisticated engineering challenge that directly defines the premium character of the vehicle and the emotional engagement of the rider. While the acoustic signature of the exhaust system provides the desirable soul of the machine, suppressing parasitic mechanical emissions is essential to ensure a refined, high-quality sound profile that meets modern market expectations. This study, conducted during an industrial internship at Motori Minarelli S.p.A., focuses on the analysis of surface noise emissivity for the MM460 engine. The primary motivation behind this work was to transform noise management from a reactive troubleshooting phase into a proactive design advantage. By developing a high-fidelity dynamic multibody model, the project provides a robust virtual environment capable of predicting vibro-acoustic behaviour long before physical prototypes are built, thereby accelerating the development cycle and ensuring structural excellence. The technical activity was structured into several stages, beginning with the detailed preparation of component geometries and finite element discretization using Altair HyperMesh. This was followed by the application of dynamic condensation within Altair OptiStruct to generate the necessary flexible bodies, ensuring that the model retained its dynamic accuracy while remaining computationally efficient. Rather than moving directly to full-scale testing, the project followed a strategy of increasing complexity, starting from simplified sub-systems and progressively integrating components into a comprehensive simulation environment within AVL Excite™ PowerUnit. This methodical buildup allowed for a detailed comprehension of how internal excitation forces and operational dynamics propagate through the engine structure. A key part of the analysis involved identifying specific engine regions with the highest noise emissivity by visualizing energy radiation patterns. By utilizing Altair HyperView to pinpoint structural weaknesses and extracting data via AVL Impress™ Chart, the study moved beyond mere observation toward active optimization, using MATLAB post-processing to evaluate overall performance and propose targeted structural solutions. This integrated approach demonstrates how advanced simulation effectively isolates noise sources and guides the selection of the most appropriate design modifications. For instance, the model allows for the evaluation of stiffening measures, such as the strategic addition of structural ribs designed to create a more rigid architecture; this suppresses the 'drumming effect' of thin-walled casings, preventing them from acting like diaphragms that radiate noise. Beyond structural reinforcement, the simulation provides the necessary data to implement insulation through sound-absorbent materials or acoustic shielding to dampen residual radiation. Finally, the analysis can target the root of the problem by reducing excitation at its source, exploring the optimization of reciprocating masses or the introduction of geometric offsets to minimize inertial impacts and piston slap. Ultimately, this virtual tool serves as the essential link between identifying a noise issue and proposing the most effective technical solution to refine the acoustic footprint of future high-performance powertrains.