Genetic Algorithms for the Vibroacoustic Optimization of the Thin Parts of I.C. Engine

The Noise Vibration Harshness (NVH) behavior of engines is one of the predominant factors for market acceptance of vehicles. To reach this goal it is necessary to reduce the absolute noise level and also the noise level in specific frequency ranges. The design of radiating structures for minimal sound radiation is a multidisciplinary problem that involves complex objective functions and expensive computations. In this paper, genetic algorithm is used as a promising tool for numerical optimization of such problems. The objective of the study is to determine effective, general design methods for determining the optimal design of thin parts of I.C. engine that minimizes the total radiated acoustic power. Variable attached discrete masses are considered. Acoustic response is minimized either at a single frequency or first five natural frequencies. Radiated sound power is calculated using a boundary element method, in conjunction with a finite element solver ANSYS for the solution of the structural acoustical problem.