Abstract
This study focuses on the vibration response estimation for an exhaust gas system of passenger vehicles using the neural network technique. At first, FEM (Finite Element Method) calculations of the transient response for an impact load were carried out to the representative exhaust gas systems. These calculated data were used as the teacher's data of the neural network, and a computer program was developed to obtain the vibration amplitude for similar exhaust gas system. If a designer repeats the input to this neural network model for various support conditions, the optional support position can be gained without carrying out such complex FEM analysis repeatedly.
