Virtual product development process for reducing noise, vibration, and harshness of vehicle based on substructuring and artificial neural network

Abstract

In this study, virtual product development method for reducing vibration and noise is proposed for designing at the concept development stage of a vehicle. To this end, the vibration characteristics of the system are predicted through the Lagrange-multiplier frequency-based substructuring technique. The concepts of contact, blocked and transmitted force, and force transmissibility were used for determining the improvement subsystem or combination of subsystems when using the modular platform. Moreover, after the subsystems to be improved were determined, Artificial Neural Network was used as a method of predicting vibration characteristics according to the change of design variables. To verify this, the prediction of the blocked force was performed by changing the young’s modulus of the simplified substructure. Finally, the reduction in response was confirmed by applying the blocked force of the simplified subframe to the simplified structure, and a vehicle development process using a database at the concept setting stage is proposed.