Engine air induction shell noise is a structure borne noise that radiates from the surface of the air induction system. The noise is driven by pulsating engine induction air and is perceived as annoying by vehicle passengers. The problem is aggravated by the vehicle design demands for low weight components packaged in an increasingly tight under hood environment. Shell noise problems are often not discovered until production intent parts are available and tested on the vehicle. Part changes are often necessary which threatens program timing. Shell noise should be analyzed in the air induction system design phase and a good shell noise analytical process and targets must be defined.Several air induction clean side ducts are selected for this study. The ducts shell noise is assessed in terms of material strength and structural stiffness. A measurement process is developed to evaluate shell noise of the air induction components. Noise levels are measured inside of the clean side ducts. The surface vibrations of the duct are measured using accelerometer and laser vibrometer. Measured noise levels drive a speaker and sound intensity measurements are made on a test bench to compute the surface radiated sound power. A simulation process is developed to predict the radiated sound power for the same air induction components. Internal acoustical pressure of the component is used to drive a frequency response finite element model. Then, two methods are utilized to predict the radiated sound power. One method is using Rayleigh Integral method to estimate radiated sound power directly from surface velocity calculation. The other method is using Boundary Element Method (BEM) to apply calculated surface velocities to a boundary element model and radiated sound power is predicted.Results of the bench study are compared with the prediction data from the developed simulation process. Results of the correlation conclude that the simulation process is good for predicting radiated sound power of engine air induction components. The BEM method has better correlation than Rayleigh method.