Battery packs used in electrified vehicles exhibit high modal density due to the repeated cell substructures. If the excitation frequency is in the region of high modal density, small commonly occurring structural variations can lead to drastic changes in the vibration response. Because the battery pack fatigue life depends strongly on the vibration response, a statistical analysis of the vibration response with structural variations is important from a design point of view. The parametric reduced order models (PROMs) are developed to efficiently and accurately predict the vibration response with structural variations. Additionally, an iteration approach to handle material nonlinearities has also been built into reduced order models, since these material nonlinearities are common within battery cells.