Application of Statistical Model Validation to a Truck Frame

A statistical model validation process was used to determine predictive confidence bounds for mass and natural frequency predictions applied to 1) stamped crossmember and 2) truck frame finite element models.

For the crossmember, modal performance measure sensitivity to variation parameters was determined using a designed experiment based on a set of 121 finite element model analyses and field tests of 7 hardware samples, A similar experiment based on a set of 8 finite element model analyses and field tests of 8 hardware samples of the truck frame was conducted.

Statistical measures were used to compare and test distributions of computer model and field outputs with the intent of computing confidence bounds for the selected performance statistics.