Investigation of a Statistical Metric for Validation of a Structural Dynamics Model

Structural vibration is an important component of spacecraft launch and operation. Consequently, improved models have been developed to predict physical system responses during launch and operation. However, because model errors can be significant they are typically not used without first being accredited via validation methods. Model validation is the act of demonstrating the adequacy or accuracy of predictions of mathematical models for use within a specific area. This project investigates the effectiveness of a general statistical technique used to validate structural vibration models. To carry out this analysis we compare experimental and numerical responses by forming probabilistic representations for model predicted structural behaviors. From these distributions we calculate P-valued probability intervals used to recast the problem as a serious of Bernoulli experiments from which a minimum and actual validation score are derived. In order to execute this algorithm and determine the effectiveness of our vibration model a MATLAB module will be developed and implemented for a selected set of test data. The results will be compared to the conclusions of a more routine modal analysis comparison