Using Model Uncertainty to Reduce Verification and Validation in Noise and Vibration Problems
Verifying and validating that a mechanical system meets the design requirements is often a costly iterative activity. This is particularity true, for example, with complex vehicle systems that must meet noise and vibration requirements to ensure vehicle occupant comfort. We show here how analysis of model uncertainty can speed verification testing by bounding and guiding hardware prototype redesign. Vibration and acoustic model uncertainty and residual errors are estimated, and then analysis derived to ensure that this uncertainty range is covered by a planned set of design changes. We further use these results to define a complexity metric based on uncertainty, and an adaptability metric based on the domain of available adjustment. We then propose a capability metric by comparing the range of uncertainty against the range of adaptability provided. We demonstrate the efficacy with an example from the elevator system design, rapidly meeting noise and vibration requirements with only one prototype iteration.