Experimental validation of a hybrid FE-SEA-experimental model

Abstract

Modern manufacturers usually need to know at a design stage if their products will meet certain vibro-acoustic requirements. In many cases the use of physical prototyping becomes significantly expensive, and engineers opt for the development of predictive computational models. In these cases, a hybrid deterministic-statistical approach can overcome the challenges that arise when predicting the response of built-up systems in the mid-frequency range. These models, however, may face limitations when complex vibration sources need to be considered. These limitations were recently addressed using a hybrid statistical-experimental model for the case where the vibration sources are directly coupled to statistical components. In this work the approach is extended to the more realistic scenario where the structure of interest contains also deterministic components. The approach is tested in a laboratory case study consisting of a complex vibration source, rigid beams, and a statistical plate receiver. The developed model is validated by comparing the predicted vibration response to the one obtained by experimentally randomising the receiver plate. The results show that a good agreement is obtained for the statistics of the response of not only the statistical plate receiver, but also the deterministic rigid beams. It is concluded that FE-SEA-experimental models can be a suitable tool for predicting the statistics of the response of complex systems