A Data-Driven Approach to the Impedance Matched Multi-Axis Test Method

Abstract

Environmental testing is critical in certifying systems to operate and survive in harsh vibration environments. An example of this would be cargo rockets destined to deliver supplies to a low-earth orbiting space station. Aerodynamic effects would impose a distributed random excitation load onto the cargo rocket. It is imperative that test engineers perform an environmental test to replicate, as best as possible, the anticipated vibration loads on the rocket without subjecting it to the actual operational environment. Traditional in-lab environmental tests have shown to poorly reproduce the true response of a system subjected to distributed excitation loads. In recent literature, the Impedance Matched Multi-Axis Test (IMMAT) was developed to mitigate some of the current limitations of environmental tests through the use of finite element models (FEM) and multi-input multi-output (MIMO) control. As an extension of IMMAT, the present work investigates the use of a data-driven approach to supplement the creation of a numerical model used to predict optimal excitation locations and forces. The main advantage of this alternative approach to IMMAT removes the need for a FEM of the system being tested. This extends IMMAT to cases where a model of the subject being tested is not readily available and, yet, provides an opportunity for IMMAT to be deployed for certification. Additionally, a data-driven approach has the potential to capture more realistic test parameters, such as hard-to-recreate boundary conditions, material properties, and optimal excitation locations, as it is built directly from test data. There may be cases where this is advantageous, but may require better test designs. A preliminary numerical simulation is implemented to test the effectiveness and practicality of using vector fitted accelerance frequency response functions (FRF) to perform IMMAT.