Generation of design allowables on the open hole tensile strength of composite laminates using sensitivity analysis and uncertainty quantification framework

Abstract

Aeronautical
industries are concerned about the cost effective generation of design allowables
for composite laminates. Design allowables take into account the variabilities
arising from different sources (material, manufacturing, defects etc.,) which
are determined using expensive and time consuming experimental campaigns. For
rapid certification and costs reduction, it is of high interest for the
aeronautical industries to use high fidelity numerical models to compliment the
testing. In this work, we use a high fidelity numerical model to simulate open
hole tension (OHT) of composite laminate, followed by an efficient global sensitivity
analysis and uncertainty quantification and management framework to generate design
allowables. In a first step, Morris sensitivity analysis is used to screen the
sensitive input material properties that affect the OHT strength. In the second
step, machine learning technique is used to create a surrogate model, which is
used to obtain the B basis design allowable on the OHT strength.