Allowable generation and generic basis value for multiple loading and environmental conditions

Abstract

Engineers conduct expensive testing in different environmental conditions to obtain the necessary material allowables. These values consider scatter in composite material constituents and structure according to the “Building Block Testing Approach”. There is a need of a robust methodology that can be implemented to predict material properties from Room Temperature to Elevated Temperatures including presence of moisture as well as determine A- and B- basis allowables with reduced testing. The objective is to reduce tests and risk associated with the use of composites in aerospace structures. In this paper a robust methodology is implemented to predict and validate mechanical properties of IM7-8552 material system at various environmental conditions. The proposed methodology will be also used to generate A-basis and B- Basis allowables with limited test data. A multifactor based mathematical model is implemented to calculate mechanical properties at different environmental conditions. The proposed methodology also uses multi-scale progressive failure integrated with probabilistic and statistical approach that consists of: a) Constituent Variation Method (CVM) and Bayesian statistics; and b) Generic Basis strength values for polymer composites to generate allowables with available limited test data. Mechanical properties of IM7-8552 were predicted and validated with NIAR test data for different environmental conditions. Next, cumulative distribution functions, probabilistic sensitivities, and A- and B- Basis Allowables for unidirectional in tension and compression for un-notched composite were validated with physical testing for HEXCEL 8552 IM7 tape in Cold Temperature Dry and Elevated Temperature Wet. The fiber and the matrix properties were calibrated using deterministic Cold Temperature Dry and Elevated Temperature Wet condition strength and stiffnesses. The constituents' coefficients of variation were initially set to 5% and the software computed real values that likely existed in the test conditions using sensitivities. All simulations were within an acceptable amount of error (less than 2%) for longitudinal tensile, transverse tension and compression and shear loading. Finally, the generic acceptance and equivalence areas for multiple loading and environment conditions were computed and plotted.