Design Optimization of a Multi-Layer Aircraft Canopy Transparency Plate Against Bird Strike

Abstract

Bird strikes in aviation affect flight safety and can lead to financial losses or even fatalities. In this study, a machine learning based optimization approach is used to carry out design optimization of a canopy transparency plate for a fighter aircraft against bird strike. The canopy plate is designed to have a multi-layered structure such that polycarbonate (PC) and stretched polymethyl methacrylate (SPMMA) materials are laminated with a thermoplastic polyurethane (TPU) adhesive. To model PC and SPMMA materials, the Johnson-Cook material model is used. A finite element model is generated for the canopy plate subject to bird strike test conditions, and the lightest structure that provides good collusion performance is investigated. For this purpose, a training data set is created with the Latin hypercube sampling method and a support vector machine (SVM) model that could predict the collision outcome is created. Using the constructed SVM model, optimization is made using genetic algorithm and the optimum transparency design is determined. Finally, the optimum design is subjected to bird strike tests for validation. It is found that the optimum transparency design successfully satisfies the test requirements.