Date of Award


Publication Type


Degree Name



Mechanical, Automotive, and Materials Engineering

First Advisor

Y. Kim

Second Advisor

A. Rahini

Third Advisor

A. Cherniaev


Artificial neural network, Ballistic limit equation, Hypervelocity impact, Orbital debris, Sandwich panels, Spacecraft shielding




Sandwich panels are widely used in the design of uninhabited satellites and, in addition to having a structural function can often serve as shielding, protecting the satellites’ equipment from hypervelocity impacts (HVI) of orbital debris, and micrometeoroids. This thesis aims to provide: a comprehensive review of HVI experimental studies for honeycomb- and open-cell foam-cores; an examination of available predictive models used to assess the panels’ ballistic limits; as well as signify the influence of honeycomb-core parameters, such as cell size and foil thickness, as well as core material, on the ballistic performance of honeycomb-core sandwich panels (HCSP) when subject to HVI scenarios.

To study the influence of HCSP parameters, two predictive models: a dedicated ballistic limit equation (BLE)—based on the Whipple shield BLE—and an artificial neural network (ANN) trained to predict the outcomes of HVI on HCSP were developed. A database composed of physical and numerical simulations allowed for BLE fitting and ANN training. The ANN was developed using MATLAB’s Deep Learning Toolbox framework and was tuned using a comprehensive parametric study to optimize the ANN architecture, including such parameters as the activation function, the number of hidden layers and the number of nodes per layer. The predictive models were verified using a new set of simulation data and achieved low error percentage in comparison when predicting the ballistic limits of HCSP, ranging from 1.13% to 5.58% (BLE) and 0.67% and 7.27% (ANN), respectfully.