Combat helmet liner design for blunt impact absorption using multi-output Gaussian process surrogates

Abstract

A finite element based computational model simulating the standard drop tower test for military helmets was created and used in conjunction with a multi-output Gaussian process surrogate to seek different designs of helmets for improved blunt impact performance. Experimental drop test results were used for the validation of the model’s ability to simulate impact. The influence of foam stiffness, impact velocity, strap tension, as well as pad placement and size on parameters on the peak linear acceleration (PLA) of the headform was investigated for the first time through a surrogate model trained by strategically choosing simulation points. Impact velocity was found to have the greatest effect. The strap tension and foam pad stiffness ranges examined within this sampling plan were found to have less of an effect on the performance of the helmet than the pad size and shape parameters examined. The surrogate modeling approach was used to quantify the influence of design parameters and can lead to not only improved helmet designs but also new data-driven design metrics and testing standards to accelerate the development of TBI-mitigating helmets.