USING NEURAL NETWORKS TO OPTIMIZE LETHALITY OF FRAGMENTING WARHEADS

Ayisit O., Saran S., Erbil B. A., Kaya C.

33rd International Symposium on Ballistics, BALLISTICS 2023, Bruges, Belgium, 16 - 20 October 2023, vol.1, pp.1200-1209, (Full Text) identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 1
  • City: Bruges
  • Country: Belgium
  • Page Numbers: pp.1200-1209
  • Middle East Technical University Affiliated: Yes

Abstract

In this paper, we propose a novel approach for optimizing the lethality of fragmenting warheads using neural networks against a 6-wheel generic military truck. Our approach involves training a neural network on a dataset of Pk (Probability of kill) values of the target and maximizing lethality. The neural network takes in four independent warhead parameters as input: fragment type, number of fragment layers, fragment size, and warhead diameter. We used Kuzgun Tech in-house Fragmentation Analysis Software (FAST) to determine warhead fragmentation patterns. We also utilized VERSUS lethality/vulnerability software developed by Kuzgun Tech for the calculation of dynamic warhead-target interactions and to predict Pk values of the target. We setup vulnerability model of the generic military truck based on its 3D CAD data, identified critical components, specified component kill criteria and kill tree of the truck.