Modeling of Performance of an artillery shell Using Neural Networks

摘要

This study presents an alternative approach to mathematical mod- eling used hitherto for predicting shell performance in terms of the firing angle needed for the required range. The neural model is shown to be a viable way of modeling many input variables that affect the relationship between the range and the firing angle. It is envisaged that the measured data for shells could be used for de- veloping neural models that would be useful in field applications, including finding the firing angle, the time of flight (this might be required for fuse setting), and the drift angle for the specified range.