Analysis of Radial Dependence of the Localized Magnetic Field using Artificial Neural Networks

The measurements of the angular distributions of charged particles have a long history in atomic and molecular collision studies. To detect all electrons originating from collision has great importance in experimental studies. Due to the physical constraints of the experimental instruments, electrons in definite angles can be detected. Magnetic angle changer is designed to steer electrons scattered at undetectable angles. The magnetic angle changer is a source of the localized magnetic field. A well-controlled magnetic field in the interaction region changes the angles of the electron trajectories. In this study, artificial neural networks have been performed to obtain variation of the magnetic field strength as a function of radial distance calculated from boundary element method. A stringent quality filter is used for data to produce more robust artificial neural network based prediction. The results indicate that the well-trained artificial neural networks can predict the effect on the radial dependence of the localized magnetic field with tremendous precision. It is believed that this study will introduce a new insight into collision studies.