A Time-Delayed Neural Network Model for Neutron TOF Spectrum Measurements

Abstract

Abstract-A novel approach for performing the reverse time of flight (RTOF) analysis of the diffraction spectra through the use of time-delayed neural network (TDNN) is introduced. The new design applies a data acquisition system, a special interface card and software program installed in a PC computer, to perform the cross-correlation functions between signals received from the chopper-decoder and detector devices. This paper focuses on theoretical analysis of the neural network architecture to be embedded in the data acquisition system. The implementation of the suggested TDNN along with the experimental measurements will be deferred to a subsequent paper.