novel neutron texture diffractometer analyzing method with Convolutional Neural Networks for Face-Centered Cubic Material

In this study, we propose a convolutional neural network (CNN) model aimed at inferring texture types and their volume fractions from neutron diffractometer data. The model is trained using labeled texture data of face- centered cubic (FCC) materials, sourced from X-ray diffraction (XRD) measurements. The effectiveness of the model is evaluated using data obtained from neutron diffraction. Compared to traditional data analysis methods, the CNN model not only offers fast and accurate predictions of texture components and their volume fractions, but also demonstrates strong generalization ability. Even under a certain signal-to-noise ratio, the CNN model maintains high accuracy in inferring texture types and their volume fractions. This capability could facilitate the operation of neutron texture diffractometers at lower neutron beam intensities in the future, thus improving both the efficiency of texture analysis.