Application of neural network for predicting photon attenuation through materials. Issue 3 (3rd April 2019)

• Record Type: Journal Article
• Authors: Medhat, M. E.
• Is Part Of: Radiation effects and defects in solids. Volume 174:Issue 3/4(2019)
• Page Start: 171
• Page End: 181
• Publication Date: 2019-04-03
• Subjects: Neural network -- composite materials -- mass attenuation coefficient; Radiation chemistry -- Periodicals; Crystals -- Defects -- Periodicals; Crystal lattices -- Periodicals; 530.416
• DOI: 10.1080/10420150.2018.1547903
• Languages: English
• ISSNs: 1042-0150

Abstract

ABSTRACT: Photon attenuation prediction in composite materials usually requires a great expert knowledge and time-consuming calculations with complex procedures especially in experimental arrangements. An artificial neural network can be applied to predict exactly the value of mass attenuation at different energies. For training of the network a neural model was designed with chemical composition and molecular cross-section of samples as neural input, while the photon attenuation coefficient was its output. The method was applied for different composite materials with different chemical compositions. The results of mass attenuation coefficients were compared with the experimental and theoretical data for the same samples and a good agreement has been observed. The results indicate that this process can be followed to determine the data on the attenuation of gamma-rays with the several energies in different materials.