Experimental and DBN-Based neural network extraction of radiation attenuation coefficient of dry mixture shotcrete produced using different additives

ALKAN ÇAKIROĞLU M., KAPLAN A. N., SÜZEN A. A.

Radiation Physics and Chemistry, vol.188, 2021 (SCI-Expanded, Scopus) identifier

  • Publication Type: Article / Article
  • Volume: 188
  • Publication Date: 2021
  • Doi Number: 10.1016/j.radphyschem.2021.109636
  • Journal Name: Radiation Physics and Chemistry
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Aerospace Database, Chemical Abstracts Core, Chimica, Communication Abstracts, Compendex, EMBASE, INSPEC, Metadex, Pollution Abstracts, Civil Engineering Abstracts
  • Keywords: DBN, Deep learning, Dry mixture shotcrete, Neural network, Radiation attenuation coefficient
  • Isparta University of Applied Sciences Affiliated: Yes

Abstract

In this study, the radiation attenuation coefficients (μm) of different proportions of additives were produced in dry mixture shotcrete both by experimental processes and by deep neural network based on DBN. Fly ash, silica fume, and polypropylene fiber were used as additives of dry mix shotcrete. In the first part of the two-part study, μm values were obtained from seven samples produced and a data set was created along with the input parameters of the experiment. In the second part, a model was developed for predicting the value of μm with input parameters using the DBN deep neural network Algorithm. Experimental data obtained in accordance with both applications and data generated by the Deep Belief Network (DBN) model were analyzed. As a result, the DBN model prediction μm values with an accuracy performance of 87.86%.