APPLICATION OF MACHINE LEARNING IN PREDICTING THE THICKNESS OF MATERIAL PLATES BASED ON THE GAMMA TRANSMISSION TECHNIQUE
Main Article Content
Abstract
This study proposes an approach using machine learning to predict the thickness of material plates based on the gamma transmission technique. First, the data generated from the Monte Carlo simulation was utilized to train and optimize the machine learning model. Then, the experimental ratio R corresponding to different thicknesses of aluminum, PMMA, and iron plates will be input into the optimal model to predict the thickness. The predicted thicknesses of material plates from the linear calibration curves (LCCs) constructed from simulation data are compared with the predictions from the ML model to evaluate the reliability of this approach. The results obtained for seven of nine samples show that the average relative deviation between the reference thickness values and those predicted by the ML model is less than 2%, while it is about 4% when the LCCs are used. This is the basis for further applications of machine learning to predict the thickness of various materials.
Keywords
gamma transmission, machine learning, Monte Carlo simulation, MLP-ANN, NaI(Tl), thickness
Article Details
References
Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. MIT Press.
Hoang, D. T., Huynh, D. C., Tran, T. T., & Chau, V. T. (2016). A study of the effect of Al2O3 reflector on response function of NaI(Tl) detector. Radiation Physics and Chemistry, 125,
88-93. https://doi.org/10.1016/j.radphyschem.2016.03.020
Huang, Z., Zhu, J., Zhuo, L., Li, C., Liu, C., Hao, W., & X. W. (2022). Non-destructive evaluation of uneven coating thickness based on active long pulse thermography. NDT & E International, 130, 102672. https://doi.org/10.1016/j.ndteint.2022.102672
Huseyin Sahiner, X. L. (2020). Gamma spectral analysis by artificial neural network coupled with Monte Carlo simulations. Nuclear Instruments and Methods in Physics Research Section A: Accelerators, Spectrometers, Detectors and Associated Equipment, 953, Article 163062. https://doi.org/10.1016/j.nima.2019.163062
Huynh, D. C., Le, T. N. T., Le, H. M., Nguyen T. T. L., Hoang, D. T., & Tran, T. T. (2022). Thickness determination of material plates by gamma-ray transmission technique using calibration curves constructed from Monte Carlo simulation. Radiation Physics and Chemistry, 190. https://doi.org/10.1016/j.radphyschem.2021.109821
Huynh, D. C., Truong, T. S., Le, T. N. T., Nguyen, T. T. L., Le, H. M., & Hoang, D. T. (2023). Thickness measurement of material plates using low-activity sources with various energies in gamma-ray transmission technique. Applied Radiation and Isotopes, 194. https://doi.org/10.1016/j.apradiso.2023.110729
Kanja, J., Mills, R., Li, X., Brunskill, H., Hunter, A. K., & Dwyer-Joyce, R. S. (2021). Non-contact measurement of the thickness of a surface film using a superimposed ultrasonic standing wave. Ultrasonics, 110, Article 106291. https://doi.org/10.1016/j.ultras.2020.106291
Song, J., Guo, D., Jia, J., & S. T. (2022). A new on-line ultrasonic thickness monitoring system for high-temperature pipes. International Journal of Pressure Vessels and Piping, 199, Article 104691. https://doi.org/10.1016/j.ijpvp.2022.104691
Nguyen, V. H., Chuong, H. D., Thanh, T. T., & Chau, V. T. (2018). New method for processing gamma backscattering spectra to estimate saturation depth and to determine thickness of aluminum and steel materials. Journal of Radioanalytical and Nuclear Chemistry, 315(1),
293-298. https://doi.org/10.1007/s10967-017-5667-0.
Smolensky, P., Chauvin, Y., & Rumelhart, D. E. (1995). Backpropagation: Theory, architectures, and applications. Lawrence Erlbaum Associates.
Truong, T. S., Huynh, D. C., & Dinh, H. T. (2021). An artificial neural network based approach for estimating the density of liquid applied in gamma transmission and gamma scattering techniques. Applied Radiation and Isotopes, 169, Article 109570. https://doi.org/10.1016/j.apradiso.2020.109570
Xue, Z., Fan, M., Cao, B., & Wen, D. (2021). Enhancement of thickness measurement in eddy current testing using a log–log method. Journal of Nondestructive Evaluation, 40(2), 1-10. https://doi.org/10.1007/s10921-021-00773-x