Akademik Veri Yönetim Sistemi
Turkish Journal of Science and Health, cilt.2, sa.1, ss.133-138, 2021 (Hakemli Dergi)
Purpose: The use of radioisotopes in diagnosis and treatment in medicine is increasing day by day. In order to make the production of these radioisotopes efficiently, the production cross-sections must be calculated correctly. In the absence of experimental data, cross-sections are calculated in various theoretical ways and the data corresponding to the desired energy value is obtained. In our study, using artificial neural networks as a different approach, an alternative model is presented to estimate cross-sections at unknown neutron energies. Material and Methods: Artificial neural networks method was used to obtain cross-sections of 51Cr radioisotopes produced by neutron-induced reactions. By taking this cross-section data available in the literature, 80% of it was used in the training of the network and the remaining 20% was used in the test. The inputs of artificial neural networks are the incident neutron energies and the output is the cross-section. Hidden layer neuron number 20 was used that gave the best results after many trials. Results: According to the results we have obtained, the artificial neural network method can be used as an alternative method to estimate the radioisotope production cross-sections. While the MSE value of the estimations made over the training data is 0.178 barn, the MSE value on the test data is 0.155 barn. Correlation coefficient values of the predictions of the network on training and test data were found as 0.93 and 0.95, respectively. Conclusion: When compared with the experimental results in the literature, it is concluded that the results of artificial neural networks can be used as an alternative to estimate the cross-section. An advantage of this method is that it allows to obtain results quickly without going into complex mathematical formulation. The results obtained from this study are an indication that cross-sections of any reaction to be performed using any isotope can be obtained by using artificial neural networks method.