Loading...

Unfolding of the Gamma-Ray Spectrum of the Scintillator Detectors Using the Neural Network Method

Pezeshki, Ali | 2011

528 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 42146 (46)
  4. University: Sharif University of Technology
  5. Department: Energy Engineering
  6. Advisor(s): Vosoughi, Naser
  7. Abstract:
  8. The analysis of gamma-ray spectra of low resolution detectors is difficult and sometimes impossible, as a result of photo peak’s overlapping. The usual methods of radiation spectra analysis based on fitting of peaks to mathematical curves or detecting of peaks with numerical calculation, are valid for high resolution detectors. However these methods are less successful for lower resolution detectors such as the common scintillators. The wide peaks in the spectrum maybe overlap and make it difficult to analysis. To solve this problem, we test here a method, based on the use of the artificial neural network. At first spectra of elements are converted to the patterns which are suitable for the neural network and after that the net is trained through Error backpropagation perceptron algorithm. At last unknown spectra can be analyzed with associative memory according to the matrixes of weights, have been calculated before. Furthermore, the pattern recognition is based on the whole spectrum instead of each individual peak. It can be used with advantage for low resolution detectors. Finally we built a database consist of 8 isotopes (198Au ,28Al, 54Mn, 56Mn 59Fe,92Sr, 65Zn, 65Ni) for benchmarking of our program. These isotopes have partial and also complete overlapped peaks, and the program was successful for whole exams which are prepared for testing it
  9. Keywords:
  10. Neural Network ; Scintillation Detector ; Error Backpropagation ; Gamma Ray Spectrum

 Digital Object List

 Bookmark

No TOC