特集 放射線計測の新しい手法とデバイス
放射化法を用いた中性子ドシメトリーヘのニューラルネットワークの適用
1995 年 21 巻 3 号 p. 2-11
詳細
1995 年 21 巻 3 号 p. 2-11
A neural network technique has been applied to neutron dosimetry using activation detectors. The performance is checked through analyses of mock-up data for fission reactor neutron fields, including a survey of the network structure suitable for this application. The results from measured data are also compared with those obtained by a conventional method for neutron spectrum adjustment (J1 method or NEUPAC). It is shown that the neural network adjustment has some potential advantages over the conventional methods in computaional simplicity and solution validity.
