研究目的
Investigating the use of LSTM neural networks for predicting future events inside a nuclear reactor core based on FBG measurements.
研究成果
The LSTM based neural network accurately predicts the behavior of an optical sensor in a nuclear reactor core, offering potential for early failure detection in nuclear facilities.
研究不足
The study is limited by the challenging conditions of FBG inscription on RAL fibers and the nonlinear fashion of grating intensity loss under neutron irradiation.
1:Experimental Design and Method Selection:
The study utilized an LSTM based neural network for predicting the behavior of an optical sensor in a nuclear reactor core.
2:Sample Selection and Data Sources:
FBG measurements collected from a neutron reactor core were used.
3:List of Experimental Equipment and Materials:
A Ti:sapphire regenerative amplifier, phase mask inscription method, and a specialty random airline (RAL) fiber were used for FBG fabrication.
4:Experimental Procedures and Operational Workflow:
The FBG was inscribed using a femtosecond laser, inserted into a nuclear reactor core, and its responses to harsh conditions were recorded.
5:Data Analysis Methods:
The LSTM system was trained on previous time steps to predict future sensor behavior, with accuracy measured using Root Mean Square Error (RMSE).
独家科研数据包,助您复现前沿成果,加速创新突破
获取完整内容