Deep-Learning and Dynamic Time Warping-Based Approaches for the Diagnosis of Reactor Systems DOI Creative Commons
H. Y. Jeong,

Jihyun Kim,

Doyun Jung

et al.

Sensors, Journal Year: 2024, Volume and Issue: 24(23), P. 7865 - 7865

Published: Dec. 9, 2024

The degradation of clamping force in the core support barrel, which forms internal structure a nuclear power plant, has potential to significantly impact plant's safety and reliability. Previous studies have concentrated on detection but been constrained their ability identify precise size position. This study proposes novel methodology for diagnosing position barrels, combining deep-learning techniques dynamic time warping (DTW) algorithms. DTW is applied magnitude data ex-core neutron noise signal obtained frequency domain, thereby enabling effective learning changes sensor values. Moreover, autoencoder-based (AE-based) representation utilized extract features data, preventing overfitting thus enhancing robustness model. experiment results demonstrate that can be accurately predicted. It expected this research will contribute precision efficiency monitoring plants.

Language: Английский

Deep-Learning and Dynamic Time Warping-Based Approaches for the Diagnosis of Reactor Systems DOI Creative Commons
H. Y. Jeong,

Jihyun Kim,

Doyun Jung

et al.

Sensors, Journal Year: 2024, Volume and Issue: 24(23), P. 7865 - 7865

Published: Dec. 9, 2024

The degradation of clamping force in the core support barrel, which forms internal structure a nuclear power plant, has potential to significantly impact plant's safety and reliability. Previous studies have concentrated on detection but been constrained their ability identify precise size position. This study proposes novel methodology for diagnosing position barrels, combining deep-learning techniques dynamic time warping (DTW) algorithms. DTW is applied magnitude data ex-core neutron noise signal obtained frequency domain, thereby enabling effective learning changes sensor values. Moreover, autoencoder-based (AE-based) representation utilized extract features data, preventing overfitting thus enhancing robustness model. experiment results demonstrate that can be accurately predicted. It expected this research will contribute precision efficiency monitoring plants.

Language: Английский

Citations

0