Application of SPEA2-MMBB for Distributed Fault Diagnosis in Nuclear Power System DOI Open Access
Ying Xu, Jie Ma,

Jinxiao Yuan

et al.

Processes, Journal Year: 2024, Volume and Issue: 12(12), P. 2620 - 2620

Published: Nov. 21, 2024

Accurate fault diagnosis in nuclear power systems is essential for ensuring reactor stability, reducing the risk of potential faults, enhancing system reliability, and maintaining operational safety. Traditional diagnostic methods, especially those based on single-system approaches, struggle to address complexities composite faults highly coupled data. In this paper, we introduce a distributed method that leverages Strength Pareto Evolutionary Algorithm 2 (SPEA2) multi-objective optimization modified MobileNetV3 neural network with Bottleneck Attention Module (MMBB). The SPEA2 algorithm used optimize sensor feature selection, data are then input into MMBB model training. outputs accuracy rates each subsystem overall system, which subsequently as targets guide refining selection process diagnosis. experimental results demonstrate significantly enhances accuracy, an average 98.73%, achieves comprehensive 95.22%, indicating its superior performance compared traditional network-based approaches.

Language: Английский

One-dimensional decoupled convolutional autoencoder with sparse self-attention mechanism for process monitoring DOI

Yuguo Yang,

Hongbo Shi, Bing Song

et al.

Process Safety and Environmental Protection, Journal Year: 2025, Volume and Issue: unknown, P. 107156 - 107156

Published: April 1, 2025

Language: Английский

Citations

0
Forecasting Dissolved Gas Concentration in Transformer Oil Using the AdaSTDM DOI
Weiqing Lin, Rui Zhao, Jing Chen

et al.

IEEJ Transactions on Electrical and Electronic Engineering, Journal Year: 2025, Volume and Issue: unknown

Published: April 21, 2025

Abstract Accurately forecasting dissolved gas concentration (DGC) in transformer oil is crucial for ensuring the safety and reliability of power transformers facilitating early anomaly warning. Current methods DGC demonstrate limited effectiveness non‐stationary characteristics with data‐distribution shifts. To address this, this paper presents a novel adaptive segmented temporal distribution matching (AdaSTDM) model, consisting Toeplitz inverse covariance‐based clustering (TICC) algorithm time (TDM) algorithm. effectively adapt to different state data, TICC used segment domain sequence, Jensen‐Shannon (JS) divergence as an indicator evaluate segmentation results. The TDM module designed mitigate mismatches by learning common knowledge among states. Experimental results across two real‐world cases illustrate that proposed AdaSTDM outperforms various advanced predicting both stationary data. © 2025 Institute Electrical Engineers Japan. Published Wiley Periodicals LLC.

Language: Английский

Citations

0
Forecasting in-core power distributions in nuclear power plants via a spatial–temporal hierarchical-directed network DOI
Weiqing Lin, Xiren Miao, Chen Jing

et al.

Progress in Nuclear Energy, Journal Year: 2025, Volume and Issue: 186, P. 105795 - 105795

Published: May 10, 2025

Language: Английский

Citations

0
Tool State Recognition Based on POGNN-GRU under Unbalanced Data DOI Creative Commons
Weiming Tong,

Jiaqi Shen,

Zhongwei Li

et al.

Sensors, Journal Year: 2024, Volume and Issue: 24(16), P. 5433 - 5433

Published: Aug. 22, 2024

Accurate recognition of tool state is important for maximizing life. However, the sensor data collected in real-life scenarios has unbalanced characteristics. Additionally, although graph neural networks (GNNs) show excellent performance feature extraction spatial dimension data, it difficult to extract features temporal efficiently. Therefore, we propose a method based on Pruned Optimized Graph Neural Network-Gated Recurrent Unit (POGNN-GRU) under data. Firstly, design Improved-Majority Weighted Minority Oversampling Technique (IMWMOTE) by introducing an adaptive noise removal strategy and improving MWMOTE alleviate problem Subsequently, POG construction multi-scale multi-metric basis Gaussian kernel weight function solve one-sided description single metric basis. Then, construct POGNN-GRU model deeply mine better identify tool. Finally, validation ablation experiments PHM 2010 HMoTP datasets that proposed outperforms other models terms identification, highest accuracy improves 1.62% 1.86% compared with corresponding optimal baseline model.

Language: Английский

Citations

0
Application of SPEA2-MMBB for Distributed Fault Diagnosis in Nuclear Power System DOI Open Access
Ying Xu, Jie Ma,

Jinxiao Yuan

et al.

Processes, Journal Year: 2024, Volume and Issue: 12(12), P. 2620 - 2620

Published: Nov. 21, 2024

Accurate fault diagnosis in nuclear power systems is essential for ensuring reactor stability, reducing the risk of potential faults, enhancing system reliability, and maintaining operational safety. Traditional diagnostic methods, especially those based on single-system approaches, struggle to address complexities composite faults highly coupled data. In this paper, we introduce a distributed method that leverages Strength Pareto Evolutionary Algorithm 2 (SPEA2) multi-objective optimization modified MobileNetV3 neural network with Bottleneck Attention Module (MMBB). The SPEA2 algorithm used optimize sensor feature selection, data are then input into MMBB model training. outputs accuracy rates each subsystem overall system, which subsequently as targets guide refining selection process diagnosis. experimental results demonstrate significantly enhances accuracy, an average 98.73%, achieves comprehensive 95.22%, indicating its superior performance compared traditional network-based approaches.

Language: Английский

Citations

0