A deep transfer learning method for system-level fault diagnosis of nuclear power plants under different power levels

2022 ◽  
Vol 166 ◽  
pp. 108771
Author(s):  
Zhichao Wang ◽  
Hong Xia ◽  
Jiyu Zhang ◽  
M. Annor-Nyarko ◽  
Shaomin Zhu ◽  
...  
1994 ◽  
Vol 44 (3) ◽  
pp. 225-235 ◽  
Author(s):  
Qin Zhang ◽  
Xuegao An ◽  
Jin Gu ◽  
Binquan Zhao ◽  
Dazhi Xu ◽  
...  

2015 ◽  
Vol 2015 ◽  
pp. 1-15 ◽  
Author(s):  
Xinghui Zhang ◽  
Jianshe Kang ◽  
Lei Xiao ◽  
Jianmin Zhao

Gear and bearing play an important role as key components of rotating machinery power transmission systems in nuclear power plants. Their state conditions are very important for safety and normal operation of entire nuclear power plant. Vibration based condition monitoring is more complicated for the gear and bearing of planetary gearbox than those of fixed-axis gearbox. Many theoretical and engineering challenges in planetary gearbox fault diagnosis have not yet been resolved which are of great importance for nuclear power plants. A detailed vibration condition monitoring review of planetary gearbox used in nuclear power plants is conducted in this paper. A new fault diagnosis method of planetary gearbox gears is proposed. Bearing fault data, bearing simulation data, and gear fault data are used to test the new method. Signals preprocessed using dilation-erosion gradient filter and fast Fourier transform for fault information extraction. The length of structuring element (SE) of dilation-erosion gradient filter is optimized by alpha stable distribution. Method experimental verification confirmed that parameter alpha is superior compared to kurtosis since it can reflect the form of entire signal and it cannot be influenced by noise similar to impulse.


2021 ◽  
Author(s):  
Le Zhang ◽  
Wei Cheng ◽  
Xue Liu ◽  
Xuefeng Chen ◽  
Fengtian Chang ◽  
...  

2021 ◽  
Author(s):  
Jiangkuan Li ◽  
Meng Lin

Abstract With the development of artificial intelligence technology, data-driven methods have become the core of fault diagnosis models in nuclear power plants. Despite the advantages of high flexibility and practicability, data-driven methods may be sensitive to the noise in measurement data, which is inevitable in the process of data measurement in nuclear power plants, especially under fault conditions. In this paper, a fault diagnosis model based on Random Forest (RF) is established. Firstly, its diagnostic performance on noiseless data and noisy data set containing 13 operating conditions (one steady state condition and 12 fault conditions) is analyzed, which shows that the model based on RF has poor robustness under noisy data. In order to improve the robustness of the model under noisy data, a method named ‘Train with Noisy Data’ (TWND) is proposed, the results show that TWND method can effectively improve the robustness of the model based on RF under noisy data, and the degree of improvement is related to the noise levels of added noisy data. This paper can provide reference for robustness analysis and robustness improvement of nuclear power plants fault diagnosis models based on other data-driven methods.


Sign in / Sign up

Export Citation Format

Share Document