Research Article

Research on gear fault diagnosis for nuclear power circulating water pump based on deep neural network

¹ School of Mechanical Engineering, Xi’an Jiaotong University, Xi’an 710049, PR China
² National Key Lab of Aerospace Power System and Plasma Technology, Xi’an Jiaotong University, Xi’an 710049, PR China
³ China Nuclear Power Engineering Co. Ltd., Beijing 100000, PR China

^* e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.

Received: 27 June 2025
Accepted: 13 November 2025

Abstract

In nuclear power circulating water pump systems, gears, as critical transmission components, operate in low-speed, heavy-load environments and are crucial for maintaining the stable operation of the pump. To significantly enhance the efficiency and accuracy of fault diagnosis, this paper proposes a gear fault diagnosis technology based on deep neural networks. Efficient Fast Fourier Transform (FFT) techniques are employed to extract frequency domain samples containing rich fault information. Subsequently, a deep neural network model is designed and trained to automatically extract deep features from the frequency domain signals and accurately identify different fault types. By comparing the performance of various deep neural network models, precise diagnosis of gear faults in nuclear power circulating water pumps is achieved. This method is of significant importance for improving nuclear power safety, optimizing plant operational efficiency, and enhancing economic benefits.