Research on equipment fault diagnosis model based on gan and inverse PINN: Solutions for data imbalance and rare faults.

Journal: PloS one
Published Date:

Abstract

In the field of medical imaging equipment, fault diagnosis plays a vital role in guaranteeing stable operation and prolonging service life. Traditional diagnostic approaches, though, are confronted with issues like intricate fault modes, as well as scarce and imbalanced data. This paper puts forward a fault diagnosis model integrating digital twin technology and Inverse Physics - Informed Neural Networks (Inverse PINN).The practical significance of this research lies in its potential to revolutionize the engineering aspects of medical imaging equipment management. By constructing a physical model of equipment operation and leveraging inverse PINN to deal with imbalanced datasets, the model can accurately identify and predict potential faults. This not only optimizes the full lifecycle management of the equipment but also has the potential to reduce maintenance costs, improve equipment availability, and enhance the overall efficiency of medical imaging services.Experimental results show that the proposed model outperforms in fault detection and prediction for medical imaging equipment, especially making breakthroughs in data generation and fault detection accuracy. Finally, the paper discusses the model's limitations and future development directions.

Authors

  • Jian Deng
    Department of Mechanics, Zhejiang University, Hangzhou 310027, People's Republic of China.
  • Zheng Cheng
    Key Laboratory of Mesoscopic Chemistry of Ministry of Education, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing 210023, P. R. China.
  • Aiming Gu
    Jiaxing First Hospital, Jiaxing, China.
  • Shibohua Zhang
    School of economics and management, Xi'an University of Technology, Xi'an, China.