Circuit Fault Detection of an AC Stable Power Supply Based on a Data Driven Method

Authors

  • Xiaoyu Chen Electronic and Information Engineering College, Henan Institute of Technology, Xinxiang 453000
  • Pengfei Zheng College of Electronic and Electrical Engineering, Henan Normal University, Xinxiang 453000, China

Abstract

When the circuit fault detection of an AC stable power supply based on data driven is studied, theAC stable power circuit acquisition system, composed
of the intelligent controller, signal conditioning module and data sensor, is adopted to collect the data of the AC stable power circuit. The circuit
fault detection method based on principal component analysis is used to construct an original observation matrix according to the acquired AC stable
power circuit data. The co-variance matrix of the original observation matrix is carried out using singular value decomposition, and the singular
value and eigenvalue are decomposed into the residual subspace and principal component space, the observation matrix, constructed according to
the new AC stable power circuit data, is then projected into the residual subspace and the principal component space. When the statistic value of
the new observation matrix is larger than that of the original observation matrix, the circuit is judged to be faulty, otherwise the circuit is normal. The
simulation results show that this method can accurately detect the specific fault circuit of an AC stable power supply, and it takes less time to detect
the faults of each circuit. Moreover, after completing the circuit fault detection of an AC stable power supply, the service life of the AC stable power
supply increases, which can improve the economic benefits of the AC stable power supply.

Keywords: data driven; stable AC; power circuit; fault detection; observation matrix; principal component analysis

Cite As

X. Chen, P. Zheng, "Circuit Fault Detection of an AC Stable Power Supply Based on a Data Driven Method", Engineering
Intelligent Systems,
vol. 29 no. 1, pp. 11-17, 2021.


Published

2021-01-01