Online Correlation Analysis and FCM-based Fault Detection of Current and
Speed Signals for Underwater Thruster Entanglement
Abstract
The fault diagnosis with one single sensor cannot comprehensively use
the multi-sensor correlation information of fault signals for underwater
thruster entanglement. To solve the real-time fault diagnosis problem
that in the case of entanglement, a kind of fault diagnosis method based
on the combination of current and rotational speed signal correlation
analysis and Fuzzy C-means clustering (FCM) is proposed. Firstly, the
collected current and speed signals of underwater thrusters under
different states are normalized; Secondly, the correlation of normalized
current and speed signals is calculated, and the correlation matrix is
formed, and the sliding window of sampling series is adopted to
calculate the correlation coefficient change over time between current
and speed. Then a kind of Fuzzy C-means clustering method based on
improved distance index is used to diagnose fault with correlation
matrix elements. To verify the effectiveness of the proposed method, the
sampling data in the case of propeller entanglement is used to verify
the proposed fault detection method. The results show that the proposed
method can fully extract the multi-sensor correlation information of
underwater thrusters compared with the fault diagnosis method using only
one signal using Support Vector Machine, and the feature extraction is
more sufficient, effectively improving the accuracy of underwater
thruster fault diagnosis.