A New Feature Extraction Method Based on Instantaneous Phase for Clamp Looseness on Pipes

F388 extraction is a key step in mechanical fault diagnosis and the phase is a significant parameter that is more sensitive to the regularity of signals than amplitude. This work proposed a new method to extract feature of complexity from the phase of signals, which is the auto-correlation scatterplot of instantaneous phase (ASIP). Moreover, the high diversity of ASIP represents that the signal has complex components. According to the simulation analysis, the ASIP is a more sensitive way to represent the appearance of random component in signals, so that the diversity of ASIP is a good feature for regularity detection. And in the experimental analysis on detection of clamp looseness on pipes, it demonstrates that the looseness of the clamp would affect the strain of pipe on its upstream side, and the distribution of the ASIP could visually reflect the complexity of the signal collected by fiber Bragg grating sensors placed on the pipe. Therefore, ASIP is an effective method to describe the regularity of a complex signal.