Multi-objective matched synchrosqueezing chirplet transform for fault feature extraction from marine turbochargers

Turbocharger is one of the vital parts a diesel engine causing high failure rate. Its surface vibration signal contains important time-varying features. To better process nonstationary signals with features and perform time-frequency transformation on turbocharger signal, novel multi-objective matched synchrosqueezing chirplet transform method proposed in this paper. The based Linear Chirplet Transform to optimize selection demodulation Parameters such as Rayleigh entropy signal-to-noise ratio are used targets select value optimal Then local maximum post-processing for rearrangement signal. This improves energy concentration result while maintaining ability reconstruction. On test stand, fault samples were obtained. time-domain at 1×, 2×, 9× frequencies reconstructed results, characteristic parameters extracted from them. effectiveness feature parameter identification was validated by Principal Component Analysis. study showed that our MOMSSCT capable processing analysing strongly turbochargers. results have aggregation, clear trajectories, 37.5% reduction frequency spread width. visualization show data different types more clearly differentiated two-dimensional graph than common good classifiability under various operating conditions. Using MOMSSCT, diagnostic accuracy rate can reach 85%.