Abstract:The vibration signals collected from aero-engine accessory casing suffer from serious background noise and strong harmonic interference caused by a combination of nonstationary operating conditions, complex transmission paths, and time-varying airflow environments, which ultimately lead to weak early fault signature information. In recent years, blind deconvolution is considered as a very powerful approach for weak signal detection and attracted much attention in dynamic signal processing area. Moreover, when suppressing the above interference, the traditional blind deconvolution (TBD) method exists the issue of manually solving the filter coefficients, which leads to extraction errors. In order to solve the aforementioned issues, an aero-engine accessory casing weak fault signature mining method based on adaptive global noise suppression (AGNS) deep deconvolution is proposed. First, a noise suppression deep convolution network is constructed as an AGNS deep deconvolution model, and an adaptive fault period block operator is established by utilizing the autocorrelation function to match with the fault period, which is used to perform the adaptive block operation of aero-engine accessory casing vibration signals. Then, a global noise energy amplitude ratio loss function is constructed as the objective function of the AGNS deep deconvolution, which is solved by iterative optimization of the filter coefficients by utilizing the Adam optimization algorithm, so as to gradually suppress the background noise interference and the strong harmonic interference components in the vibration signals of the aero-engine accessory casing. Finally, the effectiveness and generalization of the AGNS deep deconvolution method are verified by utilizing a rolling bearing outer race fault simulated signal under heavy background noise and strong harmonic interference and an aero-engine accessory casing vibration signal under real operating conditions, which indicates that the AGNS deep deconvolution method has the capability of detecting weak faults in the aero-engine accessory casing while comparing it with the latest reported TBD method, which demonstrates that the AGNS deep deconvolution method is superior to the latest reported TBD method, especially in terms of the ability of detecting weak faults in the aero-engine accessory casing.

FAULT DIAGNOSIS OF MULTISTAGE CENTRIFUGAL PUMP UNIT USING NON-LOCAL MEANS-BASED VIBRATION SIGNAL DENOISING

Research on noise reduction method of centrifugal pump coupling fault signal based on LMD-FE-IWT

Application of multi-layer denoising based on ensemble empirical mode decomposition in extraction of fault feature of rotating machinery

Fault Diagnosis in Centrifugal Pumps: A Dual-Scalogram Approach with Convolution Autoencoder and Artificial Neural Network

Fault Diagnosis for Centrifugal Pumps Using Deep Learning and Softmax Regression

A Technique for Centrifugal Pump Fault Detection and Identification Based on a Novel Fault-Specific Mann-Whitney Test

A Signal-Filtering and Feature-Enhancement Method Based on Ensemble Local Mean Decomposition and Adaptive Morphological Filtering

A New Method of Denoising of Vibration Signal and Its Application

A Fault Feature Extraction Method Based on LMD and Wavelet Packet Denoising

Off-Design Operation and Cavitation Detection in Centrifugal Pumps Using Vibration and Motor Stator Current Analyses

Investigation on Vibration Signal Characteristics in a Centrifugal Pump Using EMD-LS-MFDFA

Multistage Centrifugal Pump Fault Diagnosis Using Informative Ratio Principal Component Analysis

Group-Based K-SVD Denoising for Bearing Fault Diagnosis

Multimodal fusion fault diagnosis method under noise interference

A noise-enhanced machine fault diagnosis method with multi-channel information fusion

Multisensor signal denoising based on matching synchrosqueezing wavelet transform for mechanical fault condition assessment

Vibration amplitude normalization enhanced fault diagnosis under conditions of variable speed and extremely limited samples

Adaptive Global Noise Suppression Deep Deconvolution for Weak Fault Signature Mining of Aero-Engine Accessory Casing

Fuzzy Entropy Assisted Singular Spectrum Decomposition to Detect Bearing Faults in Axial Piston Pump

Adaptive frequency-domain nonlocal means for seismic random noise attenuation

Iterative feature mode decomposition: A novel adaptive denoising method for mechanical fault diagnosis