Period-oriented multi-hierarchy deconvolution and its application for bearing fault diagnosis

Area of Science:

• Signal Processing
• Mechanical Engineering
• Condition Monitoring

Background:

• Deconvolution methods are vital for extracting excitation sources from noisy signals.
• Traditional deconvolution methods often suffer from incomplete information extraction.
• Accurate fault period identification is a prerequisite for many existing deconvolution techniques.

Purpose of the Study:

• To introduce a novel deconvolution method, period-oriented multi-hierarchy deconvolution (POMHD), to address limitations in current techniques.
• To develop a method capable of extracting multiple latent fault components without requiring prior fault period knowledge.
• To enhance the accuracy and robustness of fault diagnosis in rotating machinery.

Main Methods:

• Adaptive filter design using an iterative algorithm to update filter coefficients.
• Utilizing harmonic-to-noise ratio as the deconvolution orientation.
• Introducing a normalized proportion of harmonics index for fault feature evaluation.
• Constructing a harmonics proportion diagram for diagnostic decisions.

Main Results:

• The proposed POMHD method successfully overcomes limitations of traditional deconvolution techniques.
• POMHD can simultaneously extract multiple latent fault components from signals.
• The method provides an intuitive graphical representation of fault information via the harmonics proportion diagram.
• Validation using simulated and experimental data, including single and compound bearing faults, demonstrates feasibility and robustness.

Conclusions:

• The period-oriented multi-hierarchy deconvolution (POMHD) offers a significant advancement in signal processing for fault diagnosis.
• POMHD effectively extracts multiple fault signatures even without precise prior knowledge of the fault period.
• The method's ability to present diagnostic information in a clear, hierarchical diagram enhances its practical applicability in condition monitoring.