Hilbert-Huang Transform (HHT): A Novel Non-Stationary Signal Processing Technique with Implementation

This paper investigates non-stationary signal processing using the Hilbert-Huang Transform (HHT), presenting both comprehensive research documentation and practical source code implementation. The HHT methodology integrates two core algorithmic procedures: Empirical Mode Decomposition (EMD) and Hilbert Spectral Analysis. The EMD algorithm implements an adaptive time-series decomposition process where any non-stationary signal undergoes iterative sifting to extract multiple Intrinsic Mode Functions (IMFs), each representing distinct oscillatory modes embedded within the original signal. Following decomposition, the Hilbert transform algorithm processes each IMF component to generate corresponding Hilbert spectra containing instantaneous frequency information. The integration of all component spectra produces a complete Hilbert spectrum that characterizes the original signal's frequency content through instantaneous energy-frequency distributions.

The research further explores HHT-based transient power quality disturbance detection, detailing fundamental HHT principles and implementation methodologies for identifying multiple power quality disturbances. Algorithm implementation includes real-time detection mechanisms for disturbance onset/termination timing, duration calculation, and amplitude quantification. Simulation experiments validate the method's effectiveness in monitoring and identification systems for complex power quality scenarios.

In summary, this work provides detailed exposition of HHT principles, methodological implementations, and practical applications, accompanied by robust source code that serves as valuable reference material for researchers in related fields. The code implementation demonstrates proper handling of boundary conditions, sifting stopping criteria, and spectral aggregation techniques essential for accurate HHT analysis.