A new modified sliding window empirical mode decomposition technique for signal carrier and harmonic separation in non-stationary signals: Application to wind turbines
| Autor: | Erik Etien, Sebastien Cauet, Anas Sakout, Laurent Rambault, Jack P. Salameh |
|---|---|
| Přispěvatelé: | Laboratoire d'Informatique et d'Automatique pour les Systèmes (LIAS), Université de Poitiers-ENSMA, Laboratoire des Sciences de l'Ingénieur pour l'Environnement - UMR 7356 (LaSIE), Université de La Rochelle (ULR)-Centre National de la Recherche Scientifique (CNRS) |
| Rok vydání: | 2018 |
| Předmět: |
0209 industrial biotechnology
Signal processing [PHYS.PHYS]Physics [physics]/Physics [physics] Computer science Applied Mathematics 020208 electrical & electronic engineering 02 engineering and technology 7. Clean energy Fault detection and isolation Hilbert–Huang transform Computer Science Applications Mechanical system 020901 industrial engineering & automation Control and Systems Engineering Colors of noise Control theory Harmonics Sliding window protocol 0202 electrical engineering electronic engineering information engineering Time domain Electrical and Electronic Engineering Instrumentation ComputingMilieux_MISCELLANEOUS |
| Zdroj: | ISA Transactions ISA Transactions, Elsevier, 2019, 89, pp.20-30. ⟨10.1016/j.isatra.2018.12.019⟩ |
| ISSN: | 1879-2022 0019-0578 |
| DOI: | 10.1016/j.isatra.2018.12.019⟩ |
| Popis: | Modern control applications justify the need for improved techniques capable of coping with the non-stationary nature of measured signals while being able to monitor systems in real-time. Empirical Mode Decomposition (EMD) is known for its efficiency in time domain analysis of multi-component signals through Intrinsic Mode Functions (IMFs) extraction. Recent years witnessed the introduction of Sliding Window EMD (SWEMD) capable of analyzing signals in real time applications. However, complex signals require several sifting iterations while a rather increased number of IMFs might result in impracticality for on-line applications. This paper introduces a new modified faster SWEMD capable of extracting harmonics from non-stationary signals in real-time operation. The method uses the traditional EMD properties in the first pass for a small number of sifting processes. In addition, a new section is added to the algorithm based on inflection point tracking of the residue derivative from the first pass is added, in order to track low frequency waves and render the analysis faster. The method is validated for non-stationary signals with and without added colored noise and applied on measured turbine side angular velocity for harmonic extraction in wind turbines as an application. The proposed method may well be used for fault detection and disturbance rejection in mechanical systems. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect