Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CEEMDAN-Based Multiscale CNN for Wind Turbine Gearbox Fault Detection

Nejad Alagha, Anis Salwa Mohd Khairuddin, Obada Al-Khatib, Abigail Copiaco

TL;DR

This paper addresses gearbox fault detection in wind turbines by tackling non-stationary vibration signals with a hybrid CEEMDAN–MSCNN framework. CEEMDAN decomposes signals into intrinsic mode functions (IMFs) across multiple time-frequency scales, which are then processed by a multiscale CNN with parallel branches to extract scale-specific features for robust healthy/faulty classification. The approach achieves high diagnostic accuracy (F1 up to 0.9895 on 0.25 s segments) while maintaining favorable computational efficiency, outperforming several baselines in training speed and offering practical potential for edge deployment. These results demonstrate the viability of end-to-end, multiscale deep learning on CEEMDAN-derived features for reliable wind turbine gearbox fault diagnostics in real-world settings.

Abstract

Wind turbines play a critical role in the shift toward sustainable energy generation. Their operation relies on multiple interconnected components, and a failure in any of these can compromise the entire system's functionality. Detecting faults accurately is challenging due to the intricate, non-linear, and non-stationary nature of vibration signals, influenced by dynamic loading, environmental variations, and mechanical interactions. As such, effective signal processing techniques are essential for extracting meaningful features to enhance diagnostic accuracy. This study presents a hybrid approach for fault detection in wind turbine gearboxes, combining Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and a Multiscale Convolutional Neural Network (MSCNN). CEEMDAN is employed to decompose vibration signals into intrinsic mode functions, isolating critical features at different time-frequency scales. These are then input into the MSCNN, which performs deep hierarchical feature extraction and classification. The proposed method achieves an F1 Score of 98.95\%, evaluated on real-world datasets, and demonstrates superior performance in both detection accuracy and computational speed compared to existing approaches. This framework offers a balanced solution for reliable and efficient fault diagnosis in wind turbine systems.

CEEMDAN-Based Multiscale CNN for Wind Turbine Gearbox Fault Detection

TL;DR

This paper addresses gearbox fault detection in wind turbines by tackling non-stationary vibration signals with a hybrid CEEMDAN–MSCNN framework. CEEMDAN decomposes signals into intrinsic mode functions (IMFs) across multiple time-frequency scales, which are then processed by a multiscale CNN with parallel branches to extract scale-specific features for robust healthy/faulty classification. The approach achieves high diagnostic accuracy (F1 up to 0.9895 on 0.25 s segments) while maintaining favorable computational efficiency, outperforming several baselines in training speed and offering practical potential for edge deployment. These results demonstrate the viability of end-to-end, multiscale deep learning on CEEMDAN-derived features for reliable wind turbine gearbox fault diagnostics in real-world settings.

Abstract

Wind turbines play a critical role in the shift toward sustainable energy generation. Their operation relies on multiple interconnected components, and a failure in any of these can compromise the entire system's functionality. Detecting faults accurately is challenging due to the intricate, non-linear, and non-stationary nature of vibration signals, influenced by dynamic loading, environmental variations, and mechanical interactions. As such, effective signal processing techniques are essential for extracting meaningful features to enhance diagnostic accuracy. This study presents a hybrid approach for fault detection in wind turbine gearboxes, combining Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and a Multiscale Convolutional Neural Network (MSCNN). CEEMDAN is employed to decompose vibration signals into intrinsic mode functions, isolating critical features at different time-frequency scales. These are then input into the MSCNN, which performs deep hierarchical feature extraction and classification. The proposed method achieves an F1 Score of 98.95\%, evaluated on real-world datasets, and demonstrates superior performance in both detection accuracy and computational speed compared to existing approaches. This framework offers a balanced solution for reliable and efficient fault diagnosis in wind turbine systems.
Paper Structure (13 sections, 5 equations, 2 figures, 4 tables, 1 algorithm)

This paper contains 13 sections, 5 equations, 2 figures, 4 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methodology
  4. Dataset Description
  5. Data Pre-processing
  6. CEEMDAN-MSCNN Approach
  7. Signal Processing using CEEMDAN
  8. MSCNN Architecture
  9. Results and Analysis
  10. Impact of CEEMDAN parameters on training-validation
  11. Algorithm Evaluation Across Different Signal Durations
  12. Comparative Analysis with Existing Methods
  13. Conclusion

Figures (2)

  • Figure 1: The architecture of the proposed CEEMDAN-based Multiscale CNN model for wind turbine gearbox fault detection.
  • Figure 2: Performance Metrics Across Different Signal Durations