22 April 2024 in IEEE Internet of Things Journal
Efficient information extraction enhances condition monitoring and fault diagnosis for bearings. The graph model (GM) has been proven to be a practical approach to extracting signal information within the temporal dynamic of frequencies. This article proposes an early fault detection method based on graph entropy (GE) for the dynamical bearing degradation process, considering the structure differences between graph dynamic changes. First, the complete GM (CGM) is constructed by a short-time spectrum generated from the original signal. In the fault detection phase, the GE, highly correlated with the health condition, is extracted from the GM to check any change in the machine state. Subsequently, the adaptive threshold of short-term month-over-month is used to judge the final decision making in an automated way. Finally, the validation experiment on the XJTU-SY data set and FEMTO-ST data set, as well as compared with the state of the art demonstrates its excellent detection performance. The proposed method extracts an effective 1-D index, which affords an excellent detection ability on early fault occurring in noisy environments, indicating a good potential for identification in the practical dynamic operation of engineering applications.
17 May 2025 in Engineering Applications of Artificial Intelligence
17 May 2026 in Measurement
1 January 2025 in Shock and Vibration
The gearbox is widely used in various machines, and their fault diagnosis can reduce unexpected downtime and maintenance costs. Although deep learning networks (DLNs) can automatically extract features from original series data and perform diagnostic analysis, their ability to distinguish complex or imbalanced fault patterns may be unsatisfactory. This study introduces a fault diagnosis technique that integrates wavelet packet decomposition (WPD) with a long short‐term memory (LSTM) enhanced by an attention mechanism. The fault information at different scales is extracted using WPD, and the wavelet packet energy is employed as input data for the LSTM network. Subsequently, the LSTM network searches for the long‐term dependence between the fault feature information to represent the fault state. An attention mechanism is introduced, allowing the network to selectively emphasize important time steps in the input data by applying varied levels of importance, which enhances the representation of critical features. Finally, the effectiveness of the proposed approach is validated on a standardized dataset, with experimental outcomes exceeding 98.2% average precision. This indicates that the proposed approach effectively utilizes both feature extraction based on expert knowledge and DLN.