A5012223467- Machine Fault Diagnosis Techniques
- Engineering Diagnostics and Reliability
- Fault Detection and Control Systems
- Advanced Sensor Technologies Research
- Imbalanced Data Classification Techniques
- Grey System Theory Applications
- Gear and Bearing Dynamics Analysis
- Mechanical Failure Analysis and Simulation
Nanjing University of Aeronautics and Astronautics
2022-2023
Abstract Long short-term memory (LSTM) based prediction methods have achieved remarkable achievements in remaining useful life (RUL) for aircraft engines. However, their performance and interpretability are unsatisfactory under complex operating conditions. For engines with high hazard levels, it is important to ensure the of models while maintaining excellent accuracy. To address these issues, an interpretable RUL method conditions using spatio-temporal features (STFs), referred as iSTLSTM,...
Currently, the monitoring of health status mechanical systems is becoming more and critical, actual data massive high-dimensional, these are characterized mainly by imbalance. To solve problem low fault recognition rate caused imbalance, this article proposed a novel local non-local information balanced neighborhood graph embedding interpretable deep autoencoder (LGBNGEDAE) method for rotating machinery diagnosis. Specifically, embedded into original objective function to smooth manifold...
Abstract Rotating machinery is widely used in industrial production facilities, and once a failure occurs, it can be catastrophic. Alerting to potential defects time prevent further equipment degradation challenging task. In this paper, novel two-stage fault warning framework proposed for early of rotating machinery. Specifically, new method based on intra-class inter-class neighborhood information graph embedding orthogonal discriminant projection firstly adopted extract the global...
T Conducting gearbox health status evaluation can effectively obtain the operating of equipment so as to develop maintenance plans. A condition method based on deep fuzzy clustering neural networks (DFCNN) is proposed. The proposed DFCNN incorporates high-level feature extraction layer Generalized Supervised Deep Autoencoders (GSDAE) into an improved C-Means (KMDFCM) algorithm. GSDAE used eliminate variation parameters due differences in conditions, and then KMDFCM for unsupervised...