Table 7 Comparison of the proposed study with previous studies.
From: An efficient bearing fault detection strategy based on a hybrid machine learning technique
N | References | Methodologies | Accuracy (%) |
|---|---|---|---|
1 | Long Wen et al.38 | Signal-to-image transformation, Transfer learning, TCNN(ResNet-50) structure | 98.95% ± 0.0074 (bearing damage dataset), 99.99% ± 0 (motor bearing dataset), 99.20% ± 0 (self-priming centrifugal pump dataset) |
2 | Liangsheng Hou et al.39 | IFMs-based deep ResNet, signal-to-IFMs method | 99.7% (artificial damage), 99.7% (real damage), 99.81% (mixed damage) |
3 | Qiumin Wu et al.40 | FSWT, deformable convolution, DC-ResNet | 93.90% (printing press bearings under actual working conditions) |
4 | Zia Ullah et al.41 | Transfer learning-based VGG-16 network, irreversible-demagnetization fault (IDF), bearing fault (BF) | 96.65% |
5 | Yuxing Li et al.42 | Genetic Algorithm (GA) optimized Variational Mode Decomposition (VMD), Center Frequency-based feature extraction | Not specified |
7 | Yuanhang Chen et al.43 | Deep Inception Net with Atrous Convolution (ACDIN) | 95% |
8 | Yuanhang Chen et al.44 | Hybrid deep-learning model (CNN and gcForest) based on continuous wavelet transform (CWT) | Not specified |
9 | Proposed method | DL Models and ML | 95.51 |
