Research Article
Fault Diagnosis and Detection in Industrial Motor Network Environment Using Knowledge-Level Modelling Technique
Table 5
Performance of different architectures for classification.
| | Architecture | Sample | MSE performance | Number of epochs | Accuracy (%) | Classification error |
| | [4 × 3 × 3] | S1 | 6.69 × 10−3 | 70 | 92.5 | 7.5 | | S2 | 6.19 × 10−3 | 75 | 93.4 | 6.6 | | S3 | 7.37 × 10−3 | 67 | 92.6 | 7.4 | | S4 | 8.01 × 10−3 | 100 | 94.0 | 6.0 |
| | [4 × 10 × 3] | S1 | 8.49 × 10−3 | 117 | 96.2 | 3.8 | | S2 | 8.59 × 10−3 | 125 | 96.3 | 3.7 | | S3 | 8.72 × 10−3 | 132 | 97.4 | 2.6 | | S4 | 8.99 × 10−3 | 131 | 97.1 | 2.9 |
| | [4 × 15 × 3] | S1 | 8.01 × 10−3 | 329 | 94.1 | 5.9 | | S2 | 6.23 × 10−3 | 327 | 93.5 | 6.5 | | S3 | 7.85 × 10−3 | 346 | 83.1 | 16.9 | | S4 | 8.56 × 10−3 | 387 | 91.8 | 8.2 |
|
|
