Research Article
Sensor Fault Detection and Diagnosis Method for AHU Using 1-D CNN and Clustering Analysis
Table 4
Detection efficiency of the proposed method and other three existing neural network-based methods.
| Type | Bias | Index | BSC-NN (%) | AUX-NN (%) | CB-NN (%) | Proposed (%) |
| Fixed bias of Twr sensor | 1.0 | MAR | 0.00 | 0.00 | 0.00 | 0.00 | Fixed bias of Tws sensor | 0.5 | MAR | 0.00 | 0.00 | 0.00 | 0.25 | Fixed bias of Mw sensor | −5% | MAR | 34.60 | 0.00 | 0.00 | 1.98 | Positive fixed bias of Tsup sensor | 1.5 | MAR | 10.20 | 4.20 | 2.50 | 0.00 | Negative fixed bias of Tsup sensor | −1.5 | MAR | 50.80 | 0.00 | 0.00 | 0.00 | Normal | — | FAR | 0.00 | 2.44 | 3.25 | 3.16 | Average | — | — | 15.93 | 1.11 | 0.96 | 0.90 |
|
|