Research Article
Damage Localization of Cable-Supported Bridges Using Modal Frequency Data and Probabilistic Neural Network
Table 8
Summary of correct identification for Ting Kau Bridge (TKB).
|
Noise level | Identification accuracy (IA) | Using 20 frequencies | Using 10 frequencies | Using 5 frequencies |
| ε = 100% | 46.23% | 33.35% | 27.90% | ε = 80% | 57.08% | 45.25% | 33.05% | ε = 60% | 62.55% | 51.85% | 43.40% | ε = 40% | 78.15% | 62.90% | 51.33% | ε = 20% | 84.33% | 71.70% | 58.75% | ε = 10% | 86.42% | 73.53% | 60.83% | ε = 8% | 87.63% | 75.08% | 63.15% | ε = 6% | 88.90% | 76.85% | 65.80% | ε = 4% | 89.32% | 77.23% | 67.58% | ε = 2% | 89.97% | 77.35% | 67.33% | ε = 1% | 90.00% | 80.05% | 67.70% |
|
|
Note: IA (identification accuracy) is defined as the ratio of total number of correct identification to total number of testing samples (4000).
|