Research Article

Damage Identification for Large Span Structure Based on Multiscale Inputs to Artificial Neural Networks

Table 19

The identification errors of forecasted cases based on acceleration damage parameters.

NumberDamage extentsLoad casesNoise levelsMaximum error
0.000.010.020.040.080.10

1 50.1470.0120.2090.1450.1180.2100.210
2 100.1080.1120.0770.2440.0860.0410.244
3 150.0020.0650.1590.0070.0540.2320.232
4 200.0710.2650.1040.1460.0060.0340.265
5 250.0180.2510.0680.0780.1010.0130.251
6 310.0110.2830.0290.1080.0610.1740.283