Damage Identification of a Steel Frame Based on Integration of Time Series and Neural Network under Varying Temperatures

Table 4

The identification results of damage severities when one brace is damaged.

Damaged brace	aj (%)	T (°C)	Target output	Actual output	Relative error (%)
1	50	20	0.5	0.485	−2.99
		30	0.5	0.507	1.46
	70	20	0.7	0.728	3.99
		30	0.7	0.757	8.11
4	50	20	0.5	0.500	0.09
		30	0.5	0.519	3.89
	70	20	0.7	0.742	6.04
		30	0.7	0.779	11.31
5	50	20	0.5	0.455	−8.91
		30	0.5	0.457	−8.63
	70	20	0.7	0.734	4.87
		30	0.7	0.718	2.57