Predicting Carbonation Depth of Prestressed Concrete under Different Stress States Using Artificial Neural Network

Table 3

Normalized input/output pairs and the training and testing results.

Number	Input					Output/DC(mm)
	SL	TA	W/C	C/FA	C/CA	Experiment	BPNN	RBFNN
Training
A1	−0.309	0	0.4	0.669	0.410	23.20	23.47	23.57
A3	−0.309	1.0	0.4	0.669	0.410	34.45	34.62	34.10
A5	−0.515	0.5	0.4	0.669	0.410	26.38	26.38	28.23
A6	−0.515	1.0	0.4	0.669	0.410	31.56	32.17	31.37
A7	−0.721	0	0.4	0.669	0.410	19.95	21.22	20.38
A8	−0.721	0.5	0.4	0.669	0.410	25.13	24.62	25.84
B10	0.313	0.25	0.4	0.669	0.410	29.60	30.07	29.28
B12	0.563	0.25	0.4	0.669	0.410	30.20	30.35	29.92
B13	0.563	1.0	0.4	0.669	0.410	39.10	39.03	38.54
B15	0.701	1.0	0.4	0.669	0.410	40.50	40.43	40.22
C17	0	0.25	0.4	0.669	0.410	29.10	29.80	28.87
C19	0	1.0	0.4	0.669	0.410	36.45	36.23	36.69
Testing
A2	−0.309	0.5	0.4	0.669	0.410	31.90	28.43	30.47
A4	−0.515	0	0.4	0.669	0.410	22.50	22.31	22.06
A9	−0.721	1.0	0.4	0.669	0.410	29.19	31.21	28.50
B11	0.313	1.0	0.4	0.669	0.410	37.50	37.15	37.69
B14	0.701	0.25	0.4	0.669	0.410	31.15	30.24	31.48
C16	0	0	0.4	0.669	0.410	25.45	27.26	24.73
C18	0	0.5	0.4	0.669	0.410	31.92	31.10	32.43