A New Procedure for Damage Assessment of Prestressed Concrete Beams Using Artificial Neural Network : Table 3

Table 3: Summary of the results for various input cases.


Input parameter to the ANN	Number of inputs	Expected damage level (%) ED	Minimum predicted damage level		Maximum predicted damage level		Average predicted damage level		No. of epochs taken for training the ANN	Performance accuracy
Input parameter to the ANN	Number of inputs	Expected damage level (%) ED	Predicted damage value (OD)	Difference (ED–OD)	Predicted damage value (OD)	Difference (ED–OD)	Predicted damage value (OD)	Difference (ED–OD)	No. of epochs taken for training the ANN	Performance accuracy

p-f	2	40 50 70	32.19 33.37 68.09	−7.81 −16.63 −1.91	47.98 66.10 72.17	+7.98 +16.10 +2.17	37.69 57.56 70.40	−2.31 +7.56 +0.40	1203	1
s-f	2	40 50 70	33.01 31.77 58.98	−6.99 −18.23 −11.02	52.95 77.68 77.68	+12.95 +27.68 +7.68	42.05 48.63 71.44	+2.05 −1.37 +1.44	423	0.06
f	1	40 50 70	20.13 52.59 57.60	−19.87 +2.59 −12.40	58.77 58.82 59.00	+18.77 +8.82 −11.00	35.82 56.09 59.00	−4.18 +6.09 −11.00	5000	0.25