Title:
Analysis of concrete fatigue durability by the neuro-fuzzy network FWNN
Author(s):
M. Jakubek, Z. Waszczyszyn
Publication:
KILW
Volume:
53
Issue:
1
Appears on pages(s):
3-22
Keywords:
fatigue,durability
DOI:
Date:
1/1/2007
Abstract:
The paper is related to [2] where the standard BPNNs (Back-Propagation Neural Networks) were applied to the analysis of concrete fatigue durability. Failure is related to the number N of compressive load cycles causing fatigue damage of laboratory specimens. About 450 results on laboratory concrete specimen tests were taken from [1]. The main goals of the paper are to improve the neural approximation performed by the standard BPNN and to extend neural simulation also for data given in intervals of concrete strength and cycle frequencies. That is why a neuro-fuzzy NN called for short FWNN (Fuzzy Weight NN) was applied. This approach enables us to be closer to the experimental reality.