INVESTIGATION OF WELDING EFFECT ON REBARS USING NEURAL NETWORKS

Typically material modeling has involved the development of mathematic al models of material behavior derived from human observation of exper imental data. An alternative procedure, discussed in this paper, is to use a computation and knowledge representation paradigm, called a neu ral network, to model material behavior. The main benefits in using a neural network approach are that all behavior can be represented withi n the unified environment of a neural network and that the network is built directly from experimental data using the self-organizing capabi lities of the neural network, meaning that the network is presented wi th the experimental data and learns the relationships between stresses and strains. Such a modeling strategy has important implications for modeling the behavior of complex materials. In this paper, the mechani cal behavior of rebars affected by welds is modeled with a back-propag ation neural network. The results of using networks to study the effec t of welds on rebars look very promising.