INVESTIGATION OF PRESTRESSED CONCRETE FRAME BEHAVIOR WITH NEURAL NETWORKS

To date, concrete structure modeling has involved the development of m athematical models of concrete structure behavior derived from human o bservation of, and reasoning with, experimental data. An alternative, discussed in this paper, is to use a computation and knowledge represe ntation paradigm, called neural networks, developed by researchers in connectionism (a subfield of artificial intelligence) to model concret e structure behavior. The main benefits in using a neural-network appr oach are that all behavior can be represented within a unified environ ment of a neural network and that the network is built directly from e xperimental data using the self-organizing capabilities of the neural network, i.e., the network is presented with the experimental data and learns the relationships between stresses and strains. Such a modelin g strategy has important implications for modeling the behavior of mod ern, complex concrete structures. In this paper, the behavior of prest ressed concrete frames is modeled with a back-propagation neural netwo rk. The preliminary results of using networks to model concrete struct ure look very promising.