ABSTRACT
This chapter discusses the commonly available neural network (NN) models for classification and regression. NN models are specified by the net topology, node characteristics, and training or learning rules. The chapter presents a detailed survey of applications related to civil infrastructure. Active structural health monitoring methods analyze the health of a structure by exciting the structure and recording the corresponding response. A computing device based on NN has a greater fault tolerance than a classical sequential computer due to the increased number of locally connected processing nodes. An NN must be trained with sufficient training data to diagnose damage correctly. The NN was used to identify the map from static strain data to a subjective measure of the damage. NNs are extensively used in signal processing because they provide nonlinear parametric models with universal approximation power as well as adaptive training algorithms.
