Structural Dynamics Model Identification Using Heuristic Search

An application of artificial intelligence methodology, such as object-oriented programming and knowledge organization, to the problem of identifying time-varying structural dynamic systems has been presented in this paper. Possible changes in boundary conditions or configurations of a structure are organized into a taxonomy of models, and a variation of best-first, heuristic search is used to find the model whose simulated response best matches that of the current physical structure by using an output error approach in a discontinuous model space and an equation-error approach in parameter space. The unique advantages of some A1 methods for implementing k nowledge structuring a nd inheritance are discussed. An example problem is implemented in which both the time-varying model and its new parameters are identified when changes occur.