Abstract
The paper describes the hybrid computer system dedicated to identification of models of materials subjected to thermomechanical processing. The functionalities of the system consist of plastometric tests data processing and application of the inverse analysis. The latter functionality is realized unconventionally, instead of the finite element method the metamodel is implemented using artificial neural network. The metamodels, used for simulations of the plastometric tests, are imported to the proposed computer system as external plugins, what guarantees flexibility and possibility of further development. On the other hand, application of rich optimization libraries assures the best possible solution of the problem. Basic principles of the inverse analysis with metamodels and mentioned optimization procedures are described in the paper. Selected examples of identification of models for various metallic materials recapitulate the paper.
HYBRID COMPUTER SYSTEM FOR THE IDENTIFICATION OF METALLIC MATERIAL MODELS ON THE BASIS OF LABORATORY EXPERIMENTS
