Applications of Soft Computing Techniques in Structural System Identification

Abstract eng:
The seismic behaviour of masonry structures strengthened with fibre-reinforced polymer (FRP) materials has received very little attention experimentally and theoretically. The non-linear nature of these systems often results in mechanical responses that are difficult to predict via classic analytical methods. A neural network (NN) approach for dynamic system identification is presented here. This method addresses aspects such as system non-linearity, dependence on past loading history and noise contamination. Full-scale seismic tests conducted at Bristol University provided a large dataset of measured and computed dynamic state variables. The NN is capable of predicting the system response under a wide range of seismic inputs and for various user-specified reinforcement ratios. The results indicate that the NN non-parametric approach has an important potential in dynamic system identification.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket