engleski

STRENGTH AND DEFORMATION CAPACITY OF STRUCTURAL ELEMENTS UNDER HORIZONTAL LOADING

Understanding the behavior of structural elements under in-plane horizontal loading is essential for any performance based design procedure. That applies especially to columns and walls as main lateral load resistingelements. There are many parameters that influence the prediction of the elements behavior, especially at theultimate stage. The outlined problem presents an ideal example where neural networks (artificial intelligence)can be used. An extensive library of tested columns and walls has been collected. Work on that database considers devising aprotocol of presenting the research data in the performance form. The relationship between qualitativeperformance description and engineering parameters that can be considered in design has been established. Theuse of neural networks, taught on the collected data base, enables development of improved procedures for assessment of strength and deformation capacities of the columns and walls at all performance levels. It enablesprognostic behavior of the lateral load resisting elements under horizontal in-plane loading. The element geometry and material data are used for input and performance behavior of the elements is obtained as theoutput of such an expert system that can be used as problem solver. The work on the collected data base increases the understanding of elementbehavior, their performance at various demand levels and eliminates complicated detailing that were encouraged by lack of knowledge. That isin direct support of development a multi level performance design methodology for structural systems.

Neural networks; Performance based design procedure; Seismic capacity; Reinforced concrete columns and walls; Experimental database ;

nije evidentirano