Using artificial neural networks in the design of orthotropic bridge

Welded tubular connections are widely used in most steel platform structures. As a result of cyclic wave loading, the fatigue strength of welded tubular joints becomes one of the most important factors for safeguarding the structural integrity of these structures. The most effective factor in the behavior of the joints is the significant stress concentration at the weld toe. In order to quantify the stress concentration, an important parameter called the Stress Concentration Factor (SCF) is used. With the advances in modern computing techniques, finite element analysis has become a practical tool for engineering analysis and design which can be used to satisfy the SCF. Development of structural design code equations requires a wide range of experimental studies which is costly, timeconsuming and uneconomical. That¿s why, this problem can be simplified using COMSOL, which is a finite element analysis program.