عنوان مقاله [English]
In this study, the behavior of end plate connections with various rigid has been studied. After the creation of the finite element models. model accuracy is evaluated using the experimental results. Maximum error of finite element models generated is estimated at about 8 percent. With parametric analysis semi rigid connections, an important factor influencing the behavior of the connections is detected. The prying action is one of the most significant parameters in connections behavior; lack of attention to this phenomenon in the design phase will cause errors in calculation. To calculating the prying action force in T connections, two experimental methods exist. The computational method of Smith and AISC, the amount of computational error in smith method,is 80 to 160 percent depending rigidity of end plate connections. In second method, AISC, in connection with semi-rigid end plate, the error is estimated at about 40 percent. To understand how to set formulation AISC, models of the T Connection with variable end plate thickness is made and parametric analysis was performed. Failure mechanism of these connections is shown that in connection with flexible end plate, much of the prying action force is created. In connection with the very rigid plate can be neglected under the conditions of the effect of prying action forces. With comparing the stress distribution in flexible, rigid and T connections, the geometric coefficient of 3 row bolted end plate connections is considered 1 25 and in 2 row bolted 0.5 The maximum error of calculating the prying action force is about \%5. Using the improved method, in semi rigid connections, the need to calculate the prying action and no need to calculate it is characterized. Having the exact amount of prying action forces. animproved design method has been authorized based on the philosophy and methods of allowed stress. In this method, 22 percent reduction in design error has been observed.