Fragility curves are one of the useful tools to achieve probabilistic eva‎luation of seismic structures. Fragility curves show the potential vulnerability of structures at risk of earthquakes as the probability of exceeding the desired damage range in terms of seismic parameters. They are used for possible analysis of structures at earthquake risk, engineering decisions by estimating damage costs, and eva‎luation of various reinforcement scenarios regarding the final performance of structures. Fragility curves are generated by conventional methods such as Monte Carlo, Cornell and HDMR. To draw the fragility curve with high accuracy in the basic Monte Carlo method, many random numbers and analyzes are required. In the Cornell method, due to the selection of a limited number of random parameters, the accuracy of the fragility curve decreases, and also in the HDMR method, the acceleration uncertainty parameter is selected at all the same performance levels and increases the error, especially at higher performance levels. As a result, to produce brittleness curves at different levels, it is necessary to select the selected points appropriate to each functional level to increase the accuracy of the curve created at each functional level. The novelty of this study is the development of the SI-HDMR method for the production of Fragility curves of steel structures, so that with a smaller number of analyzes than the Monte Carlo simulation method, seismic eva‎luation with sufficient accuracy on this structure Performed in terms of random parameters. For this purpose, 4- and 9-story steel structures in the city of Los Angeles were selected and modeled in OpenSees software in a concentrated hinge method. Then, to accelerate the analysis of time history, 15 Accelerogram were used according to the construction conditions. Finally, the Fragility curve at different performance levels for 4- and 9-story steel structures was generated by SI-HDMR method and the results were compared with Cornell and HDMR methods. The results showed that the SI-HDMR method is in good agreement with the basic Monte Carlo method. Also, the error rate in the production of fragility curves by the proposed SI-HDMR method has been reduced at all different performance levels compared to other methods. Also, the average error at all performance levels is about 7%.

پيام اسدي

حسين تاجمير رياحي، فرهاد بهنام‌فر