In recent years, stainless steel columns have found widespread application in engineering while their reliability and safety are not well understood. This paper adopted an improved probabilistic model suitable for reliability analyses of stainless steel columns under axial compression. A database comprising 342 austenitic and 163 duplex stainless steel tubular columns under axial compression was established by collecting experiment results conducted in the past. By utilizing this extensive database, a thorough statistical analysis was performed to identify the most optimal probability distribution that can effectively fit the uncertainty of resistance model theta R and its associated distribution parameters. Subsequently, reliability analyses were conducted to evaluate the reliability levels of European, American, Chinese, and Australian/New Zealand stainless steel codes. The results indicate that European code exhibits the highest reliability index, followed by the Chinese code. The Australian/ New Zealand code ranks third, while the American code demonstrates the lowest reliability. The bias and variability of theta R , material strength and geometric dimensions were considered in reliability analyses, as well as load combination and load ratio. Additionally, a sensitivity analysis was conducted to study the effects of these variables on reliability index. Finally, based on the findings showing that the Chinese code cannot meet the target reliability index [ beta ] CN = 3.2, the resistance partial coefficient of material gamma R = 1.165 in the prediction model of bearing capacity for duplex stainless steel columns was modified to 1.32. The research results of this paper can provide reference for more accurate reliability analysis of stainless steel columns.
