| In the analysis and design of steel structures, it is vital to know the accurate stress-strain relationship of steels. Modeling the accurate material propertied required accurate modeling of the stress-strain relationship of the material. Generally, the stress-strain relationship of steels was represented by Ramberg-Osgood relationship. However, the use of Ramberg-Osgood relationship for stress higher than the yield stress can lead to serious inaccuracy. In the present study, a number of test data of tension coupon tests in the existing literature were collected and analyzed. Two stress-strain models were proposed for light gauge steels and all carbon steels respectively based on these test data. In addition, a stress-strain model was also proposed for the carbon steels with a yield plateau in the present study. These proposed models were all estimated by only three basic Ramberg-Osgood parameters. This result is particularly useful for the design and numerical modeling of structural members and elements where the stress-strain relationship is specified in terms of the three basic Ramberg-Osgood parameters. These proposed models were based on a careful interpretation of existing experimental data. The main work and conclusion ars described below.1. A large number of stress-strain data obtained from tensile coupon tests reported in existing literature have been collected and analyzed. These coupons were cut from carbon steel sheets and the flat portions of cold-formed and hot-formed carbon steel sections with various steel grades.2. The thesis presents new stress-strain models for light guage steels and all carbon steels respectively. It was found from the comparison between the models and the test stress-strain curves that, these models can generally describe the full range stress-strain relationship of steels without a yield plateau accurately, especially the stress-strain behavior of steels after yielding. In addition, a new stress-strain model was also proposed to describe the full range stress-strain relationship of steels with a yield plateau. The models proposed in the present study were all estimated by only three basic Ramberg-Osgood parameters, which was very convenient for use.3. It was found that the accuracy of the full-range stress-strain models highly depend on the accuracy of the ultimate stress and the ultimate strain. Thus, the effect of variation in predicted the ultimate stress and the ultimate strain on the proposed stress-strain models of light gauge steels, all carbon steels and steels with a yield plateau were given and discussed based on parameter analysis, especially the effect on the upper bound strain (i.e. the total strain 0.02) for general structural response. |