| Reinforced concrete, as a typical composite material easy to draw and with high compression strength, has been applied extensively in engineering since exploited in 19 th century. Compared with the steel structural calculating theory, the experimental data of reinforced concrete is discrete, with immature RC calculating theory and diverse RC codes in different countries. Except for the unified recognition of theories about axial tense and pure bending, there has been no agreement on the research of initial theory of torsion and shear and most codes adopted the calculating formulas based on experiments. Among the diverse theories, the truss model theory is more mature and capable of the overall process analysis of RC members, but it has too complex formulas and is difficult to be applied to engineering practice. The ultimate equilibrium theory has successfully solved problems of RC members under pure bending and pure shear. With simple formulas and clear concept, it can be well approvable to engineering practice if well improved to solve problems of RC members under bending, shear and torsion.This paper designed seven rectangular RC members under various combinations of bending, shear and torsion. The deforming progress, destruction form and bearing capacity of RC members were described. On the basis of experiments, initial stress mechanism analysis and the ultimate analysis of three-dimension inclined torsion destruction surfaces, this paper has improved the ultimate equilibrium theory and developed a series of formulas for the failure strength of RC member under bending, shear and torsion. The theoretical solutions were verified with the experimental results of seven RC members as well as other experimental results and both showed a good coordination. The comparison was also made between the theoretical solutions and solutions by the equations in Chinese code GB 2010-50010. The theoretical solutions turned out more agreeable to the experimental results. In the end, this paper built numerical models in ANSYS to expand the research parameters. |