Deficiencies of the normal moment yield criterion for RC slabs

Design methods for the ultimate load design of slabs, for example yield line analysis, Hillerborg's strip method and the Wood-Armer equations, require a yield criterion. The criterion usually used is the 'stepped' or normal moment criterion. The sections must also be ductile. A substantial number of experimental tests and theoretical studies have been carried out to verify the criterion for relatively low percentages of reinforcement. However, more recently, experimental studies have shown that the criterion may be unconservative for cases of pure twisting, even with relatively low areas of reinforcement. This paper examines the reasons for the unconservatism. Numerically determined yield criteria are given for a number of isotropically reinforced sections and the loading conditions for which the criterion is unconservative are identified. The implications for current methods of slab design and assessment are discussed. In order to provide safe solutions the sandwich approach method proposed by Morley, combined with the Clark-Nielsen equations, is considered to be a suitable solution.