The results of about 8000 measurements of crack width on eight composite T-beams subjected to negative (hogging) bending are analysed, and related to the established relationships that are the basis for current methods of crack width prediction for reinforced concrete beams, slabs, and tension members. Many departures from these relationships were found, especially at low tensile strains. The influence of shrinkage of concrete on crack width is greater than in reinforced concrete members. New methods were devised for predicting both mean and design crack widths. The observed crack widths are much greater than those predicted by the method of BS 5400: Part 5. This finding needs to be checked by study of continuous members in service. Professor R.P. Johnson and R.W. Allison
An approximate method is presented for the analysis of the distribution of lateral forces among the components of a 3-dimensional building structure that consists of assemblies of shear walls, coupled walls, rigidly jointed frames and cores, subjected io both bending and torsion. The load distribution on each assembly is assumed to be represented sufficiently accurately by a polynomial in the height coordinate, together with a concentrated interactive force at the top. The presence of the latter is essential for obtaining accurate results as simply as possible. A set of flexibility influence coefficients, relating the deflection at any level to any particular load component, is established for each assembly, the continuum approach being used to analyse individual cores and coupled shear walls, and the shear cantilever analogy for the frame elements. By making use of the equilibrium and compatibility equations at any desired set of reference levels, the load distribution on each assembly may be determined. Good results appear to be achieved.for regular structures by using no more than about six reference levels. Professor A. Coull and T.H. Mohammed
A deterministic approach was used to assess the strength of centrally loaded H-columns, taking into consideration the effects of residual stresses, initial out-of-straightness, and small end restraints. The column types used in the study include hot-rolled wide-flange shapes and flamecut H-shapes. The initial crookedness at midheight was taken as one thousandth of the length of the column. Five types of residual stress distribution were considered and four common types of beam-to-column connection were used as end restraints. Eric M. Lui and Professor Wai-Fah Chen