The analysis of slender reinforced concrete columns is complicated because the
non-linearities o f the materials (caused by the cracking o f concrete and time-dependent
effects) are combined with the geometric non-linearity which characterises the behaviour
of such columns.
A simple, analytical method based on a graphical technique, originally proposed by
B eal, is developed. The method takes account o f the material and geometric
non-linearities and allows rapid and accurate analysis o f slender pin-ended reinforced
concrete columns, concentrically or eccentrically loaded, without the need for iterative
procedures or simplification o f section behaviour. The method allows for sustained load
effects and enables the reduction in the short-term ultimate capacity to be predicted.
The theoretical analysis is backed by nineteen short and long-term full-scale tests
on pinned reinforced concrete columns having slenderness ratios between 18 and 63,
loaded eccentrically. The experimental results substantiated the fact that instability is the
primary failure criterion for slender columns; and it occurs at relatively low compressive
concrete strains of the order o f 0.001-0.002. Material failure eventually follows, but for
slenderness ratios o f 33 and above this requires considerable bending to occur.
Creep was found to strongly influence the buckling load; with a sustained load of
60% of short-term capacity, creep causes a considerable reduction in the load capacity of
a slender column and can be as much as 40%. Initial imperfections are inevitable during
column construction. This was accounted for in the theoretical approach and the results
obtained demonstrated the sensitivity of slender columns to such imperfections.
The accuracy of the proposed method is further established when comprehensive
comparisons are made with the experimental work o f other investigators. Significant
improvement is noted over the existing design methods in BS8110 and ACI318. The
design procedures adopted in the two Codes are based on strength calculations which are
proved to be almost irrelevant for slender columns. The BS8110 approach was found to
be unsafe in predicting the long-term buckling loads o f the columns tested because it
does not make allowance for creep effects. This is in contrast to ACI318
recommendations which were found to be conservative in predicting the failure loads.