Scielo RSS <![CDATA[Journal of the South African Institution of Civil Engineering]]> vol. 50 num. 1 lang. es <![CDATA[SciELO Logo]]> <![CDATA[<b>Why stabilise? Using triaxial tests for determining pavement stiffness and shear strength parameters of mechanically modified layers</b>]]> Triaxial testing of naturally occurring, slightly silty, medium-graded, coarse sand, derived from completely weathered granite (with some gypsum), compacted to 95 % of Mod AASHTO density generated unload/reload Young's modulus E-values of about 300 MPa at a representative confining stress of 90 kPa. This is some 80 % higher than what would usually be expected for this type of G5 material. Mohr-Coulomb shear strength parameters of c = 15,9 kPa and φ = 51,4° were obtained from the high-quality triaxial tests. Taking into account the variability of the materials, it is suggested that these be downgraded to c = 12,7 kPa and φ = 48,8° for use as design parameters. It is common practice to stabilise natural gravel materials to generate weakly cemented sub-base layers. However, the presence of naturally occurring gypsum within the in-situ granite generated concern as negative effects were observed on elements of past construction. The use of cement as a stabilising agent was eliminated by generating a nearly equally strong layer by mixing the naturally occurring gravels with varying quantities of crushed stone, crusher waste and dune sand, the latter to combat a high plasticity index. The best result was obtained by using a blend of 50 % natural gravel, 30 % crusher waste and 20 % dune sand. These blended materials, generated unload/reload Young's modulus E-values of some 560 MPa at 95 % Mod AASHTO compaction and a confining stress of 247 kPa. This E-value is very similar to what is thought would be attained for a cracked and hydrated cement-stabilised layer but without the disruptive effects of a lowered confining stress that would be the case when a stabilised layer shrunk and cracked on hydration of the cement stabilising agent. <![CDATA[<b>Finite element fracture modelling of concrete gravity dams</b>]]> A smeared crack model, based on non-linear fracture mechanics, was developed which allows for either linear or bilinear softening and assumes shear retention dependent on the strain normal to a crack. A mesh objectivity verification study proves that the proposed crack modelling method is mesh objective. The crack model and its computational procedure is verified for a benchmark concrete gravity dam model and an existing concrete gravity dam by comparing the results with those of numerical investigations obtained by other researchers. Furthermore, an existing concrete gravity dam in South Africa is analysed and evaluated with regard to dam safety in terms of the maximum overflow level. A higher imminent failure flood is predicted in the analysis than that obtained by classical strength-based methods. The study proves the usefulness and applicability of the proposed crack model and implementation procedure in predicting crack response and evaluating the safety of concrete gravity dams. A sensitivity study on the material fracture properties and fracture parameters is included for the purpose of investigating the uncertainties often encountered in this type of analysis. The influence of the fracture properties and parameters on the cracking response and the overall structural behaviour is discussed. <![CDATA[<b>The South African durability index tests in an international comparison</b>]]> Over the last decade, an approach to improving the durability of reinforced concrete construction has been developed in South Africa. The durability index test methods applied in this approach are unique to the country. Internationally, similar trends can be observed and a number of test methods for the quality assessment of hardened concrete, particular the cover zone have been developed world-wide. A group of researchers from different parts of the world carried out comparative tests on the quality of concrete cover, applying the most commonly accepted international test methods. The results indicate that the South African oxygen permeability and chloride conductivity test methods are successful in characterising the concrete cover in respect of important deterioration mechanisms. A good correlation exists between results obtained from these index test methods and those obtained from related test procedures applied overseas.