ABSTRACT
ABSTRACT. Increased awareness of the consequences of uncontrolled leakage of hazardous liquids to the environment, and of the ingress of contaminating sub stances from the environment into a retaining structure, has renewed the engineer’s interest in numerical models and design procedures to predict cracking of concrete structures and leakage through penetrating cracks. Engineering models for crack pattern prediction in cylindrical walls, submitted to combined hydrostatic and ther mal loads, are dealt with in this paper. The crack pattern comprises the width and depth of the cracks and the number of (vertical) cracks. Crack pattern predictions, in which the effect of thermal cycling has been accounted for, have revealed to be quite reliable. Less reliable are numerical predictions of leakage through cracks in real structures! These predictions are very sensitive to variations in the crack width. These variations can be quite large when alternating thermal loads, shrinkage and creep effects are involved, as is often the case in real structures. Conclusions are drawn as regards reliable and economic solutions for ensuring a liquid tight beha viour. It is emphasized, that for economic solutions one has to consider the "degree of tightness" that has to be designed for. Initial investments, required to ensure tightness, are compared with the costs for repairing leaking cracks.
