Recommendations for the use of concretes having various shrinkage levels are given in Table 1. They are based on findings for a wide range of shrinkable concretes examined in the laboratory, coupled
with a more limited experience of concrete in the field. These recom- mendations cannot be regarded as final but offer some guidance until further experience is gained.
Appendix I
EXPERIMENTAL WORK ON
Shrinkage of rocks and gravels
In the dolerite group of rocks, the shrinkage of the rock on losing water varied between 0 and 0.065%. The shrinkage of dolerites was found to be related to the absorp tion of the rock; although there was considerable scatter of individual measurements around the main line, the shrinkage increased approxima tely linearly from 0 to 0.065% as the water absorption increased from 0 to 1.2%.
For rocks outside the dolerite group it was not possible to establish a relationship between shrinkage and absorption because of the wide variety of types encoun- tered and the few samples available that were representative of each type. However, the specimens examined had a comparatively low shrinkage with the exception of a felsite and certain mudstone speci- mens which had high shrinkages in the range 0.051-0.088%.
In one
laminated mudstone, the shrinkage at right angles to the bedding planes was six times greater than that parallel to them.
The shrinkage of the gravels was found
between 0 vary 0.075%.
to
Dense concrete specimens
and
The shrinkage of dense concrete made with various aggregates was found to increase as the shrinkage of the aggregate itself increased, and ranged from 0.03 to 0.10%. The relationship was observed to he linear and indicated that the shrink- age of a
concrete made with a shrinkable aggregate was at least 0.03% plus the shrinkage of the rock. Both fractions of the aggre- gate. i.e. the coarse and the fine, contributed to the shrinkage. but the coarse made the greater contribu- tion; in fact, it set the level of shrinkage to be expected.
116
SHRINKABLE AGGREGATES
By comparison with concrete made with ordinary Portland cement, the use of high-alumina cement reduced the shrinkage by about 10% but rapidhardening and Portland blast furnace cements had no such effect.
Water curing of the concrete specimens for 28 days was found to increase the shrinkage from 20 to 25% by comparison with specimens that were air cured for the same length of time. The shrinkage was found to increase slightly as the water/cement ratio and the amount of entrained air were increased, but, as noted later, the durability was improved by air-entrainment.
Although no relationship has been found between shrinkage and workability
a wide or strength. range of strengths was recorded in the many concretes tested despite the standard proportions and water/ cement ratio adopted.
Shrinkage of concrete specimens made with shrinkable aggregates has been shown to be about three times as great as the thermal movement induced by a rise in temperature of 60°F. and it has been concluded that though drying shrinkage has a far greater influence on the movement of concrete, the combined effect may well be important.
No-fines concrete specimens
The use of shrinkable instead of non-shrinkable aggregate to make no-fines concrete results in much greater shrinkage than occurs with dense concrete. When certain dolerite aggregates were used, the shrinkage of no-fines concrete
was
six to seven times that of a similar concrete made with non-shrinkable aggregate, whereas use of these aggregates in dense concrete only increased the shrinkage by three or four times. The results are given in Table 2.
Wetting and drying cycles
In addition to the test for shrink- age, some specimens of dense concrete were subjected to 25 cycles of wetting and drying. Measure. ments revealed that, with each cycle, the wet and dry lengths of all speci- mens made with shrinkable aggre- gates increased. This progressive lengthening varied with the type of cement used, being greatest in the case of high-alumina cement, and was from half to twice the recorded drying shrinkage of the specimens. On the other hand, specimens made with non-shrinkage quartz gravels decreased slightly in length over the first few cycles of wetting and dry- ing and then remained constant.
Cracking of exposed specimens
Cracking was attributed mainly to the shrinkage of the aggregate used and sometimes was seen to have begun within a few weeks of ex- posure. even before the onset of frost. In some cases the growth of the cracks appeared to be checked. possibly by the self-healing of the
concrete.
A large number of the specimens exposed eventually cracked, some disintegrated. It was not possible, however, to relate the extent of the cracking to the shrinkage level of the aggregates used; for instance, in a few cases extensive cracking oc- curred in specimens made with low- shrinking aggregate. It was noted that the tendency to crack depended more on the properties of the fine than the coarse fraction of the ag- gregate. For a limited number of tests this tendency increased as the grading of the fine fraction became
coarser.
Cracking was, on the average, slightly greater for ordinary Port- land and rapid-hardening Portland cements than for Portland blast furnace and high-alumina cements; even so, it was apparent that the aggregate had a much greater in- fluence than the cement in this respect. In concrete specimens made with an aggregate of medium shrinkage, the tendency for cracks to develop increased as the water/ cement ratio increased. At a water/ cement ratio of 0.70 the specimens disintegrated soon after exposure.
THE HONG KONG & FAR EAST BUILDER-VOLUME 18. NUMBER 2