The structural permeability of concrete at cryogenic temperatures

The thesis describes a programme of research designed to identify concretes for application at cryogenic temperature, in particular for storage of Liquefield Natural Gas which is maintained at a temperature of -165oC. The programme was undertaken in two stages. Stage 1 involved screening tests on seventeen concrete mixes to investigate the effects of strength grade (and water/cement ratio), air entrainment, aggregate type and cement type. Four mixes were selected on the basis of low temperature strength, residual strength after thermal cycling and permeability at ambient temperature. In Stage 2 the selected mixes were subjected to a comprehensive range of tests to measure those properties which determine the leak tightness of a concrete tank at temperatures down to -165oC. These included gas permeability; tensile strength, strain capacity, thermal expansion coefficient and elastic modulus, which in combination provide a measure of resistance to cracking; and bond to reinforcement, which is one of the determining factors regarding crack size and spacing. The results demonstrated that the properties of concrete were generally enhanced at cryogenic temperature, with reduced permeability, reduced crack proneness and, by virtue of increased bond to reinforcement, better control of cracking should it occur. Of the concretes tested, a lightweight mix containing sintered PFA aggregate exhibited the best performance at ambient and cryogenic temperature, having appreciably lower permeability and higher crack resistance than normal weight concretes of the same strength grade. The lightweight mix was most sensitive to thermal cycling, but there was limited evidence that this behaviour would not be significant if the concrete was prestressed. Relationships between various properties have been identified, the most significant being the reduction in gas permeability with increasing strain capacity. The structural implications of the changing properties of the concrete have also been considered.

The susceptibility of various UK aggregates to alkali silica reaction

A group of lithologically varied UK aggregates have been incorporated into concrete prisms of variable alkali content to ascertain the alkali levels at which significant ASR first occurs at 38oC and 100% RH. Petrographical analysis was used to establish the source of reactivity. The results of these expansion tests showed that significant ASR can develop with certain aggregates at initial alkali levels as low as 3.5 kg/m3 Na2Oe. Similar prisms were made at initial alkali levels, well above, on and just below the alkali thresholds for each aggregate. These prisms were placed in salt solution to establish the effects of ASR. The results showed that an external source of NaCl does accentuate ASR in high alkali mixes. However, in low alkali mixes the ASR initiated was even greater than that developed by the high alkali mixes. It was proposed that an `initial alkali pessimum' existed for each aggregate type for specimens placed in salt solution. Electron microprobe analysis of the ASR gels from concretes immersed in salt solution, showed that two compositionally varied gel suites develop. The first suite was derived from ASR caused by the initial alkalis in a concrete mix and was identical to ASR gels derived from the various concretes when immersed in distilled water. The second suite was developed by alkalis derived from a reaction between NaCl and the C3A component of the cement paste. It was demonstrated that the `initial alkali pessimum' was probably due to a combination of these two ASR types at the alkali threshold point where both suites of ASR gel can develop. Equivalent mixes were made with a 25% replacement of the cement by pulverised fuel ash (pfa) to establish whether alkalis released from the pfa could initiate ASR in otherwise non-reactive low alkali mixes. The addition of air entrainment to reactive concrete mixes was also examined as a method of suppressing ASR.