The fire resistance of high-strength (greater than 60 MPa) concretes has been reported to be reduced when compared to normal strength concrete. This behavior has been attributed to the very dense concrete matrix usually associated with high-strength concrete. This dense matrix does not allow rapid transmission of water vapor within the concrete exposed to high temperatures, thus leading to disruptive vapor pressures. This problem is aggravated further when the fire is a hydrocarbon fire which reaches 880 C within three minutes. Offshore concrete platforms, which are typically built with high-strength concrete, are therefore at risk from hydrocarbon fires. This paper presents the results of two test programs involving hydrocarbon fires and high-strength concretes. Both lightweight aggregate concrete and modified normal density concrete (blends of normal density and lightweight aggregate coarse aggregates) were evaluated. Small beams of lightweight concrete, using different lightweight aggregate types, were evaluated for spalling resistance. Polypropylene fibers, used in some beams, were successful in greatly reducing spalling. Large wall sections of modified normal density concrete experienced significant spalling, but retained adequate concrete strength behind the reinforcing bars because of the low rates of heat transfer within the large sections.

This paper deals with CANMET investigations on the performance of concrete, with and without supplementary cementing materials, in a marine environment. A series of more than 250 concrete prisms, 305 x 305 x 915 mm in size, were cast over a period of 16 years and installed at Treat Island, Maine, an outdoor exposure facility operated by the U. S. Army Corps of Engineers, U. S. A. The prisms of the first phase of the investigation were installed at the site in 1978, with the remaining specimens being installed at almost yearly intervals. The specimens of the latest phase were installed in 1994. The test prisms are installed at mid-tide level on a rack and are exposed to repeated cycles of wetting and drying and to about 100 cycles of freezing and thawing per year. The test specimens will be kept at the exposure site until at least the year 2005. The test prisms are evaluated annually. The evaluation includes visual examination and rating and ultrasonic pulse velocity testing. Also, a complete photographic record is kept. Some of the principal conclusions based upon up to 17 years of exposure of some of the test prisms are as follows. The use of non air-entrained concrete is not recommended for the exposure conditions experienced at Treat Island. For the exposure conditions experienced at Treat Island, the percentage of silica fume in concrete should be limited to 10 percent. Both the normal weight and semi-lightweight concretes incorporating fly ash or slag or silica fume or a combination of these materials are in good to excellent condition, provided water-to- cementitious materials ratio is kept below 0.50 and portland cement is kept at a certain minimum level. There is no significant difference in the performance of concrete made with ASTM Types I, II, and V cements.

The marine environment provides a severe durability test for reinforced concrete structures with premature deterioration often being associated with steel corrosion. The rate of chloride ingress from the sea through the concrete cover is of primary importance since the depassivation of steel and subsequent corrosion are largely controlled by the chloride concentration at the reinforcement. Accurate service-life predictions are made by defining the material, assessing the severity of exposure, and monitoring the durability performance in that environment. Concrete, therefore, needs to be characterized in terms of early age properties that control the diffusion of chlorides through the cover concrete. These characterized values may then be related to long-term characteristics which determine durability for different environmental conditions. Early age tests should only be used as indicators of potential durability once suitable relationships have been established with the durability performance of concrete under marine conditions. Results from a research program are presented, in which concrete specimens were initially characterized at 28 days before being exposed to four marine environments in South Africa. The concrete was tested using a newly developed chloride conductivity test which determined the chloride resistance of concrete using an accelerated technique. Chloride contents were measured after 24 months of exposure and the diffusion coefficients were related to the initial characterization values. Results indicated that the severity of exposure has a major influence on the relative rate of chloride ingress into the concrete. The chloride conductivity test was found to be a useful indicator of chloride resistance, but the results are specific to the type of concrete being tested. Comparisons of potential durability of concretes based solely on the results of rapid chloride tests at early ages may be misleading and should be used with caution.

SP-163 This specialized publication offers 24 papers presented at the Third CANMET/ACI International Conference held on August 4 -9, 1996 in New Brunswick, Canada on the subject of Performance of Concrete in Marine Environments.

Research on high-performance concrete structures in the field have revealed large variations in chloride penetration into the concrete and in corrosion rates of the reinforcement. High-performance concrete is considered to be highly resistant to the initiation of reinforcement corrosion. The rate of chloride transport decreases with time to extremely low values (10 -13/10 - 14 m 2/sec). However, since concrete often is cracked, the occurrence and effect of cracks are also important factors for corrosion initiation in chloride environment. The effect of cracks on the corrosion rate of steel in high-performance concrete is under study in a Swedish research program. Reinforced concrete slabs with cracks of different widths are placed on floating platforms in the sea; the lower part of each slab is submerged, while the upper part is exposed in the splash zone. Corrosion rates, half-cell potentials (CSE), and concrete resistivities are measured regularly. Parallel specimens are similarly exposed in saline solution in the laboratory. This paper presents interim results of this project. The results so far show little effect of crack width and concrete cover thickness. Addition of microsilica (silica fume) is positive, by increasing the concrete resistivity and, thus, reducing the rate of corrosion; the corrosion behavior of such concretes is close to passive, even with cracks.