Evaluation of fire damage to high performance concrete

Abstract

This thesis describes a study to evaluate and explain the nature of the damage caused by fire to high performance concrete (HPC) material. The study includes the two main parts, (1) experimental investigation of fire damage to HPC, and (2) the establishment of a mechanism explaining the fire damage to HPC, which takes into account the characteristics of the various material properties, the behavior of the concrete observed and the temperatures to which HPC is exposed. The investigation methodology involved a series of experiments, including fire resistance tests, spalling tests, mechanical property tests, crack growth resistance curve (R-curve) tests, quantitative X-ray diffraction (QXRD) tests, cracking observations, mercury intrusion porosimetry (MIP) measurements, scanning electron microscope (SEM) observations, and rebound hammer tests. Methods for testing spalling and establishing the R-curve test were developed during this investigation by the writer. Explosive spalling is the foremost type of damage that could be suffered by HPC due to fire. The spalling tests revealed that the moisture content and the strength grade of HPC are the two main factors governing spalling, and the results confirmed the existence of a vapor pressure mechanism for explosive spalling. The measured strength loss of HPC exposed to fire was basically similar to that of normal strength concrete (NSC). In the range 400 to 800 C, HPC lost most of its original compressive strength, which almost had been maintained up to 400 C. On the other hand, tensile splitting strength and elastic modulus of HPC sharply reduced above 200 C. After exposures up to 600 C, the HPC R-curves exhibited a flattening tendency. The flattening effect means that HPC becomes more ductile but undesirably weaker after exposure to elevated temperatures. Consistent with the cracking observations, the R-curve reveals that a stable crack growth process proceeds prior to explosive spalling. The vapor pressure mechanism for explosive spalling was confirmed from the cracking observations. From the QXRD tests, the decomposition of C-S-H gel was found to initiate at 600 C, whereas pore structure coarsening was found from the MIP measurements to occur above 200 C. Based on all the experimental data obtained, a mechanism explaining the fire damage to HPC has been established. The fire damage to HPC material has two components. One is the mechanical property loss caused by the thermally induced changes to the microstructure of HPC, and the other is the stable crack growth inside HPC driven mainly by internal vapor pressures.