Erik J. Sellevold

Ice formation in hardened cement paste, Part I — room temperature cured pastes with variable moisture contents

Abstract Ice formation in room temperature cured mature hardened Portland cement paste with different moisture contents have been measured. The measurements were carried out continuoslly in the temperature range from +20° C to −60° C. For moisture contents higher than that corresponding to approximately three equivalent BET-monolayers, the initial freezing temperature increases with increasing moisture contents. Moisture contents below approximately three monolayers is not freezable. For specimens containing freezable water, there is an increase in the amount of non-frozen water content with increasing moisture contents. Further, there is a slight increase in the amount of non-frozen water with increasing water/cement ratios for specimens in equilibrium with the same relative water vapor pressure. It is postulated that frost problems only can be expected in specimens containing more water than corresponding to a relative water vapor pressure of about 0.9.

On the measurement of free deformation of early age cement paste and concrete

Abstract Autogenous deformation and thermal dilation produce stresses, which may lead to cracking in early age concrete subjected to external restraint. To quantify the two types of “free deformation” by measurements is a prerequisite to fundamental understanding, as well as to formulate numerical models for use in stress calculations. However, results reported in the literature reveal large deviations and also inconsistencies between different measuring methods. The present paper discusses free deformation measurements and show that different types of measuring errors are involved, where, for instance, reabsorption of bleed water is an important one, and a standard test procedure should therefore describe how to handle the effect of bleeding. Furthermore, it is possible to obtain fairly good reproducibility within one laboratory using the same test rig, whereas a Round-Robin test program showed that it is far more difficult to produce similar results from different laboratories measuring on the same concrete. The implication is that measuring errors were present. Hence, there is a need for more calibration work and better control of test rig behavior.

Ice formation in hardened cement paste, Part II — drying and resaturation on room temperature cured pastes

Abstract Ice formation in partly dried and resaturated room temperature cured mature Portland cement paste has been measured. The measurements were carried out continuously in the temperature range from +20° C to −60° C. The results demonstrate that the drying-resaturation treatment increases the volume of large pores and the continuity of the pore system. The increase in the volume of large pores occurs at the expense of small pores and necks in the pore system. As the relative water vapor pressure at which drying takes place is decreased to 0.58, the number of continuous large pores is increased. Drying at lower relative water vapor pressures does not further increase the effect. The present findings point out the importance of the moisture history of concrete before testing such properties as water permeability and frost resistance.

Ice formation in hardened cement paste, part III — Slow resaturation of room temperature cured pastes

Ice formation in partly dried and very slowly resaturated room temperature cured mature Portland cement pastes has been measured in the temperature range from +10° C to −60° C. It is demonstrated that there is an aging effect on the ice formation pattern. During resaturation the amount of non-frozen water increases with time and thereby the amount of micropores. Furthermore, there are indications that also the amount of non-freezable water increases. The water content at resaturation decreases irrecoverably with increasing severity of the previous desorption, but it is almost constant when desorption has been carried out to equilibrium at relative water vapor pressures below 0.58.

Mechanical properties of young concrete: Part I: Experimental results related to test methods and temperature effects

Results of several test series on mechanical properties of young concrete are presented. Six different concrete mixes were tested systematically, and a number of other concrete mixes less extensively, all with w/b≈0.40. The program included compressive strength, tensile strength, splitting tensile strength and E-modulus determination both in compression and tension. Because temperature influences the rate of property development, and also the “final value” of a given property, the specimens were subjected to realistic temperature regimes the first few days as well as isothermal conditions. The test methods are described, and results obtained by the different test methods are compared. It is recommended that the temperature sensitivity constants should be determined from compressive strength tests on specimens exposed to realistic temperature histories. These parameters depend strongly on the cement type and the silica fume content. It was found that high performance concretes were quite robust to the negative effects of elevated temperatures. This was particularly true for concretes with pozzolana. In part two of the paper model parameters for an equation to be used in calculation programs are determined, and a test program for crack risk estimation is proposed.RésuméDans ce rapport sont présentés les résultats de plusieurs séries d’essais sur les propriétés mécaniques du béton jeune. Six mélanges différents ont été testés de façon systématique, et quelques autres de manière moins approfondie. Le but de ces expériences était de déterminer la résistance en compression et en traction, la résistance à la rupture en traction ainsi que le module d’élasticité, à la fois en compression et en traction. Comme la température influence la façon dont évoluent les caractéristiques du béton, ainsi que la «valeur finale» d’une caractéristique donnée, les échantillons ont été soumis durant les premiers jours d’une part à des régimes de températures semblables aux conditions naturelles, et d’autre part à des conditions isothermiques.Sont ici données une description des méthodes expérimentales utilisées et une comparaison des différents résultats obtenus. Il est recommandé de déterminer les constantes de sensibilité à la température à partir des essais de résistance en compression sur les échantillons soumis aux variations de températures. Ces paramètres dépendent beaucoup du type de ciment et de la quantité de fumée de silice contenus dans le mélange. On a remarqué que les bétons à hautes performances résistaient plutôt bien aux conséquences négatives des températures élevées. Ceci est vrai en particulier pour les bétons contenant du ciment pouzzolanique.Dans la seconde partie de ce rapport, les paramètres de modélisation pour l’équation MC1990 modifiée sont déterminés, pour les six mélanges, dans le cas de la résistance en compression et en traction, et du module d’élasticité. En outre, il est proposé un programme expérimental destiné à l’évaluation des risques de fissuration.

Mechanical properties of young concrete: Part II: Determination of model parameters and test program proposals

The paper concerns testing and modelling of the mechanical properties required as input to calculation programs made for crack risk estimation of hardening concrete structures. The results from several test series on mechanical properties of young concrete as described in Part I of this paper, are further evaluated. Model parameters for the modified CEB 1990 Model Code-equations, are determined for six concrete mixes, all having a w/b-ratio on 0.40, for compressive strength, tensile strength and E-modulus. To make the models applicable for young concrete, at0-parameter is introduced to fix the time at which significant mechanical properties are present. A test program to determine the model parameters is proposed, based on the experience that compressive strength tests have the smallest statistical scatter and that they are simplest to carry out.RésuméCe rapport fait état des essais et de la modélisation concemant les propriétés mécaniques du béton, données nécessaires aux programmes de calcul pour l’estimation des risques de fissuration des structures fraîchement coulées. Les résultats de plusieurs séries d’essais sur les propriétés mécaniques du béton au jeune âge, décrits dans la première partie de ce rapport, font l’object d’une analyse pluspoussée. Les paramétres de modélisation pour l’équation MC1990 modifiée sont déterminés, pour les six mélanges, dans le cas de la résistance en compression et en traction, et du module d’élasticité. Pour que ces modèles puissent s’appliquer au béton jeune, on introduit un paramètret0 représentant l’instant à partir duquel on considère les propriétés mécaniques comme significatives. Il est proposé un programme expérimental visant à déterminer les paramètres de modélisation, basé sur le fait que les essais de résistance en compression montrent une très bonne reproductibilité et sont les plus simples à mettre en œuvre.

Effects of Silica Fume and Temperature on Autogenous Deformation of High Performance Concrete

This paper uses four high-performance concretes with the same water-to-binder (w/b) ratio to investigate the effects of different silica fume content and temperature on autogenous deformation. The four concretes with a w/b ratio of 0.40 and from 0 to 15% silica fume were tested under 20 deg C isothermal conditions and under realistic (semi-adiabatic) temperature developments with maximum temperature in the range 60 to 65 deg C. The coefficient of thermal expansion is not very sensitive to silica fume content and its time/temperature dependence may be expressed by the maturity concept. The autogenous shrinkage is extremely temperature dependent, and, importantly, isothermal data cannot be used to predict the behavior during realistic temperature histories. Findings show that an increase in silica fume (1:1 replacement of cement) generally increases early autogenous shrinkage; however, the increase depends strongly on the temperature history and occurs primarily in the first 2 days. The consequence of the autogenous deformation data on cracking risk depends on a large number of other factors, and must be calculated for each particular structure. However, the data from this study is important input data for such a calculation as long as other factors are considered.