Restrained autogenous shrinkage can lead to considerable tensile stresses in concrete with low water/cement-ratio, especially in ultra-high strength concrete that was focused upon herein. To avoid cracking, it is essential to comprehensively investigate free and restrained autogenous shrinkage. However, the nature of these phenomena, in particular their very early onset, brings about some formidable challenges in terms of methodology.
Hydration is a complex process, in Portland cement paste and even more so in concretes with numerous constituents influencing hydration. It is very difficult and laborious to completely capture the very early development of hydration and of micro-structural properties. Therefore micro-analytical approaches, aiming at the complete understanding and modeling of underlying mechanisms, will require much time and large resources.
This study was to contribute to a more pragmatic and immediate solution. The strategy behind the chosen phenomenological approach was, on the one hand, to assess the autogenous shrinkage cracking propensity for the comparably simple case of constant temperature, and, on the other hand, to select, improve and develop methods suitable for investigating those parameters whose influence is most difficult to capture up to now, namely temperature and viscoelasticity. The basic objective of this strategy is to provide the methods for measuring all relevant quantities as precisely and efficiently as possible under all relevant conditions. The long-term goal is the empirical modeling of the autogenous shrinkage cracking propensity from as few and as simple input parameters as feasible and reasonable.
As indicated, the present knowledge about the influence of temperature on autogenous shrinkage and related cracking is insufficient. The corrugated tube method for measuring the autogenous shrinkage, recently standardized in the U.S., is precise and reliable, as was confirmed by the own tests. However, it is limited to isothermal conditions. Besides, with sticky or stiff mortars the casting is hampered by the small inlet diameter of the tubes. Therefore the ‘shrinkage cone method for measuring the autogenous shrinkage’ was developed. The investigation of repeatability and reproducibility proved that the method is precise, reliable and easy to use. The very good agreement with results obtained with the corrugated tube method evidenced the accuracy. Since the method is also suitable for tests under non-isothermal conditions, it will facilitate the required clarification of the role of temperature.
Tensile stresses due to autogenous shrinkage were investigated by the restrained ring test. The method proved to be very efficient and sufficiently precise. Most tests were carried out at a constant temperature of 20 °C and evaluated by means of an existing stress solution. In addition, an analytical solution for the thermal stress component due to different thermal expansion of steel and concrete ring at changing temperatures was derived. This will facilitate the future evaluation of restrained ring tests under non-isothermal conditions. Furthermore, the potential stress redistribution in restrained ring tests at early and very early age was pointed out.
The development of the stress-strength ratio was used to determine the autogenous shrinkage cracking propensity of four different ultra-high strength concretes under quasi-isothermal conditions. The applicability of the stress-strength criterion was analyzed and confirmed. Results of splitting tension tests at very early age needed to be multiplied by a factor of 2 to 3 to correspond to the maximum stress at failure observed when rings dried out at very low relative humidity from respective ages. These values as well as the factor’s age-dependent disproportional decrease were in good agreement with a recently proposed function for the development of the ratio of uniaxial to splitting tensile strength.
The influence of tensile creep and relaxation on restraint stresses is known to disproportionally increase with increasing stress level and decreasing age at loading. However, there is hardly any quantitative knowledge as to this influence at very early age. An ample experimental investigation of tensile creep at very early age with very high stress levels and of its influence on the cracking propensity would be extremely demanding and has not been attempted, yet. Common stress rigs like the particularly large temperature-stress testing machines are hardly suitable for comprehensive parameter studies; actually, most investigations were found to have included too few tests to allow for any statistically backed conclusion. This appears particularly critical in view of the two round-robin tests with temperature-stress testing machines on free and restrained autogenous shrinkage which failed to prove a sufficient reproducibility of results.
Therefore it was analyzed whether the restrained ring test could be an alternative also for this task. In its elementary form it is a passive test method, i.e. the strains cannot be controlled and the stress history is always conditional on the autogenous shrinkage and the development of the mechanical properties. The influence of time-dependent deformations can be investigated by comparing theoretical-elastic and actual stress, however, not always in a satisfactory manner. A more versatile investigation is feasible with an additional stress component which can be created at arbitrary points in time and with different rates of stress increase by either drying of the concrete to the ambient air or by a specific variation of temperature. If the additional free concrete strains are known, the creep factor and the ratio of creep factor to stress level may be suitable to describe the influence of time-dependent deformations.
The main conclusions regarding the methodology can be summarized as follows:
- The shrinkage cone method is suitable for testing the autogenous shrinkage of hardening cement paste and fine-grained concrete. This includes sticky and stiff materials. Tests under non-isothermal conditions are possible as well.
- The restrained ring test is suitable for determining the autogenous shrinkage cracking propensity of concrete. It is more appropriate for determining the cracking propensity of fine-grained ultra-high strength concrete than temperature-stress testing machines.
- The stress-strength failure criterion is applicable at very early age. Splitting tension tests can be used to determine the strength, however, results require a strong age-dependent correction.
- An additional stress component in restrained ring tests may be used to investigate the residual stress capacity and the influence of stress relaxation on cracking. The additional stress can be produced by temperature changes or by drying.