HARDENING PROCESS OF BINDER PASTE AND MICROSTRUCTURE DEVELOPMENT
7.6. DEVELOPMENT OF STRESSES DUE TO RESTRAINT
The test results presented in the foregoing sections showed free, non-restricted de- formations of cement pastes and concretes made with and without SAP. The mag- nitudes of these deformations are expected to give an indication on the possible in- tensities of tensile stresses which would be found in concrete structures if the shrinkage strains were restrained. However, the real values of tensile stresses de- pend also on the development of concrete stiffness, creep deformations and on thermal dilation due to the heat of hydration. In order to estimate the actual tensile stresses due to restraint, two types of apparatus are commonly used: uniaxial- restrained shrinkage testing machine [49], and an instrumented ring [50].
In parallel to their investigations on free autogenous shrinkage, Igarashi and Watanabe [30] studied the development of stresses due to restrained shrinkage us- ing a closed-loop, computer-controlled, uniaxial-restrained shrinkage (CLCCURS) test apparatus, cf. Figure 7.23 [49]. The mixture proportions and the test condi- tions are presented in the previous Sections 7.4.1 and 7.4.3. According to [30], the addition of SAP successively reduced stresses due to restraint. No tensile stress was generated in the concrete with 0.70% SAP, which was consistent with the complete elimination of free autogenous shrinkage in this mix. The tensile stresses measured for the reference mixtures were between 0.7 and 1.1 MPa at an age of 7 days. It should be mentioned that unexpectedly the stress in the reference concrete with the water-to-cement ratio of 0.3 was greater than the stress developed in the concrete with the w/c = 0.25. This finding requires further investigation. For the mix containing 0.35% SAP, the tensile stress at the age of 7 days reached 0.3 MPa.
Fig. 7.23. Uniaxial restrained-shrinkage (CLCCURS) testing machine [49].
Jensen and Hansen [28] evaluated the cracking susceptibility of mortars with and without SAP addition using an instrumented ring test. The setup typically consists of an inner steel ring that partially restrains the autogenous shrinkage of the annular test specimen cast around the ring. Due to the specimen shrinkage, this type of restraint gives rise to a uniform, radial pressure on the steel ring and in- duces tensile hoop stresses in the specimen. The stress build-up can be measured by strain gauges on the inside of the inner steel ring. Figure 7.24 shows the mortar test results. Significant tensile hoop stresses are induced in the mortar without SAP addition: After approximately 3 days of hardening, this mortar cracked as in- dicated by the sudden drop in stress. Internal curing based on SAP addition is seen to be effective in counteracting the stress build-up, which was counteracted partly in the specimen with 0.3% SAP addition and almost fully in the specimen with 0.6% SAP addition. The small stress, approximately 0.15 MPa, measured for the 0.6% SAP addition may have been induced by the slight temperature change dur- ing hydration. In any case, neither of the two mortars at 0.3% and 0.6% SAP addi- tion cracked within the 3-week period of testing.
Fig. 7.24. Equivalent hydrostatic pressure developed in the inner ring by mortars with different amounts of SAP and, consequently, different amounts of entrained curing water. SAP additions are given by weight of cement, where 0.6% corresponds to an entrained w/c of 0.075. The basic w/c was 0.3 for all mixtures [28].
Mechtcherine et al. [46, 47] also used the instrumented ring test to assess the magnitude of the tensile stresses developed due to restrained autogenous deforma- tions. They estimated quantitatively the cracking tendency of fine-grain UHPC with and without SAP addition. A concrete annulus was cast around a steel ring of dimensions standardized by ASTM C1581-04 [50], cf. Figure 7.25a. The strains were measured continuously by four strain gauges glued to the inner side of the steel ring. These strain values were used for calculating the tensile stresses in con- crete according to the equations proposed in [51]. Figure 7.25b shows the results
2.0 1.5 1.0 0.5 equivalent hydrostatic pressure [MPa] 0
time [d]
0 7 14 21
0% SAP
0.6% SAP 0.3% SAP
-2 0 2 4 6 8 10
0 20 40 60 80 100 time [d]
tensile stress [MPa]
obtained from ring tests for the reference concrete F-R and SAP-enriched mixtures F-S.4.07 (SAP 0.4% of cement mass, wIC/c=0.07) and F-S.6.08 (SAP 0.6% of ce- ment mass, wIC/c=0.08) over the first 100 days after mixing. The mitigation of autogenous shrinkage using internal curing caused a dramatic reduction of the stresses due to restraint. For example, at a concrete age of 50 days, the stresses in- duced in the specimens made with the mix F-S.4.07 containing SAP were ap- proximately 2 MPa and so 4 times smaller than the corresponding stresses in the reference concrete at the same age. For a higher SAP content and a higher content of additional water used in mixture F-S.6.08, the developed stresses were even lower. This clearly demonstrates the considerable reduction of the cracking poten- tial of UHPC as a result of internal curing.
(a) (b)
Fig. 7.25. a) Setup for the instrumented ring test; b) development of tensile stresses due to re- straint in sealed concrete specimens made of fine-grained UHPC without SAP addition (F-R) and with SAP (F-S.4.07: SAP 0.4% of cement mass, wIC/c=0.07; F-S.6.08: SAP 0.6% of cement mass, wIC/c=0.08) in the first 100 days after casting. The graph includes data from [46, 47] sup- plemented by additional measurements from [52].
As can be seen in Figure 7.25b, the development of stresses is very unsteady at early concrete ages, which likely results from temperature changes due to the heat of hydration; see also Figure 7.26a showing the tensile stress levels in the first 24 hours after casting. In order to estimate the possible effects of hydration heat on stress development, temperature was measured on the inner steel ring and in con- crete. Figure 7.26b shows the development of temperature in the experiments on
F-R
F-S.4.07
F-S.6.08
the fine-grain UHPC with and without internal curing produced and cast on the same day. Note that no calibration of the measuring sensors with regard to abso- lute temperature was carried out because it was not necessary, i.e., only the rela- tive temperature changes need be considered. The minimum temperatures re- corded with each sensor (the lowest point of each curve) roughly corresponded to the room temperature in the lab (approximately 20°C).
Both concretes, which warmed up during the intensive mixing process, showed a gradual decrease in temperature during the first 7-8 hours (mixture without SAP) and approximately in the first 10-11 hours (mixture with SAP) after casting, which was due to the cooling of the mixes down to the temperature of the steel rings and the ambient atmosphere in the lab. Subsequently, the temperature rose again due to the heat of hydration and reached its maximum at a concrete age of approxi- mately 18-19 hours for the reference mixture and at about 22-23 hours for the concrete containing SAP and extra water. It appears as well that the rise of tem- perature is less pronounced when internal curing is used.
-2 0 2 4 6 8 10
0 4 8 12 16 20 24 time [h]
tensile stress [MPa]
19 20 21 22 23 24 25 26
0 4 8 12 16 20 24 time [h]
temperature developed [°C]
Fig. 7.26. Results of instrumented ring tests on sealed concrete specimens made of fine-grained UHPC without SAP addition (F-R) and with SAP (F-S.4.07: SAP 0.4 mass-% of cement, wIC/c=0.07; F-S.6.08: SAP 0.6 mass-% of cement wIC/c=0.08) in the first 24 hours after casting:
a) development of tensile stresses due to restraint in numerous tests performed [46, 47]; b) typi- cal temperature changes measured in concrete and at the inner surface of the inner steel ring due to boundary conditions and hydration heat [52].
The extent of the thermal stresses which can be developed in a concrete ring test depends mainly on the difference in the coefficients of thermal expansion (CTE) of steel and concrete. Given that the CTE of the steel is higher than that of the concrete, any increase in temperature in the system would induce tensile stresses in concrete, and cooling would cause the opposite effect. In the experi-
F-R ring F-R conc.
F-S.6.08 ring F-S.6.08 conc.
F-R
F-S.4.07
F-S.6.08
(a) (b)
ments cited above, the ring steel had a CTE of 12x10-6 1/K according to the steel producer [52]. The CTE of UHPC with and without internal curing depends – as for any other concrete – on the concrete age and the type of aggregates. The au- thors of [46, 47] did not measure the CTE of the UHPC, but this was done in an- other study conducted on very nearly the same UHPC compositions, one of which also contained SAP, indeed without extra water [53]. The measured CTE values for the very young UHPC were between 11.0 and 11.5*10-6 1/K, and hence slightly below that of the steel used in fabrication of the rings. Even if the differ- ence in the CTE was not pronounced, it seems to have been large enough to in- duce thermal stresses due to the hydration heat observed in the experiments.
In order to minimise the effects of temperature, Eppers et al. [54] modified their test setup by reducing the rectangular cross-section of the concrete ring to 24 mm x 25 mm, and the geometry of the steel ring was adapted accordingly. The maximum temperature rise did not exceed 1.5K in their test. The authors worked with the same UHPC reference composition (indicated with 1A in Figure 7.27) as in [46, 47] with only very slight changes in mix proportions. Furthermore, another type of SAP at a dosage of 0.3% by mass of cement was used in the SAP-enriched mixture (1A SAP), and no extra water was added. Additionally, a UHPC mixture with a shrinkage-reducing admixture (mixture 1A SRA) was investigated. Figure 7.27 shows the development of tensile stresses in the concrete specimens and their estimated cracking propensity, which was defined as ratio of the tensile stress and the measured or converted splitting tensile strength. According to these results, the addition of SAP leads to a considerable reduction of stresses developed due to re- strained autogenous shrinkage, and the cracking propensity is diminished as well.
It is worth noting that the use of an SRA agent also provided good results with re- gard to the mitigation of autogenous shrinkage and its consequences.
0.0 0.5 1.0 1.5 2.0 2.5
6 8 10 12 14 16 18 20 22 24 Age [h]
Restraint stress [MPa] . 1A
1A SAP
1A SRA
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
8 10 12 14 16 18 20 22 24
Age [h]
Ratio of stress to resistance R [-]
1A SAP 1A SRA
R: stress at failure in modified ring tests (crack/micro-crack) R: splitting tensile strength
R: converted splitting tensile strength, stress redistribution 1A
Fig. 7.27. Results of instrumented ring tests on sealed concrete specimens made of fine-grained UHPC without additives (1A), with SAP (1A SAP: SAP 0.3 mass-% of cement, wIC/c=0) and with SRA (1A SRA: SRA 4.5% by mass of water) during the first 24 hours after casting: a) de- velopment of tensile stresses due to restrained autogenous shrinkage, b) theoretical (1A, 1A SAP, 1A SRA – full lines) and corrected (1A – dashed line) cracking propensities of tested concretes [54]
In their study of high-strength mortars, Schlitter et al. [55] dealt with both the effects of internal curing by using SAP and of temperature change on the stress development of mortars. In contrast to the studies presented so far in this chapter, a particular test setup was used, which included a set of dual rings made of invar, characterized by a very low CTE. Because of this choice of material, thermal de- formation of the rings should be excluded to a great extent, which should facilitate the interpretation of the measurement data.
In the examinations internal curing was applied to two of the three mixtures with an initial w/c of 0.3. The effectiveness of internal curing was verified by the evolution of the restrained shrinkage deformations recorded by means of both the outer and inner invar rings. Supplementary examinations included free shrinkage tests as well as tests performed to determine splitting tensile strength and modulus of elasticity. Furthermore, mortars were tested under a temperature of 23°C±0.2°C, which was held constant throughout the duration of the test or re- duced at a rate of 1°C/h at the ages of 1, 1.5, 2, or 3 days.
In good agreement with the results of previous studies, the authors found a great reduction of stresses due to the application of internal curing. Reserve crack- ing capacity, defined by the authors as the magnitude of stress required by the temperature reduction to crack the sample, was improved as well. This informa- tion was consistent with the results of the free shrinkage tests. Interestingly enough, the higher resistance to cracking in the SAP-enriched mixes appeared to be in contradiction to the experienced loss of tensile strength. As a reason for this behaviour, the authors pointed out the lower modulus of elasticity which was measured for both SAP-enriched mixtures, and an expected, however not tested, increase in creep. Eventually, it was found out that higher amounts of IC water carried by SAP might not always contribute to improving cracking resistance. This conclusion was drawn on basis of leveled reserve cracking capacity in both SAP- enriched concretes.