Predition of Temperature and Aging Effets on the Properties of Conrete Jin-Keun Kim 1* and Inyeop Chu 1 1 KAIST, Republi of Korea * E-mail: kimjinkeun@kaist.a.kr ABSTRACT For the sustainable onrete strutures, it is neessary to exatly predit the material properties of onrete with age and eah plae of strutures. This paper investigates predition models estimating the temperature and aging effets on the hydration properties of onrete, suh as the ompressive strength, the splitting tensile strength, the elasti modulus, and the autogenous shrinkage. A predition model is suggested on the basis of an equation that is formulated to predit the ompressive strength. Based on the assumption that the apparent ativation energy is a harateristi property of onrete, a predition model for the ompressive strength is applied to hydration-related properties. The hydration properties predited by the model are ompared with experimental results, and it is onluded that the predition model properly estimates the splitting tensile strength, elasti modulus, and autogenous shrinkage as well as the ompressive strength of onrete. Keywords. Temperature; Aging; Hydration properties; Apparent ativation energy; Predition model INTRODUCTION Conrete is an aging material, and the hydration-related properties of onrete hange with time. Meanwhile, the heat of hydration in mass onrete strutures reates a temperature gradient between the inner and surfae. This temperature gradient affets the properties of the onrete, i.e., the ompressive strength, elasti modulus, splitting tensile strength, and autogenous shrinkage. Also, the temperature variations aused by old-weather or hotweather onditions have the same influene on onrete. Hydration properties suh as the ompressive strength, elasti modulus, splitting tensile strength, and autogenous shrinkage are fators that must be onsidered in the design and onstrution of onrete strutures. Speifially, an evaluation of the thermal raking of a mass onrete struture requires an estimation of the elasti modulus and tensile strength of early-age onrete with the temperature. Therefore, it is very important to estimate the properties of onrete aording to different temperature and aging fators (Abdel-Jawad 2006, Alexander and Taplin 1962, Freiesleben Hansen and Pedersen 19, Gardner 1990, Kim et al. 1998, Kjellsen and Detwiler 1993, Lew and Reihard 198, Zhang el al. 2008). Many existing evaluations of the influene of temperature and aging on the hydration properties fous primarily on the ompressive strength, while only a few predition models
are available for estimating the elasti modulus and splitting tensile strength (Chengju 1989, Gardner 1990, Kim et al. 2001, Tank and Carino 1991). To evaluate the validity of a predition model, not only the ompressive strength but also other properties whih are related to the hydration proess should be estimated using the predition model. Among several existing models, the model proposed by Kim et al. (2001), whih is based on a new apparent ativation energy funtion, was reported to overome the shortomings of previous predition models. The objetives of this study are to evaluate the validity of the predition model whih is proposed by Kim et al. (2001) for estimating the hydration-related properties suh as the splitting tensile strength, elasti modulus, and autogenous shrinkage as well as the ompressive strength of onrete PREDICTION MODEL Although the Arrhenius law is generally aepted (Freiesleben Hansen and Pedersen 19, Han and Han 2010, Jonathan et al. 2011, Nielsen 200, Viviani et al. 200, Waller et al. 2004) as the most suitable rate funtion for hydration of onrete among several predition models, several researhes (Chanvillard and D Aloia 199, Jonasson 1985, Kjellsen and Detwiler 1993) about the shortomings of the Arrhenius equation have been reported. A predition model was proposed by Kim et al. (2001) to estimate the ompressive strength development with temperature and aging. This model mitigated the shortomings of previous models and reasonably approximated the experimental results pertaining to the ompressive strength. This paper investigates the effetiveness of this model as a tool for prediting the hydration properties like the splitting tensile strength, elasti modulus, and autogenous shrinkage as well as the ompressive strength of onrete. The following equation is the predition model proposed by Kim et al. (2001). S 1 S 1 n Eo t E i o t i1 RTi RTi 1 A e e ( ti ti 1) i1 u e e (1) where, S is the ompressive strength, S u is the limiting ompressive strength, A is a onstant, R is the gas onstant and equal to 8.3144 J/K mole, T i is the uring temperature at time step i (K), E o is the initial apparent ativation energy (J/mole), is a onstant, t i-1 is the initial age of time step i (days) and t i is the final age of time step i (days). This equation involves four unknown parameters, t o, E o,, and A. After the predition model was introdued, the appliation of the model has been investigated by several researhers (Kim et al. 2002, Chu et al. 2012). Kim et al. (2002) applied the model to predit the elasti modulus and splitting tensile strength as well as the ompressive strength of onrete. Reasonably good agreement was shown between the predited results and the measured results. However, the apparent ativation energy funtion was determined using the results of the ompressive strength only. While Chu et al. (2012) reported that the predition model ould also be applied to predit the autogenous shrinkage of onrete, there has been no attempt to apply the predition model to as determination of all hydration-related properties (ompressive strength, splitting tensile strength, elasti modulus, and autogenous shrinkage) using the same apparent ativation energy. The apparent ativation energy an be interpreted from the two standpoints of a miro level and a maro level. At the maro-level, the apparent ativation energy is related to the inreasing development rate of the ompressive strength. Thus, if the model is applied to
other properties for whih the inreasing rate is different from that of the ompressive strength, the regression urves will have different levels of apparent ativation energy aording to different properties. On the other hand, the apparent ativation energy at the miro-level is a funtion of the degree of ement hydration. Thus, if the apparent ativation energy is onsidered as a harateristi property of onrete at the miro-level, the apparent ativation energy is a onstant value in all types of hydration-related properties. Many researhers (Chanvillard and D Aloia 199, Riding et al. 2011, Kada-Benameur et al. 2000, Kjellsen and Detwiler 1993, Wang et al. 200) have proposed that the apparent ativation energy is related to the hydration proess of ement. Therefore, it an be assumed that the apparent ativation energy is a harateristi property of onrete and that it has a onstant value in all types of hydration-related properties. The inreasing rates of the splitting tensile strength, elasti modulus, and autogenous shrinkage are dissimilar to that of the ompressive strength, and the dissimilarity must be onsidered in Eq. (1). The proportional onstant A in Eq. (1) an simulate the differenes in the inreasing rate, even when the apparent ativation energy is idential to different hydration-related properties. Therefore, it an be assumed that the differene in the inreasing rate is able to be estimated by modifying the onstant A with hydration-related properties. In this paper, based on the previous assumption, experimental results of hydration-related properties like the ompressive strength, elasti modulus, splitting tensile strength, and autogenous shrinkage are analyzed using the predition model, Eq. (1). PREDICTION OF HYDRATION-RELATED PROPERTIES OF CONCRET Experiments. For the appliation of the predition model, the experimental results reported in previous researh (Chu et al. 2012) are analysed. Table 1 shows the mix proportion used in the experiments. Ordinary Portland ement (ASTM type I) and silia fume are used as ementitious materials. In this experiment, the autogenous shrinkage under an isothermal uring ondition (20, 30, and 40 C) was measured. Additional 100 200mm ylindrial speimens were ast along with the shrinkage speimens of 100 100 400mm to investigate the development of the mehanial properties of onrete like the ompressive strength, elasti modulus, and splitting tensile strength of the onrete under different isothermal uring onditions (20, 30, and 40 C). Table 1 Mix proportion Mix ID Curing temperature, ( C) w/m S/a Unit weight (kg/m 3 ) W C SF a S G b Ad C (%) C35/SF 20,30,40 0.35 0.40 15 465 35 665 100 1.0 a Silia fume, % of total weight of ementitious materials b Maximum aggregate size of 20mm Superplastiizer (ASTM Type-F high range water-reduing admixture), % of total ementitious materials Fig. 1 shows the experimental results of eah property of onrete speimens. As shown in Fig. 1, the one- and three-day ompressive strength levels inreased with an inrease in the uring temperature. However, this tendeny was reversed with aging. The 28-day ompressive strength of onrete ured at 40 C is the lowest. These results suggest that
onrete subjeted to a high temperature at an early age attains a higher early-age ompressive strength but a lower later-age ompressive strength. Alexander and Taplin (1962) referred to this phenomenon as the rossover effet. The same phenomenon also arises in the splitting tensile strength, elasti modulus, and autogenous shrinkage with inreasing uring temperatures. However, the rossover effet of the splitting tensile strength and the elasti modulus is not as obvious as that of the ompressive strength and autogenous shrinkage. This is due to the differenes between the rates of inrease for eah property. (a) Compressive strength (b) Splitting tensile strength () Elasti modulus (d) Autogenous shrinkage Fig. 1 Experimental results of hydration-related properties Appliation of predition model. As mentioned previously, the inreasing rates of the splitting tensile strength, elasti modulus, and autogenous shrinkage are dissimilar to that of the ompressive strength. This dissimilarity should be onsidered in Eq. (1). The differene in the inreasing rate an be estimated by modifying the onstant A with determined properties. Following previous researh (Kim et al. 2001), an A value in Eq. (1) of 1.0 10 is determined as the referene for the ompressive strength. From the relationship between the relative ompressive strength and eah property, the A value of eah property an be determined. The values of the ompressive strength, splitting tensile strength, elasti modulus, and autogenous shrinkage are 1.0 10, 3.0 10, 5.0 10, and 0.3 10, respetively. Table 2 tabulates the regression results of test results for the determination of parameters in the predition model, Eq (1). Based on the results, the variation of eah parameter with the uring temperature is presented and following general equations is obtained.
Curing temperature ( C) Eo Table 2 Regression results of parameters in the predition model E o (J/mole) 20 41083 0.0048 0.22 30 40861 0.002 0.11 40 3991 0.0102 0.00 42,369 58T J/mole (2) 0.00025 T (3) to 0.440 0.011T (4) Fig. 2 depits the predited urves as well as the experimental results of the hydration-related properties, i.e., ompressive strength, splitting tensile strength, elasti modulus, and autogenous shrinkage. In this figure, the sattered points denote the experimental results and the solid lines denote the predited values. The differenes between the alulated value and the experimental result likely arise beause the regression results are based on a small number of test results exluding the results at a later-age. However, as shown in Fig. 3, the error range is less than ±10 perent, whih reflets that the predition model properly estimates the hydration-related properties of onrete. t o (a) Compressive strength (b) Splitting tensile strength () Elasti modulus (d) Autogenous shrinkage Fig. 2 Predited values and experimental results
Fig. 3 Comparison of predited values and experimental results RELATIONSHIPS BETWEEN SPLITTING TENSILE STRENGTH AND COMPRESSIVE STRENGTH OR ELASTIC MODULUS AND COMPRESSIVE STRENGTH In many existing odes, i.e., ACI 318 (Amerian Conrete Institute 1999) and CEB-FIP model ode 1990 (CEB-FIP 1993), a relationship between the splitting tensile strength and the ompressive strength or the elasti modulus and the ompressive strength are presented by following equations. sp b f a f (5) b E a f (6) where, f sp is the splitting tensile strength, E is the elasti modulus, f is the ompressive strength, a and b are empirial onstants. In Eq. (5) and Eq. (6), a dissimilarity of inreasing rate between the ompressive strength and eah property is onsidered by the value of b. ACI 318 suggests b values of 0.50 and 0.50 for the splitting tensile strength in Eq. (5) and the elasti modulus in Eq. (6), respetively. While, CEB-FIP model ode 1990 suggests b values of 0.6 and 0.33 for the splitting tensile strength in Eq. (5) and the elasti modulus in Eq. (6), respetively. The value of b is losely related to the value of A in Eq. (1). In the previous setion, different values of A for eah property were determined based on a regression analysis of the experimental results. The values of the ompressive strength, splitting tensile strength, and elasti modulus were 1.0 10, 3.0 10, and 5.0 10, respetively. To obtain values of b orresponding to determined values of A, the relationship between the relative ompressive strength and the relative values of eah property were drawn in Fig. 4 (a) and (b). The determined values of b are 0.63 and 0.46 for the splitting tensile strength and the elasti modulus, respetively. Also, the values of A orresponding to the existing model odes is determined. These values of A are tabulated in Table 3.
(a) Relationship between f sp and f (b) Relationship between E and f Fig. 12 Relationship between the ompressive strength and eah property Table 3 Values of A orresponding to existing model odes Model equation Values of A in Eq. (1) for for splitting ompressive tensile strength strength for elasti modulus ACI 318 f E sp 0.56 f 4,00 f 0.50 k 0.50 k 1.0 10 3. 10 3. 10 CEB-FIP 1990 E f sp 0.30 f 8,480 f 0.6 u 0.33 u 1.0 10 2.1 10 6.5 10 CONCLUSIONS To investigate the validity of the predition model, the model is applied to the hydrationrelated properties of the ompressive strength, splitting tensile strength, elasti modulus, and the autogenous shrinkage. Based on results in this paper, the following onlusions were drawn. 1. Based on the assumption that the apparent ativation energy is the harateristi property of onrete, a predition model for ompressive strength is applied to other hydrationrelated properties, i.e., the splitting tensile strength, elasti modulus, and the autogenous shrinkage. The predited values by the model are ompared with experimental results, and good agreement between the predited values and the experimental results was observed within an error range of ±10 perent. 2. The proportional onstant A is determined to predit different hydration-related properties of the ompressive strength, splitting tensile strength, elasti modulus, and the autogenous shrinkage. Determined values of A are 1.0 10, 3.0 10, 5.0 10, and 0.3 10 for ompressive strength, splitting tensile strength, elasti modulus, and autogenous shrinkage, respetively. 3. An empirial onstant b in existing model odes, whih defines a dissimilarity of inreasing rate between the ompressive strength and eah property, is losely related to
the value of A in the predition model. The values of A orresponding to existing model odes are determined. ACKNOWLEDGEMENTS This work (20111520100090) was partly supported by the Nulear Researh & Development of the Korea Institute of Energy Tehnology Evaluation and Planning (KETEP) grant funded by the Korea government Ministry of Knowledge Eonomy. The authors wish to express their gratitude for the finanial support that has made this study possible. REFERENCES Abdel-Jawad, Y.A. (2006), The maturity method: Modifiations to improve estimation of onrete strength at later ages, Constrution and Building Materials, 20(10), 893-900. ACI Committee 318 (1999), Building Code Requirements for Reinfored Conrete and Commentary(ACI 318-99/318R-99/318M-99), Amerian Conrete Institute. Alexander, K.M. and Taplin, J.H. (1962), Conrete strength, ement hydration and the maturity rule, Australian Journal of Applied Siene., 13, 2-284. CEB-FIP (1993), CEB-FIP MODEL CODE 1990, Thomas Telford, London, England, 24-36 Chanvillard, G. and D Aloia, A. (199), Conrete strength estimation at early ages: modifiation of the method of equivalent age, ACI Material Journal, 94(6), 520-530. Chengju, G. (1989), Maturity of onrete: method for prediting early-stage strength, ACI Material Journal, 86(4), 341-353. Chu, I. et al. (2012), Estimation of temperature effets on autogenous shrinkage of onrete by a new predition model, Constrution and Building Materials, 35, 11-182. Freisleben Hansen, P. and Pedersen, E.J. (199), Maturity omputer for ontrolled uring and hardening of onrete, Nordiska Betongfoerbundet, 21-25. Gardner, N.J. (1990), Effet of temperature on the early-age properties of type I, type III, and type I/fly ash onrete, ACI Material Journal, 8(1), 68-8. Han, M.C. and Han, C.G. (2010), Use of maturity methods to estimate the setting time of onrete ontaining super retarding agents, Cement and Conrete Composites, 32(2), 164-12. Jonasson J.E. (1985), Early Strength Growth in Conrete-Preliminary Test Results Conerning Hardening at Elevated Temperature, RILEM Symposium on Winter Conreting. Jonathan L.P. et al. (2011), Effet of Chemial Admixtures on Apparent Ativation Energy of Cementitious Systems, Journal of Materials in Civil Engineering, 23(12), 1654-1661. Kada-Benameur et al. (2000), Determination of apparent ativation energy of onrete by isothermal alorimetry, Cement and Conrete Researh, 30(2), 301-305. Kim, J.K. et al. (1998), Compressive strength development of onrete with different uring time and temperature, Cement and Conrete Researh, 28(12), 161-13. Kim, J.K. et al. (2001), Estimation of ompressive strength by a new apparent ativation energy funtion, Cement and Conrete Researh, 31(2), 21-255. Kim, J.K. et al. (2002), Effet of temperature and aging on the mehanial properties of onrete Part II. Predition model, Cement and Conrete Researh, 32(), 21-255. Kjellsen, K.O. and Detwiler, R.J. (1993), Later ages strength predition by a modified maturity method, ACI Material Journal, 90(3), 220-22. Lew, H.S. and Reihard. T.W. (198), Mehanial properties of onrete at early ages, ACI Journal, 5(10), 533-542. Neville, A.M. (1995), Properties of Conrete, 4th ed., Longman, England, 666-64.
Nielsen, C.V. (200), Modeling the Heat Development of Conrete Assoiated with Cement Hydration, ACI Speial Publiation, 241, 95-110. Riding, K.A et al. (2011), New model for estimating apparent ativation energy of ementitious systems, ACI Material Journal, 108(5), 550-55. Tank, R.C. and Carino, N.J. (1991), Rate onstant funtions for strength development of onrete, ACI Material Journal, 88(1), 4-83. Viviani, M et al. (200), Seperation of thermal and autogenous deformation at varying temperatures using optial fiber sensors, Cement and Conrete Composites, 29(6), 435-44. Waller, V. et al. (2004), Using the maturity method in onrete raking ontrol at early ages, Cement and Conrete Composites, 26(5), 589-599. Wang, J. et al. (200), Apparent ativation energy of onrete in early age determined by adiabati test, Journal of Wuhan University of Tehnology-Materials Siene Edition, 22(3), 53-541. Zhang, J. et al. (2008), New perspetives on maturity and approah for high performane onrete appliations, Cement and Conrete Researh, 38(12), 1438-1446.