J. Cent. South Univ. Technol. (2008) 15(s1): 577-581

DOI: 10.1007/s11771-008-425-9

Creep experimental test and analysis of high-performance concrete in bridge

CHEN Zhi-hua (陈志华)¹, YUAN Jian (袁 健)²

(1. School of Civil Engineering and Architecture, Wuhan University of Technology, Wuhan 430070, China;

2. School of Civil Engineering and Mechanics, Central South University of Forestry and Technology,

Changsha 410004, China)

Abstract: Factors that have effect on concrete creep include mixture composition, curing conditions, ambient exposure conditions, and element geometry. Considering concrete mixtures influence and in order to improve the prediction of prestress loss in important structures, an experimental test under laboratory conditions was carried out to investigate compression creep of two high performance concrete mixtures used for prestressed members in one bridge. Based on the experimental results, a power exponent function of creep degree for structural numerical analysis was used to model the creep degree of two HPCs, and two series of parameters of this function for two HPCs were calculated with evolution program optimum method. The experimental data was compared with CEB-FIP 90 and ACI 209(92) models, and the two code models both overestimated creep degrees of the two HPCs. So it is recommended that the power exponent function should be used in this bridge structure analysis.

Key words: cement; concrete; creep; high-performance concrete (HPC)

1 Introduction

About one hundred years have elapsed since the first observations of concrete shrinkage in the previous century and the discovery of concrete creep in 1907 by HATT (RILEM Recommendation, 1998). Much research on experiment and analysis theory has been devoted to this complex problem ever since. However, concrete is a kind of complex material, its mixture composition, curing conditions, ambient exposure conditions and element geometry certainly influence its creep and shrinkage. Specially, in the last two or three decades the development and utilization of high performance concrete (HPC) in tall buildings, offshore structures, bridges and other prestressed structures have been in focus. The HPC provides superior mechanical properties and durability, high erection speed and good workability at site, but the increased uses of such concretes are accompanied by concern regarding their creep and shrinkage. In order to get a high benefit of using these concretes there is a need for a better understanding and treatment of HPC creep. One large test series have been studied by PERSSON^[1-3]. Between 1991 and 1999, about 400 cylinders made of eight mix compositions of HPC were studied in the laboratory and in the field for creep of young or mature HPC. M?LLER et al^[4] and PERSSON^[5] had also studied the differences between laboratory test and field results of creep and shrinkage. The time-dependent behavior of early age has been studied in Refs.[6-7]. The researchers dealt with numerical modeling to characterize early age tensile, compressive creep and its associated stress relaxation.

For creep experiment takes a long time, creep in concrete structures is estimated with one of the ACI 209(92) model^[8] and CEB-FIP 90 model^[9]. In China, there is few data about concrete creep, the designers typically use CEB-FIP 90 code model to estimate concrete creep. But for important structure such as nuclear reactor shell, large bridge etc, it is recommended that concrete creep model is from experiment data of concrete mixtures used in the project. To analyze concrete creep of a bridge member accurately, laboratory tests on creep of two HPCs used in the bridge have been carried out. Creep analysis is usually finished through viscoelastic model, a power exponent function is introduced to model the creep of two HPC mixtures used in bridge prestressed members, and the corresponding parameters of the function are given by evolution program optimum method for two HPC mixtures. In the last, the experiment data were compared with CEB-FIP 90 and ACI 209(92) models.

2 Material

Based on the demand of design and construction, such as strength, good workability, low heat of hydration, mix composition of HPCs was optimized. Table 1 shows the mix composition of the HPCs and main properties. The slurry content of the gravel is 1.08% and its crush

Table 1 Mix composition of fresh HPC and its properties

index is 6.20%. The slurry content of the nature sand is 1.25% and the fineness modulus is 2.7. There are two types of plain Portland cement produced by two factories, numbered as C1 and C2.

3 Experimental

Creep of two HPCs described previously was studied in constant temperature environment (20 ℃) under laboratory conditions. The creep test was studied in a common spring-loading device. Shrinkage and creep were studied for cylinder 450 mm long with a diameter of 150 mm. Moduli of elasticity also are studied on the same specimens. DI-25 LVDT gauging and displacement transducers carried out the measurement. Other properties of the HPCs such as strength were studied on cubes with a side of 150 mm. After being cast, these testing specimens were cured in fog house. The moulds were removed after 48 h and the shrinkage and creep specimens were sealed with white sheet iron cylinder to control the moisture losses, then the cylinder specimens were cured in the curing house for creep at (20±2) ℃ and the cubes were in the standard curing house until loading age. The loading ages of the HPCs were 3, 7, 28, 90 and 180 d after specimens being cast for each HPC mixture. For each loading age, the stress σ_c loaded on specimens must be calculated with Eqn. (1).

                                  (1)

where R_f is ultimate compressive strength of the creep specimens, R is the ultimate compressive strength of the cubic specimens. Before testing the creep of the HPCs, the modulus of elasticity of the creep specimens for each loading age was studied on the spring-loading device (the ultimate stress for modulus of elasticity was σ_c). When stress reached the ultimate stress that was the creep loading stress, the testing data from the LVDT gauging and displacement transducers were looked as datum quantity of the creep specimens. The data of autogenously shrinkage should be recorded when the data of creep specimens was done. The creep tests were lasted at least half a year and up to one year. Fig.1 shows the device and the specimens. Shrinkage was studied on cylinders made from the same batch as those used in the creep tests.

Fig.1 Creep test device and specimens

4 Results

Fig.2 shows the creep degrees of concretes W1 and W2, respectively. Creep degree is concrete creep strain per stress. From these creep curves of each age, when loading age is later, creeps of HPC are less. In earlier age, the cement was hydrating and the strength was lower, so the creep of HPC was larger. However, with loading age increasing, the strength of HPC was higher, so the creep was smaller. If the creep degree of 28-day loading age, denoted as C(t,28), is looked as reference. C(t,3)/C(t,28) was 1.5-5, C(t,7)/C(t, 28) was 1.3-2.7, C(t,90)/C(t,28) was about 0.75, C(t,180)/C(t,28) was about 0.6. Creeps of two HPCs were less than those of the ordinary concrete for same loading age. The creep degrees of 360-day of two HPCs that started testing at 3 and 7 day-

Fig.2 Creep degree of concretes W1 and W2

age were about 45×10^-6 and 30×10^-6/MPa, respectively. Comparing with the creeps of two HPCs, the strengths of the two HPCs were closed and the environment conditions such as curing condition were the same, but the materials of two concretes were different. The creep of concrete W1 was larger than that of concrete W2. This difference is resulted from the ingredient of the two kinds of concrete such as proportion of granulated slag, addition agent type and proportion, strength of aggregate and fineness of cement.

5 Analyses

5.1 Creep degree

In the previous section, data about creep degree of HPC was calculated according to testing data of creep experiments. In order to use creep degree in structural analysis and design, the relationships of creep degree with the loading age and load duration should be formulated. Based on viscoelastic models, researchers (such as BURGERS, HANSEN, FLIIGGE, COWAN, ROLL, POWERS, NERILLE, et al)^[10] had brought up creep degree functions: power function, logarithm function, hyperbola function, exponential function, power exponent function, polynomial exponent function, etc. The first three type functions were used in earlier creep research of concrete, which are too simple to express much comprehensive creep characteristics of concrete. The first four type functions expressed the relationship between creep degree and load duration but loading age. It resulted in that there are different expression coefficients for the creep at different loading ages, so these functions are not convenient in use. The last two types of functions included the effects of loading age and load duration, and the creeps of different load durations for different loading ages of concrete could be calculated with one formulation. There are many factors which have effects on creep of concrete, so the creep expressions are commonly complicated and there are many parameters in the expressions. The creep function in ACI 209(92) code model includes the effects of loading age and load duration, the creep correct factors of initial moist curing, ambient relative humidity, volume-surface ratio of the concrete member, slump, fine aggregate percentage, cement content and air content. The creep functions in CEB-FIP 90 model code and NEN 6720 model code^[11] also took into account the effects of loading age, cement type, load duration, the temperature of concrete at loading age, relative humidity of ambient environment, and notational size of concrete member.

In this paper, the HPC mixture has been decided, so in the creep degree model, the effects of loading age and load duration on the creep are discussed for each HPC mixture. Because of the higher nonlinearity of power exponent function and polynomial exponent function, the least square approximation and non-liner regression methods for the parameters in the functions were not in use^[12]. The better results were from evolution program (EP) optimal methods. EPs are more general terms that include GAs as well as other derived heuristics, sharing the same parallel evolutionary principles. This method can search solution from many initial points simultaneity and obtain the global optimal solution of concrete creep model parameters with real numerical chromosome, variable probability of mutation, simple crossover and arithmetical crossover. Two series of the parameters of two HPCs were calculated with this method. Comparing with the results, the precision of power exponent function was higher than that of polynomial exponent function. The power exponent function was defined as^[13]

             (2)

                            (3)

                            (4)

where  τ is loading age (d), t-τ is the load duration (d), C(t, τ) is the tth day creep of concrete at loading age τ (10^-6/MPa), and φ₀, φ₁, p, γ₀, γ₁, q and s are the parameters which are decided by the testing data.

From the creep degree function, φ(τ) and γ(τ) show the effects of loading age. Using the evolution program, the parameters of power exponent function about the two HPCs were calculated (Table 2). The curves of the function and the testing data are shown in Fig.3 for each loading age.

Table 2 Factors of creep degree function

As shown in Fig.3, the results with the power exponent function and the parameters for the creep of two HPCs were close to the testing data at different loading ages. The errors for the later age of loading were higher, but they mostly were less than 10%. At the same time, loading is usually finished in the earlier age for concrete structure, so it is feasible to calculate creep degree using Eqn.(2) and the parameters are listed in Table 3.

5.2 Comparing ACI 209(92) model with CEB-FIP 90 model

CEB-FIP 90 model code is utilized for bridge structure analysis and design in China. Based on lab conditions, the creep predictions of CEB-FIP 90 model and ACI 209(92) model were calculated and compared with experimental curves of HPC W1 and W2 (Fig.4 and Fig.5). For HPC W1, the prediction curve of CEB-FIP 90 was closer with experimental data than that of ACI 209(92). But for earlier loading age, two code models underestimated short term creep and overestimated long term creep. For HPC W2, two code models both overestimated concrete creep. So the power exponent function should be used for structure analysis and design

Fig.3 Prediction of creep degree with Eqn.(2): (a) Creep degree of concrete W1; (b) Creep degree of concrete W2

Fig.4 Creep prediction and lab data of W1

Fig.5 Creep prediction and lab data of W2

of this bridge.

Some factors such as concrete strength, relative humidity, member size and cured condition, are considered in CEB-FIP 90 model and ACI 209(92) code model. HPC is used widely in structure in order to obtain special properties. More and more new addition agents that have efficiency on concrete creep are utilized in HPC, but CEB-FIP 90 model and ACI 209(92) creep model do not include these efficiencies. For important structures, these models should be modified by such methods as B3 model^[7].

6 Conclusions

1) Factors affecting creep of concrete include exterior conditions and interior components. A little change of some factors strongly influences the creep, for instance, ingredient and fineness, type of agent, etc. So the creep experiment of concrete made of construction material is necessary for important bridge structure.

2) The experimental results show that the creep of HPC is less than that of ordinary concrete. The creep degree of HPC W1 is greater than that of W2, and the difference results from the component of two concrete mixtures. The effect of each component on creep needs to be studied further. Based on the experiment results, the power exponent function of creep degree that can be applied in structure numerical analysis was given. This function can calculate the creep degree of different load duration at different loading ages. The results with this function are closer to those of the experiment.

3) Comparing the results of experiment with the CEB-FIP 90 model and ACI 209(92) model, at the later loading age, CEB-FIP 90 model and ACI 209(92) model are overestimated. So for the bridge member analysis and design, it is recommended to use the power exponent function model.

References

[1] PERSSON B. Basic deformation of high-performance concrete at early ages [J]. Nord Concr Res, 1997, 20: 59-74.

[2] PERSSON B. Seven-year study of the effect of silica fume in concrete [J]. Advanced Cement Based Materials, 1998, 7: 139-155.

[3] PERSSON B. Quasi-instantaneous and long-term deformations of hpc with some related properties [R]. Report TVBM-1016, Lund: Lund Institute of Technology, 1998.

[4] M?LLER H S, K?TTNER K. Characteristics and prediction of creep of HPC [C]// WITTMANN F H, SCHWESINGER P eds. Proceedings of the Fourth Weimar Workshop Held at HAB, Weimar, Freiburg and Unterengstingen, Germany and Switzerland, 1995: 145-162.

[5] PERSSON B. Correlating laboratory and field tests of creep on high-performance concrete [J]. Cement and Concrete Research, 2001, 31: 389-395.

[6] JC/T 853-1999. Clinker of Portland Cement 2000 [S]. China Building Materials Industry Association.(in Chinese)

[7] BAZANT Z P, BAWEJA S. Justification and refinements of model B3 for concrete creep and shrinkage: 1. Statistics and sensitivity [J]. Materials and Structures, 1995, 28: 415-430.

[8] ACI (1992). Prediction of creep, shrinkage, and temperature effects in concrete structures[S]. ACI 209R-92 (under revision by ACI Committee 209), 1992: 47.

[9] CEB (1993). CEB-FIP model code 1990, CEB bulletin D’Information[S]. 213/214, London: Thomas Telford, 1993: 437.

[10] NEVILLE A M, DILGER W H, BROOKS J J. Creep of plain and structure concrete [M]. London and New York: Construction Press, 1983.

[11] NEN 6720 (TGB 1990). Regulations for concrete, structural requirements and calculation methods [S]. Delft, 1995.

[12] CHEN Zhi-hua, SHAN Liang, GUAN Fu-ling. Identification of concrete creep parameters based on evolution programs [J]. Journal of Yangtze River Scientific Research Institute, 2005, 22(2): 47-49. (in Chinese)

[13] ZHU Bo-fang. Modulus of elasticity, creep degree and stress relax coefficients of concrete [J]. Journal of Hydraulic Engineering, 1985(9): 54-61. (in Chinese)

(Edited by HE Xue-feng)

Received date: 2008-06-25; Accepted date: 2008-08-05

Corresponding author: CHEN Zhi-hua, Associate professor; Tel: +86-27-63631700; E-mail: chenzh@whut.edu.cn