STUDIES IN CREEP CHARACTERISTICS OF CONCRETE AND REINFORCED CONCRETE

Abstract

Traditionally, concrete for pre-stressing work has been made using the same ingredients as for simple reinforcing concrete except that the strength requirements have been in excess of 40 MPa generally. However, owing to additional requirements of high performance concrete for durability point of view, use of mineral admixture has become more of a necessity than otherwise. One of the commonly adopted mineral admixtures in prestressed concrete particularly for marine applications is ground granulated blast furnace slag (GGBFS). It, therefore, becomes pertinent to study the impact of GGBFS on the creep characteristics of normal concrete to start with, along with its strength and other attributes. In the present research studies, this aspect (creep characteristics) on plain and reinforced concrete with and without GGBFS has been considered at the core apart from the other relevant properties. Pre-stressed concrete (PC) significantly differs from the reinforced concrete in the sense that availability of steel bars to arrest the micro-cracks due to temperature fluctuation of the environment is generally not present. From this point of view, pre-stressed concrete is induced with compressive force right from early age (3 days and beyond) so as to obviate any possibility of crack initiation and growth. As a result, pre-stressed concrete is kept under compressive stress for all its useful life. This compressive stress causes a lot of free water (moisture) to be squeezed out of the body of concrete with resultant shortening of its size (primarily length). Associated with this shortening of concrete, commonly referred to as creep, is the loss of pre-stressing force. Therefore, it becomes relevant to study the characteristics of creep in various types of concrete. The contributing factors to the phenomenon of creep are strength, porosity, water-cement ratio, and presence of fine material in the form of mineral admixture, temperature, humidity and the like. In the present studies, the impact of one parameter has been studied while keeping the remaining parameters constant. For this purpose, comprehensive experimental investigations have been carried out on the creep characteristics of plain and reinforced concrete with and without GGBFS. The basic material properties such as compressive 11 strength, tensile strength and modulus of elasticity of plain concrete with and without GGBFS were determined by carrying out tests on standard specimens in the form of cube, prism and cylinder at different ages. The creep and shrinkage strains are implicit which need to be isolated from each other by having observations on stressed specimens (creep with implicit shrinkage strains) as well as on non-stressed specimens (shrinkage strains only). Arithmetic subtraction of the latter from the first results in creep strain. This requires preparation of samples twice in number for each concrete parameter to be varied. This aspect enhances the experimental work substantially but is inevitable as it is. The creep and shrinkage strain studies were also conducted on cylindrical column specimens of plain concrete with and without GGBFS. Apart from the plain concrete exhibiting creep characteristics, the presence of steel in the body of concrete has been observed to have its own impact on the creep behaviour. All flexural structural members need to be appropriately reinforced in the tension zone and also sometimes in the compression zone. Deflection of such flexural members is considered as short term (instantaneous) and long term (time-dependent). The long-term deflection is observed to be much higher in magnitude than the one on account of short term. The long-term deflection thus becomes a matter of concern from viewpoint of fixing the limiting states of deflection, crack width and finally the integrity of the structural element itself. Creep is the major contributing factor to the long-term deflection. In the present research study, therefore, experiments have been carried out on reinforced concrete beams with and without GGBFS subjected to transverse loading of such magnitude as would not develop crack anywhere within the cross-section of the beam. Mid-span beam deflection under varying concrete strength, stress level, percent tension steel and concrete with and without GGBFS has been measured at pre-decided periods of time. Experimental details are briefly described herein with a view to indicating the parameters varied and properties studied. Concrete mixes of compressive strength 46.5, 37.0 and 27.0 MPa at 28 days were prepared. This is the range of strength commonly used for reinforced concrete. Including in more than three strengths demanded a very large number of specimens to be cast and tested for different studies which did not fit into the time-frame of the research study. The concrete mixes were also prepared with cement replacement of 20, 40 and 60 percent by GGBFS. Based on mix design considerations, the percentage of fine mineral admixture and the cement, together making a binder for the inert aggregates, has to be kept within certain proportion of the total materials. This requires equal amount of cement to be removed from the mix while adding GGBFS. As a result, the strength of concrete gets altered. Thus; the total of twelve numbers of different concrete mixes were used in the present investigation. The specimens were cured for the period of 3, 7 and 28 days and the tests were carried out at the ages of 3, 7, 28, 56, 90, 150 and 180 days to determine various strengths and modulus of elasticity. At the age of 3 and 7 days, specimens were tested to evaluate the early age development of compressive strength, split and flexural tensile strength, static and dynamic modulus of elasticity of plain and GGBFS based concrete. At 28 days, the specimens were tested to determine the 28 days compressive strength, split and flexural tensile strength, modulus of elasticity and load at first crack in cylindrical column and RC beams. After 28 days of curing, specimens were kept at room temperature and tested at the age of 56, 90, 150 and 180 days for the strengths and the modulus of elasticity. The ultrasonic pulse velocity technique was used for the determination of dynamic modulus of elasticity. The cylindrical column specimens were tested for determining the creep and shrinkage strain. The creep specimens were loaded axially by a force equal to fifty percent of the first crack load. The load was maintained for a period of 5 months. The creep and shrinkage strain measurements were recorded at the ages of 1, 3, 7, 14, 21, 28, 30, 56, 60, 90, 120 and 150 days. After 150 days of loading, the creep specimens were unloaded and creep recovery was recorded at the ages of 151, 153, 157, 164, 171, 178 and 180 days. The creep strain was thus calculated from the experimental data. Creep and shrinkage deflection of reinforced concrete simply supported beams were determined at mid span by testing under four point loading system. Prior to the IV application of sustained loading, the deflection at first crack load of the reinforced concrete beams without GGBFS was measured and different percentages of the first crack load were applied as sustained loading on RC beams for each mix. The intensity of load on RC beams to obtain creep deflection was divided into two series. In series-I, 25 percent of the first crack load was applied on the RC beams containing 20, 40 and 60 percent GGBFS and in series-II, 25 and 50 percent of the first crack load was applied on the RC beams without GGBFS. After 28 days of curing, beam specimens were tested and data for creep and shrinkage deflection were recorded for a period of five months under sustained loading at the ages of 1, 3, 7, 14, 21, 28, 30, 56, 60, 90, 120, 150 days. After five months, load was removed and creep deflection recovery was recorded for the period of one month i.e. at the ages of 151, 153, 157, 164, 171, 178 and 180 days. The creep deflection was thus calculated from the experimental data. The experimental results have been compared with the existing models for cube and cylinder compressive strength, flexural tensile strength, static modulus of elasticity, creep and shrinkage strain. The observed cube and cylinder compressive strength of plain and GGBFS based concrete has been compared with the age dependent strength prediction models of Indian Standard (IS-456), American Concrete Institute (ACI-209), Comite Euro-International du Beton -Federation Internationale de la Precontrainte (CEB-FIP) and Bezant's model (B3) respectively for all mixes. The experimentally obtained flexural tensile strength has been compared with the ACI-209 predicted model for all mixes. The empirical expressions for flexural strength of concrete given in IS-456, ACI-318, New Zealand Standard (NZS-3101) and Euro Code (EC-02) design codes have been compared with the experimentally obtained plain and GGBFS based flexural strength of concrete. A comparative study has been carried out between the observed static modulus of elasticity and ACI-209, CEB-FIP, British Standard (BS-8110), EC-02, Gardner and Lockman (GL2000) and B3 prediction models. Also, empirical expressions given in IS-456, ACI- 318, BS-8110, NZS-3101 and EC-02 design codes for static modulus of elasticity have been compared with the experimentally obtained results. A regression analysis has been carried out on the experimental data to obtain the timedependent relations for cube and cylinder compressive strength, split and flexural tensile strength, static and dynamic modulus of elasticity of plain and GGBFS based concrete. Further, semi-empirical correlations for (a) split tensile strength and compressive strength; (b) flexural tensile strength and compressive strength; (c) static modulus of elasticity and compressive strength of plain and GGBFS based concrete have been proposed. It has been observed from the present experimental investigations that there exists a linear relationship between static and dynamic modulus of elasticity for plain and GGBFS based concrete. Close agreement has been found between measured and predicted values obtained from the proposed correlations. A comparative study has also been conducted between experimentally obtained creep and shrinkage strain and strain predicted by ACI-209, CEB-FIP, B3 and GL2000 prediction models. With the help of multiple variable regression analysis of the experimental data of creep and shrinkage strain, the time-dependent hyperbolic, logarithmic and power expressions for plain and GGBFS based concrete have been proposed. Close agreement has been found between measured and predicted values of creep and shrinkage strain. The experimentally obtained creep and shrinkage deflection have been compared with deflection predicted using various design codes namely IS-456, ACI-318, BS-8110, NZS- 3101 and EC-02. vi