FLEXURAL FATIGUE BEHAVIOUR OF STEEL FIBRE REINFORCED CONCRETE

Abstract

Interest in the fatigue of concrete began at the turn of the century with the development of reinforced concrete bridges. Concrete bridges, off-shore structures and concrete pavements are the structures most influenced by the fatigue loading. Despite early interest in metal fatigue, interest in non-metallics, particularly plain and steel fibre reinforced concrete (SFRC) lagged behind for many years. Concrete fatigue investigations, like those on metals have also been motivated by practical problems. A critical review of the literature indicates that extensive studies have been carried out to investigate the fatigue characteristics of plain concrete. The concept of probabilistic analysis has been implemented and the Weibull probability distribution has been extensively examined to describe the fatigue test data of plain concrete. Most of the fatigue studies on SFRC are mainly confined to obtaining the S-N relationships and determination of endurance limits for different type/volume fraction/aspect ratio of fibres, though some studies focused attention on some other aspects. Although, the fatigue test data of SFRC shows considerable variability, even at a given stress level due to the random orientation of fibres in concrete, a little effort has been directed towards the application of probabilistic concepts to ensure adequate fatigue resistance of this material. The present investigation, therefore, was planned to study the flexural fatigue-life distributions of SFRC with different fibre contents for various fatigue stress levels and to apply the concept of equivalent fatigue-life as reported for plain II concrete, to SFRC to carry out the probabilistic analysis when both fatigue stress level S and fatigue stress ratio R are variables and to incorporate the survival probabilities into the fatigue equation S- fmax / fr - A(N)~b(i_r) by obtaining the coefficients of the above equation for SFRC with different fibre volume fractions (0.5%, 1.0% and 1.5%), corresponding to different survival probabilities, making this equation applicable to steel fibre reinforced concrete. It was also proposed to obtain the two-million cycle endurance limits corresponding to different fatigue stress ratios, for SFRC with varying fibre volume fractions. An extensive experimental investigation was planned in which 233 flexural fatigue tests and 196 static flexural tests were conducted on SFRC specimens of size 100 x 100 x 500 mm and containing corrugated Xerox type steel fibres of size 30 x 2.0 x 0.6 mm in three volume fractions i.e. 0.5%, 1.0% and 1.5%. The static flexural tests were conducted on a 100 kN INSTRON closed loop universal testing machine whereas, the flexural fatigue tests were conducted on a 100 kN MTS closed loop universal testing machine. The fatigue tests were conducted at different stress levels ranging from 0.95 to 0.60 and at three stress ratios i.e. 0.10, 0.20 and 0.30. The number of cycles to failure of each specimen under different load conditions was noted. The physical properties of the basic constituent materials viz., cement, sand and coarse aggregate were obtained as per relevant Indian Standard Specifications. The mix proportion used was 1:1.52:1.88 (by weight) with a water cement ratio of 0.46. Ill A probabilistic analysis of the fatigue test data obtained has been carried out lo achieve the objectives of the investigation. Some of the major conclusions drawn are given below. First, the flexural fatigue-life distributions of SFRC at different stress levels have been studied. It has been shown that the statistical distribution of fatigue-life of SFRC at a given stress level, can approximately be described by the two-parameter Weibull distribution, except at stress level S = 0.675, Vf = 1.5%, where the average value of the shape parameter for fatigue-life is less than unity. The Weibull distribution with shape parameter less than unity has a decreasing hazard function with time, which is contrary to the expected fatigue behaviour of engineering materials. The parameters of the distribution for different stress levels, for fatigue-life of SFRC with 0.5%, 1.0% and 1.5% fibre contents have been obtained by the graphical as well as by the method of moments. The shape parameter for fatigue-life of SFRC decreased with a decrease in the fatigue stress level and with an increase in the fibre content in concrete, thus indicating higher variability in the distribution of fatigue-life at lower fatigue stress levels and in concrete with higher fibre content. Secondly, the concept of equivalent fatigue-life has been employed to carry out probabilistic analysis when both the fatigue stress level and fatigue stress ratio are variables. Like fatigue-life of SFRC, the equivalent fatigue-life, at a particular stress level, have been found from the test data, to approximately follow the two-parameter Weibull distribution and the parameters of the distribution have been obtained by various methods. The coefficients of the fatigue equation / \—Bfl—Rl S = fmax / fr = A(N) v ' have been obtained corresponding to different survival IV probabilities for SFRC with all the three volume fractions of fibres tested in this investigation. This equation can be used to obtain the flexural fatigue strength of SFRC for the desired level of survival probability. Kolmogorov-Smirnov test has been applied as a goodness-of-fit test and it has further substantiated that the two-parameter Weibull distribution is a valid model for the description of fatigue-life as well as equivalent fatigue-life data of SFRC, at all the stress levels tested in this study. Third, the test results are presented both as S-N relationships, with the maximum fatigue stress expressed as a percentage of static ultimate flexural strength and as the relationships between actually applied maximum fatigue stress and number of cycles to failure. The two-million cycle endurance limits have been determined for different fatigue stress ratios for concrete containing different fibre volume fractions. Comparison of performance based on percentage S-N relationships give trends that differ significantly from the trends obtained when the comparisons are made based on applied maximum fatigue stress. The failure of almost all the specimens was due to the initiation and propagation of a single crack in the middle third span of the specimen. For specimens with 0.5% fibre content, the failure was due to pull-out of fibres, whereas, for specimens with 1.0% and 1.5% fibre contents, it was partly due to pull-out and partly due to breaking of fibres.