Construction and Building Materials 102 (2016) 782–791
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Analysis of flexural fatigue failure of concrete made with 100% Coarse Recycled Concrete Aggregates Sumit Arora, S.P. Singh ⇑ Department of Civil Engineering, Dr. B. R. Ambedkar National Institute of Technology, Jalandhar 144 011, India
h i g h l i g h t s Flexural fatigue performance of concrete made with 100% RCA has been investigated. Fatigue life distributions for concrete made with 100% RCA established. Higher variation in the fatigue life of concrete made with RCA has been observed. Lower two million cycles endurance limit for concrete made with RCA observed.
a r t i c l e
i n f o
Article history: Received 21 March 2015 Received in revised form 5 September 2015 Accepted 16 October 2015
Keywords: Endurance limit Fatigue life Recycled Concrete Aggregates Stress level
a b s t r a c t The paper presents results of an investigation carried out to analyze the flexural fatigue performance of concrete beams made with 100% Coarse Recycled Concrete Aggregates (RCA) and its comparison with that of concrete made with 100% Coarse Natural Aggregates (NA). The fatigue performance of concrete beams made with RCA as well as NA has been assessed in terms flexural fatigue life distributions and two-million cycle endurance limit. Experiments were conducted to obtain the flexural fatigue lives of concrete beam specimens made with 100% RCA as well as 100% NA under different stress levels and further compared with literature available on NA. Specimens of size 100 mm 100 mm 500 mm were tested under four point flexural fatigue loads applied at a frequency of 10 Hz. It has been shown that the fatigue life distributions of concrete mixes made with 100% RCA and 100% NA can be modeled by the two-parameter Weibull distribution. The values of the shape parameters of the Weibull distribution obtained for concrete made with RCA have been found to be smaller than that of concrete made with NA in present investigation and previous studies on NA, thus indicating higher variability in the distribution of flexural fatigue life of concrete made with RCA viz. a viz. concrete made with NA. The two-million cycles endurance limit for concrete made with 100% RCA has been found to be 50%, which is about 8% and 7% lower than that of concrete made with NA in present and previous studies respectively. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Concrete industry uses 12.6 billion tonnes of raw materials each year and thus is the largest user of the natural resources in the world [26]. The global market for construction aggregates is expected to increase 5.2% per year until 2015, up to 48.3 billion tonnes [8]. In the USA, the Environmental Protection Agency [9] estimated that the generation of debris, from construction, demolition, and renovation of residential and non-residential buildings in 2003, was close to 170 million tonnes. According to Eurostat [11], the total amount of waste generated in the European Union, in 2010, was over 2.5 billion tonnes, of which almost 860 million ⇑ Corresponding author. E-mail address:
[email protected] (S.P. Singh). http://dx.doi.org/10.1016/j.conbuildmat.2015.10.098 0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
tones belonged to construction and demolition activities. India produces 23.75 million tonnes of construction and demolition waste annually. Dumping of construction and demolition waste requires lot of space and is becoming a severe environmental problem. Since maintenance and protection of environment is vital for the survival of human race, the demolished concrete waste materials can be converted into valuable coarse aggregates by breaking or crushing into suitable sizes, rather than dumping these on open land. About 75% aggregates are required for the production of concrete, out of which coarse aggregates form 65%. It may be noted that the production of one tonne of Coarse Natural Aggregates (NA) results the emission of 4600 tonnes of carbon, whereas, production of one tonne of Coarse Recycled Concrete Aggregates (RCA) produces 2400 tonnes of carbon. Considering the global consumption
S. Arora, S.P. Singh / Construction and Building Materials 102 (2016) 782–791
783
Notations
a r U()
l
CoV fmax fmin fr
shape parameter standard deviation of the data sample under consideration gamma function mean of the data sample coefficient of variation of the data sample at a given stress level maximum fatigue stress minimum fatigue stress static flexural strength
of 10 billion tonnes per year of aggregates for concrete production, the carbon footprint can be determined for NA as well as for RCA [33]. When compared to NA, the density of RCA is generally lower as the density of adhered mortar is less than the underlying rocks [25]. As reported earlier, there is about 17% difference between the bulk densities of RCA and NA, with values of 2394 and 2890 kg/m3 respectively. Some authors have reported the density of RCA to be 7–9% lower than that of NA [24,37]. Also the water absorption of RCA and NA has been reported to be 4.9–5.2% and 1.0–2.5% respectively. The gradation curves for RCA lie in the range which is required for aggregates to be accepted for use in the concrete production [38]. The strength features of source concrete, effectiveness of crushing procedure, crushing method and particle size of RCA are signified by the amount of adhered mortar. The quality of RCA is inversely related to this adhered mortar [5]. The presence of attached mortar makes the RCA surface texture rough with an irregular shape. RCA possesses inferior mechanical properties such as low crushing strength, low impact resistance and low abrasion resistance than natural aggregate [17,24,37]. The characteristics and the amount of RCA in concrete can influence the strength properties of concrete. It has been reported that concrete made by replacing of 25% NA with RCA has comparable properties to concrete made with 100% NA at the same w/c ratio. However, to obtain the same compressive strength with 50–100% replacement of NA with RCA, w/c ratio needs to be lowered by 4–10% [10]. It is reported that the percentage reduction in the compressive strength of concrete made with 100% RCA ranges from 15% to 30% compared to that of concrete made with 100% NA [6,7,22]. It has also been observed that the splitting tensile strength of concrete made with RCA is less affected by RCA content. Investigations indicated that concrete made with RCA either shows comparable or superior split tensile strength than that of concrete made with NA [50,51]. This enhanced performance in tensile strength is attributed to the increased absorption by adhered mortar layer on recycled aggregates as well as an effective ITZ, consequently improving the bond between aggregates and the mortar matrix [10]. Improvement in tensile strength of concrete made with RCA has also been correlated to the higher strength of concrete source from which RCA is produced [52]. Majority of the research reported in literature on concrete made with RCA has been directed on rheological properties of fresh and mechanical properties of hardened concrete under statically applied loads. Concrete made with RCA can offer huge benefit to the construction industry by reducing the overall cost of the project. The applications of concrete made with RCA can be in bridge decks and piers, pavements, high rise buildings, rapid transport systems, tunnel linings, dams, precast structural elements, storage tanks, etc. Many of these structures are influenced by the fatigue
LN N Pf R S u RCA NA
survival probability number of cycles to failure probability of failure stress ratio = fmin/fmax stress level = fmax/fr scale parameter Coarse Recycled Concrete Aggregates Coarse Natural Aggregates
loading, thus necessitating the need to investigate the performance concrete under fatigue loads. For the last many decades, several research studies have been conducted to investigate the fatigue behavior of concrete made with NA. Major properties investigated are S-N relationships; Weibull parameters; endurance limit; mean and design fatigue lives [1]. Analysis of various prediction models were carried out in these investigations for evaluating the fatigue performance of concrete. It has been established that the range of stress influences the fatigue strength of concrete considerably [3,19,28,30,31]. The influence of the frequency of loading has been investigated by a number of researchers [2,3,4,12,21,35,43]. In general, variationinthe loading frequency in the range of 1–30 Hz has insignificant effect on the fatigue behavior of concrete if the maximum stress level is kept less than about 75% of the static strength [3,36]. Reduction of variability in the distribution of fatigue life of concrete containing cement additives as compared to control concrete has been reported [20]. It was investigated that at 75% of the maximum flexural stress level, the number of repetitions to failure was 2000 and 20,000 cycles for lean mixtures of cement and cement-fly ash concretes, respectively [13]. Similarly, it was reported that concrete with equivalent or higher compressive and fatigue strength could be obtained with cement replacement of 25% by weight of low-calcium fly ash or 50% by weight of high-calcium fly ash [45]. It has also been evaluated that the fatigue limit values were 0.65 for no-fly ash concrete and 0.70 for high-volume Class F fly ash concrete. Moreover, the plain concrete had compressive fatigue strength ratio of 56% in air at the same level of survival [32]. Static flexural and flexural fatigue strength of concrete have been found to decrease when cement replacement with fly ash increased from 15% to 55% by weight. The decrease was smaller for the 15% fly ash mixture [29]. However, very limited research has been carried out on the fatigue performance of concrete made with RCA. It has been shown that the endurance limit decreased as expected with increase in replacement of NA with RCA [18,44,47,48]. It may be noted that relatively less number of specimens have been tested in these investigations. Due to statistical nature of fatigue phenomenon, large variability usually occurs in the fatigue life data of concrete, at a given stress level, even under carefully controlled test procedures. Therefore, in investigations wherein the probabilistic analysis of the fatigue data is the prime objective, as in this work, it is desirable to test relatively more number of specimens at a given stress level to obtain data which is statistically significant. This approach has been adopted by previous investigators i.e. [27,31,40,42] for concrete made with NA. Keeping in view the wide potential of demolished concrete as source of RCA, the investigations on its fatigue behavior still lags behind and thus the present investigation has been carried out to
784
S. Arora, S.P. Singh / Construction and Building Materials 102 (2016) 782–791
evaluate the flexural fatigue performance of concrete beams made with 100% RCA. The research also aims to verify the effectiveness of RCA for possible practical applications by comparing the obtained results with that of NA used in present work and also NA used by previous researchers. 2. Research significance A brief review of literature indicates that considerable research work has been carried out to investigate the flexural fatigue behavior of concrete made with NA. It can also be observed that a lot of data has been generated on the mechanical properties of concrete made with RCA under statically applied loads. However, little information is available on the flexural fatigue characteristics of concrete made with RCA and therefore, one is prompted to work in this direction. This is important in the present scenario keeping in view the anticipated use of concrete made with RCA in civil engineering applications. Thus, the present investigation was designed to evaluate the flexural fatigue performance of concrete made with 100% RCA and to compare the same with concrete made with 100% NA. Two-parameter Weibull distribution has been verified at different stress levels to establish the probability distributions of concrete made with RCA and NA respectively. Further, the two-million cyc les fatigue strength/endurance limit of concrete made with RCA has been estimated from the S–N diagrams. 3. Experimental procedure A total of 64 flexural fatigue tests and 48 complementary static flexural tests were executed on beam specimens of size 100 mm 100 mm 500 mm under four point flexural loading. Compressive strength tests were also conducted on different batches of concrete to check the quality of each batch after 28 days of curing. 4. Materials and mix proportions Grade 43, Ordinary Portland Cement (PC) has been used in the study. Wellgraded RCA with maximum size of 12.5 mm obtained by crushing the concrete specimens available in the Concrete Laboratories of this Institute were used. Natural aggregates with comparable grading was procured from local market. The grading of RCA and NA was purposely made comparable to that of NA used in previous studies [14,27] as it was proposed to compare the results with these investigations. Fig. 1a shows the grading of RCA and NA used in current study and grading of NA used in previous studies. Table 1 shows the physical properties of RCA and NA used.
Table 1 Physical properties RCA. Aggregates used
Fineness modulus
Specific gravity
Aggregate impact value (%)
Aggregate crushing value (%)
Water absorption (%)
RCA NA
6.70 6.93
2.46 2.64
30.43 16.35
25.6 15.8
5.35 0.68
Locally available coarse sand was used as fine aggregates. The grading curve of the fine aggregates is given in Fig. 1b. Class F fly ash was used as partial replacement of PC. A polycarboxylic ether based superplasticizer was used as chemical admixture in suitable dosages to obtain the required workability of concrete mixes made with RCA and NA. The mix proportions for RCA and NA concrete mixes used for casting the specimens are shown in Table 2. Since the prime objective of this investigation is to evaluate the flexural fatigue performance of concrete made with 100% RCA compared to concrete made with 100% NA, therefore, to isolate this, it was thought prudent to only change the quantity of NA with RCA while keeping all the other mix proportions same (such as w/c ratio and quantity of fine aggregates). The static compressive strength and static flexural strength of different batches of concrete made with RCA were monitored and the reduction with respect to concrete made with NA was quantified. Such approach has also been adopted by other researchers [23,34,49].
5. Casting of specimens In this investigation, slump test was performed to control the workability of all the mixes and obtained slump was in the range of 65–90 mm. The specimens were cast in batches, wherein each batch consisted of three cube specimens of size 150 mm 150 mm 150 mm for determining 28-days compressive strength and seven standard beam specimens of size 100 mm 100 mm 500 mm for static flexural and flexural fatigue tests. The mixes were prepared in a drum mixer and poured into the standard molds followed by vibration at 3600 r.p.m to ensure good compaction. The specimens were demolded after 24 h of casting and kept in water for curing at standard temperature of 27 ± 2 °C until tested for flexure or compression. The average compressive strengths obtained after 28 days of curing for concrete specimens made with RCA and NA were 31.70 MPa and 41.77 MPa respectively. Batch-wise compressive strength of concrete made with RCA and NA after 28 days of curing is summarized in Table 3. The beam specimens were water cured under laboratory conditions for 90 days and stored in the laboratory environment for another 60 days in order to avoid possible increase in the strength during the fatigue tests.
110 Concrete made with 100% NA (Goel et al. 2012)
100 90
70 60 50 40 30 20 10 0 3
30
% Passing
80
Concrete made with100% NA (Present Study) Concrete made with 100% NA (Mohammadi and Kaushik 2005) Concrete made with 100% RCA (Present Study)
Sieve Size (log scale) (mm) Fig. 1a. Comparison of grading curve of RCA with grading curve of NA in present study as well in previous investigations.
785
S. Arora, S.P. Singh / Construction and Building Materials 102 (2016) 782–791
120 100
60
% Passing
80
40 20 0 0.1
1 Sieve Size (log scale) (mm) Fig. 1b. Grading curve of fine aggregates used in present study.
Table 2 Mix proportions per cum of concrete made with RCA and NA. Aggregate used
Cement (kg)
Fly ash (kg)
Fine aggregates (kg)
Coarse aggregates (kg)
Water (L)
RCA NA
343 343
148 148
762 762
935 1003
206 206
6. Static flexural and flexural fatigue testing To determine the maximum and minimum load limits to be applied during flexural fatigue tests on the specimens, static flexural tests were conducted on the concrete specimens made with RCA prior to fatigue tests. At least three specimens from each batch were tested to determine the static flexural strength of that batch. The average value of static flexural strengths obtained for concrete made with RCA and concrete made with NA were 4.53 MPa and 5.1 MPa respectively, and the same has been tabulated batchwise in Table 3. The graphical representation of the average compressive strength and static flexural strength showing average standard deviation for both the concrete mixes tested in the present study has been given in Fig. 4. From the graph it can be observed that the values of standard deviation corresponding to compressive strength and static flexural strength for concrete made with 100% RCA were 1.35 MPa (max. = 2.79 MPa and min. = 0.86 MPa) and 0.44 MPa (max. = 0.69 MPa and min. = 0.11 MPa) respectively. Similarly the values of standard deviation for compressive strength and static flexural strength for concrete made with 100% NA were calculated to be as 1.05 MPa (max. = 1.35 MPa and min. = 0.62 MPa) and 0.36 MPa (max. = 0.66 MPa and min. = 0.03 MPa) respectively. The remaining four specimens Table 3 Compressive strength and static flexural strength test results of concrete made with RCA and NA (present investigation). Batch No.
1 2 3 4 5 6 7 8 Average
28 days compressive strength (MPa)
Static flexural strength (MPa)
RCA
NA
RCA
NA
30.24 32.09 34.52 32.87 31.66 31.10 30.56 30.47 31.7
35.40 40.69 36.83 46.24 41.54 44.37 47.62 41.44 41.77
4.59 4.84 4.17 4.89 4.66 4.29 3.98 4.79 4.53
5.953 4.897 5.286 4.454 4.888 5.217 5.050 5.024 5.1
from a particular batch were tested for flexural fatigue once the static flexural strength was evaluated for each batch. The static flexural and flexural fatigue tests were conducted on a 100 kN Servo-controlled Actuator. The loading points in flexural fatigue tests were kept the same as in static flexural tests. The flexural tests were conducted at different stress levels S (S = fmax/fr), ranging from 0.85 to 0.55. The fatigue stress ratio R (R = fmax/fmin) was kept constant at 0.10 throughout the investigation. All the tests were conducted at constant amplitude in the form of sinusoidal loads applied at a frequency of 10 Hz. The cycle counter of the flexural testing machine displayed the number of cycles to the failure of each specimen under fatigue loading. In the present investigation, an upper limit of 2 106 cycles of fatigue loading was chosen as the number of specimens to be tested was large [15,27,41]. As soon as the failure of the beam specimen occurred or it reached the fixed upper limit of two million cycles, the test was terminated. Table 4 shows the number of specimens tested at each stress level and the corresponding flexural fatigue results of concrete made with RCA and NA. 7. Determination of fatigue life distributions 7.1. Graphical method for verification of Weibull distribution Firstly the graphical method has been employed to verify that the fatigue life data of concrete made with 100% RCA and 100% NA can be modeled by the two-parameter Weibull distribution at a given stress level. Subsequently, the graphical method, method of moments and method of maximum likelihood have been used to determine the parameters of the Weibull distribution. The reliability function LN (n)of the two-parameter Weibull distribution may be written as follows [14,16,27,30,31,46]:
LN ðnÞ ¼ exp
hnia u
ð1Þ
where n is the specific value of random variable N; a = shape parameter at stress level S; u = characteristic life at stress level S. Taking the logarithm twice on both sides of Eq. (1):
786
S. Arora, S.P. Singh / Construction and Building Materials 102 (2016) 782–791
Table 4 Laboratory fatigue life data (number of cycles to failure N, in ascending order) for concrete made with RCA and NA (present investigation).
Stress level S?
Specimen No.
0.85
0.75
0.65
0.55
RCA
1 2 3 4 5 6 7 8 9 10
567 789 1054 1188 1345 1765 1897 2098 2156 2354
192a 4353 5615 9382 9792 12,829 13,702 14,045 23,020 26,079
67,225 68,738 88,969 90,371 120,805 189,763 249,867 261,009 319,551 409,876
478640b 567390b 763984b 1167919b – – – – – –
1 2 3 4 5 6 7 8 9 10
444a 1137 1367 1678 1945 2271 2605 2647 3096 3987
10,781 13,879 18,489 21,945 25,467 31,256 36,543 42,842 46,951 51,348
100,801 142,054 187,623 220,075 260,685 323,068 360,845 456,944 512,089 558,973
– – – – – – – – – –
i kþ1
-1.0 -1.5 S 0.85
-2.5 6.5
7.0
7.5
8.0
Fig. 2a. Graphical analysis of fatigue life data of concrete made with RCA at stress levels S = 0.85.
1.5
ð2Þ
Value of ln (ln(1/LN))
1.0
ð3Þ
where, Y = ln[ln(1/LN)], X = ln (n), and b = a ln u. Eq. (2) represents a linear relationship between ln[ln(1/LN)] and ln(n), which can be used to verify the suitability of the twoparameter Weibull distribution for the description of fatigue life data of concrete made with RCA and NA. To do so, the fatigue life data at a given stress level must be first arranged in ascending order and the empirical survivorship function LN is then obtained from the following expression [14,16,27,30,31,46]:
LN ¼ 1
-0.5
Value of ln (N)
or;
Y ¼ aX b
0.0
6.0
– No specimen tested. a Rejected as outlier by Chauvenet’s Criterion, not included in analysis. b Used for S–N curves only.
1 ¼ a ln n aln u ln ln LN
0.5
-2.0
0.0 -0.5 -1.0 -1.5 S 0.75
-2.5 8.0
8.5
9.0
9.5
10.0
10.5
Value of ln (N) Fig. 2b. Graphical analysis of fatigue life data of concrete made with RCA at stress levels S = 0.75.
1.5
ð4Þ
where, i = failure order number, and k = number of fatigue data points at a given stress level S. A graph is plotted between the ln [ln(1/LN)] and ln(N)and if the test data, at a particular stress level, follows approximately straight line, it can be observed that twoparameter Weibull distribution is a reasonable assumption for the description of fatigue test data at that stress level. Figs. 2a and 3a show the plot of the fatigue life data of concrete made with RCA and NA respectively, at stress levels (S = 0.85). It can be observed that the data points fall approximately along a straight line, indicating that the two-parameter Weibull distribution is reasonably valid for the distribution of fatigue life of concrete made with RCA and NA at the considered stress level. Similarly, Fig. 2b, Fig. 2c, Figs. 3b and 3c represent the analysis of the fatigue life data of concrete made with RCA and NA at stress levels 0.75 and 0.65 respectively. The corresponding values of the correlation coefficient, Cc, are 0.989, 0.982 and 0.957 for concrete made with RCA and 0.99, 0.994 and 0.996 for concrete made with NA, at stress levels 0.85, 0.75 and 0.65 respectively. The values of the Weibull parameters obtained by this method are a = 2.090 and u = 1761; and a = 2.442 and u = 2619 for concrete made with RCA and concrete made with NA respectively at stress level, S = 0.85. Similarly, the values of Weibull parameters (a and u) obtained at other stress levels i.e., S = 0.75 and 0.65 have been tabulated in the Table 5.
r = 0.98 α = 1.647 u = 15395
0.5
-2.0
1.0 Value of ln (ln(1/LN))
NA
r = 0.99 α = 2.09 u = 1761
1.0 Value of ln (ln(1/LN))
Fatigue life data ‘N’
1.5
0.5
r = 0.96 α = 1.424 u = 218895
0.0 -0.5 -1.0 -1.5 S 0.65
-2.0 -2.5 10.0
11.0
12.0
13.0
14.0
Value of ln (N) Fig. 2c. Graphical analysis of fatigue life data of concrete made with RCA at stress levels S = 0.65.
7.2. Distribution parameters from method of moments Shape parameter a and scale parameter u can also be calculated using the following relationships [14,27,31,46]:
a ¼ ðCOVÞ1:08
ð5Þ
787
S. Arora, S.P. Singh / Construction and Building Materials 102 (2016) 782–791
1.5
r = 0.99 α = 2.442 u = 2619
1.0 Value of ln(ln(1/LN)
h ¼
0.5 0.0 -0.5 -1.0 -1.5
S 0.85
-2.0 -2.5 6.5
7.0
7.5
8.0
8.5
Value of ln(N) Fig. 3a. Graphical analysis of fatigue life data of concrete made with NA at stress levels S = 0.85.
and,
u ¼ 1
l
ð6Þ
C aþ1
where, l is the sample mean of the fatigue life data at a given stress level, COV (=r/l, r is standard deviation of data sample) is the coefficient of variation of the data, and U is the gamma function. The parameters obtained by the method of moments for concrete made with RCA are a = 2.6458 and u = 1716 and a = 2.7711 and u = 2540 for concrete made with NA respectively at stress level S = 0.85. The estimated values of the Weibull parameters at other stress levels i.e., S = 0.75 and 0.65 are listed in Table 5. 7.3. Distribution parameters from maximum likelihood estimation The probability density function of Weibull distribution may be written as follows:
f N ðnÞ ¼
a h
na1 exp
na h
ð7Þ
in which;
h ¼ ua
ð8Þ
The maximum likelihood function may be then expressed as:
Pk
a
i¼1 ðni ln ðni ÞÞ Pk a i¼1 ni
1
a
¼
k 1X ln n i k i¼1
ð9Þ
k 1X na k i¼1 i
where, a* and h* are the maximum likelihood estimators of ‘a’ and ‘h’ respectively. The parameter ‘a’ may now be obtained from Eq. (9) by a simple computer program based on iterative procedure. The parameter ‘u’ is then estimated from the Eq. (6); i.e., u = h1/a. The values estimated, for concrete made with RCA, from the method of maximum likelihood are a = 2.934 and u = 1711 for stress levels S = 0.85. Similarly, with the same method, obtained values of Weibull distribution parameters for concrete made with NA are a = 2.9505 and u = 2587 at stress level S = 0.85 and the values of Weibull parameters at other stress levels have been tabulated in Table 5. The values of the parameters estimated by various methods for concrete made with RCA as well as NA are listed in Table 5 along with their average values. A comparison of parameters obtained in the present investigation for concrete made with RCA has been made with that of the concrete made with NA in the present and previous studies. The comparisons are shown in Table 6 and Figs. 5–8. The results of the present study on static flexural strength of concrete made with RCA and NA as well as the previous studies on concrete made with NA are plotted in Fig. 5 for the purpose of comparison. The values of shape parameter ‘a’ determined for the fatigue life concrete made with RCA at stress levels of 0.85, 0.75 and 0.65 are 2.5566, 1.8667 and 1.5984 respectively, which is compared with those of concrete made with NA in current and previous studies and has been plotted in Fig. 6. In the present investigation, the values of the Coefficient of Variation (COV) of the fatigue life data of concrete made with RCA has been found to be 41.93%, 56.10% and 64.77% at stress levels 0.85, 0.75 and 0.65 respectively. Similarly COV values of the fatigue life data of concrete made with NA are found to be 39.58%, 48.39% and 51.84% at stress levels (S) 0.85, 0.75 and 0.65 respectively. The calculated values of COV for the fatigue life data of concrete made with RCA have been compared with that of concrete made with NA in the present study as well as on the previous data available on concrete made with NA, in Table 6. The corresponding increase in COV of the fatigue life data of concrete made with RCA is 45.38% and 34.75% (compared with that of [31], 30.3% and 27.70% (compared with that of [14] and 5.94% and 24.94% (compared to the present study on NA) at stress levels 0.85 and 0.65 respectively. The data has been plotted in Fig. 7 for the sake of comparison. This indicates that concrete made with 100% RCA has higher variability in the distribution of fatigue life at different stress levels as compared to concrete containing NA.
1.5
1.5 r = 0.995 α = 1.881 u = 34794
0.5
r = 0.996 α = 1.76 u = 362629
1 Value of ln(ln(1/LN))
1 Value of ln(ln(1/LN))
ð10Þ
0 -0.5 -1 -1.5 S 0.75
-2
0.5 0 -0.5 -1 -1.5 -2
S 0.65
-2.5 -3
-2.5 9
9.5
10
10.5
11
Value of ln(N) Fig. 3b. Graphical analysis of fatigue life data of concrete made with NA at stress levels S = 0.75.
11
11.5
12
12.5
13
13.5
Value of ln(N) Fig. 3c. Graphical analysis of fatigue life data of concrete made with NA at stress levels S = 0.65.
788
S. Arora, S.P. Singh / Construction and Building Materials 102 (2016) 782–791
difference (D1) of 0.1108 and 0.1208 is observed in the case of concrete made with RCA and NA respectively. The critical value for n = 10 and 5 percent significance level is found to be 0.41 from the Kolmogorov–Smirnov table. Thus, the applied model must be accepted at the 5 percent level of significance. The Kolmogorov–Smirnov test was also conducted for other stress levels i.e. 0.75 and 0.65 and it was found that the model is acceptable at 5% level of significance.
Compressive/ Flexural Strength (MPa)
45 40 35 30 25 Compressive strength
20 15
7.5. Two-million cycles endurance limit As shown in previous sections, the probability distributions for the fatigue life of concretes made with RCA and NA have been examined at various fatigue stress levels. The results show that there is increase in the variability in the distribution of fatigue life of concrete made with RCA compared to concrete made with NA. To determine the two-million cycle fatigue strength of concrete made with RCA and NA, a linear regression analysis of each set of data has been carried out and plotted in a semi-logarithmic format in the form of S–N curves. Fig. 8 represents the S–N relationship for the concrete mixes made with RCA and NA, in which the ordinate represents the maximum fatigue stress expressed as a percentage of the corresponding static flexural strength. Assuming conservatively that the test duration of two-million cycles represents safe life, the fatigue strength of concrete made with RCA can be evaluated from the S-N curve given in Fig. 8. The two-million cycles fatigue strength for concrete made with RCA has been found to be about 50% of the corresponding static flexural strength which is about 8% less than that of the concrete made with NA in present study. The fatigue test data of concrete made with NA used in present study and the data taken from literature [14,27,31,39] has also been plotted in Fig. 8 for the purpose of comparison. The twomillion cycles fatigue strength of concrete made with NA has been reported to be 58% of its static flexural strength in the present study and about 57% as reported in literature [14,27,31,39]. Thus concrete made with 100% RCA has been found to perform poorly compared to concrete made with NA, necessitating the need
Static Flexural Strength
10 5 0 100% RCA
100% NA
Fig. 4. Results of average compressive strength and average flexural strength of concrete made with 100%RCA and 100%NA.
7.4. Goodness-of-fit test The present investigation shows that Weibull distribution can be used to describe the distribution of fatigue life of concrete made with RCA. To supplement it further, Goodness-of-fit test has also been performed, apart from the graphical method. For this purpose, Kolmogorov–Smirnov Test [53] has been applied, which can be conducted by defining a parameter D by using the following expression:
D¼ki¼1 max½jF ðxi Þ F N ðxi Þj
ð11Þ
⁄
in which F (xi) = i/k = observed cumulative histogram; i = failure order number of the data point; k = total number of data points in the sample under consideration at a given stress level. A maximum
Table 5 Values of shape parameters a and characteristic life u for the fatigue life data of concretes made with RCA and NA. Aggregates used
Methods
S = 0.85
S = 0.75
S = 0.65
a
u
a
u
a
u
RCA
Graphical method Method of moments Maximum likelihood Average
2.09 2.646 2.934 2.556
1761 1715 1712 1729
1.647 1.898 2.055 1.867
15,395 14,888 14,971 15,085
1.424 1.620 1.752 1.598
218,895 210,449 210,922 213,422
NA
Graphical method Method of moments Maximum likelihood Average
2.442 2.7711 2.9505 2.7212
2619 2540 2587 2582
1.8810 2.2491 2.4405 2.1902
34,794 33,801 33,909 34,168
1.760 2.0879 2.2515 2.0331
362,629 349,097 353,990 355,239
Table 6 Comparison of COV values for the fatigue life of concrete made with RCA with the concrete made of NA in present investigation as well as in previous studies. Stress level ‘S’
0.85 0.80 0.75 0.70 0.65 – Data not available.
Oh [31]
Mohammadi and Kaushik [27]
Goel et al. [14]
Present study on concrete made with NA
Present study on concrete made with RCA
COV (%) 28.84 – 41.64 – 48.93
COV (%) 30.98 38.82 – 47.40 –
COV (%) 32.17 39.16 43.93 48.14 51.18
COV (%) 39.58 – 48.39 – 51.84
COV (%) 41.93 – 56.11 – 64.77
Average Flexural Strength
S. Arora, S.P. Singh / Construction and Building Materials 102 (2016) 782–791
789
models which have been used to predict the flexural fatigue strength of concrete are presently being examined by incorporating probability of failure into the fatigue test data of concrete made with RCA and the results will be reported in a separate paper.
5.5 5 4.5
8. Conclusion
4 Flexural fatigue tests were conducted on beams specimens of concrete made with 100% RCA with the intention to evaluate its flexural fatigue performance with reference to concrete made with NA. The flexural fatigue performance has been evaluated in terms of distribution of fatigue life data and two million cycles fatigue strength/endurance limit of the concretes made with RCA and NA. The test results indicate that probabilistic distribution of fatigue life of concrete mixes made with RCA and NA, at a particular stress level S, can be approximately modeled by the twoparameter Weibull distribution. Lower values of the shape parameter for fatigue life were obtained for concrete made with RCA compared to concrete containing NA, thus indicating higher variability in the distribution of flexural fatigue life of concrete made with RCA. The maximum increase in the COV of fatigue life data of concrete made with RCA has been found to be 45.38% at a stress
Fig. 5. Comparison of average static flexural strength of concrete made with RCA with present and previous studies on concrete made with NA.
for improving the flexural fatigue performance of concrete made with RCA. The work in this direction is currently in progress in the author’s Institute with the use of cement additions such as metakaolin and silica fume as partial replacement of PC. The
Value of Shape Parameter (α)
4
3.5
Concrete made with 100% RCA (Present Study) Concrete made with 100% NA (Goel et al. 2012)
3
Concrete made with 100% NA (Moh. and Kaushik 2005)
2.5
Concrete made with 100% NA (Oh 1991)
2
Concrete made with 100% NA (Present Study)
1.5 0.6
0.65
0.7
0.75
0.8
0.85
0.9
Stress Level S
Percentage Increase in COV
Fig. 6. Comparison of shape parameter ‘a’ of concrete made with RCA with present and previous studies on concrete made with NA.
50 45 40 35 30 25 20 15 10 5 0
Concrete made with 100% NA (Goel et al. 2012) Concrete made with 100% NA (Oh 1991) concrete made with 100% NA (Present Study)
0.65
0.75
0.85
Stress Level (S) Fig. 7. Percentage increase in coefficient of variation (COV) of concrete made with RCA with respect to present and previous data on concrete made with NA.
790
S. Arora, S.P. Singh / Construction and Building Materials 102 (2016) 782–791
0.9 0.85 Stress Level (S)
Concrete made with 100% NA (Present Study) Concrete made with 100% NA (Goel et al. 2012) Concrete made with 100% RCA (Present Study) Concrete made with 100% NA (Oh 1991) Concrete made with 100% NA (Mohammadi and Kaushik 2005) Linear (Concrete made with 100% NA (Present Study)) Linear (Concrete made with 100% NA (Goel et al. 2012)) Linear (Concrete made with 100% RCA (Present Study)) Linear (Concrete made with 100% NA (Oh 1991)) Linear (Concrete made with 100% NA (Mohammadi and Kaushik 2005))
r = 0.996 (Present Study RCA) r = 0.999 (Present Study NA) r = 0.999 (Oh 1991) r = 0.997 (Goel 2012) r = 0.996 (Mohammadi & Kaushik 2005)
0.8 0.75 0.7 0.65 0.6 3
4
5
6
Value of Log 10 N Fig. 8. Comparison of endurance limit of concrete made with RCA with the concrete made with NA in present as well as previous studies.
level of 0.85 compared to that of concrete made with NA. Incorporation of 100% RCA in concrete in place of NA also resulted in poor fatigue performance in terms of two million cycles endurance limit/fatigue strength. The two million cycle endurance limit for concrete made with 100% RCA has been found to be approximately 50% of the static flexural strength, whereas it has been found to be 58% for concrete made with NA in the present investigation and about 57% as reported in literature. Acknowledgements The authors would like to gratefully acknowledge the support of the staff of Structures Testing Laboratory at Dr B R Ambedkar National Institute of Technology, Jalandhar, India during the experimentation work reported in the paper. The financial support from MHRD to carry out the present research work is gratefully acknowledged. References [1] ACI Committee 215, Consideration for design of concrete structures subjected to fatigue loading, ACI Manual of Concrete Practice, Part 1, 1990, pp. 215R–1 to 215R–25. [2] B.A. Assimacopoulos, R.F. Warner, C.E. Ekberg, High speed fatigue tests on small specimens of plain concrete, J. Prestr. Concrete Inst. 4 (2) (1959) 53–70. [3] M.E. Awad, Strength and deformation characteristics of plain concrete subjected to high repeated and sustained loads (Ph.D. thesis), University of Illinois at Urbana, Champaign, Illinois, 1971. [4] M.E. Awad, H.K. Hilsdorf, Strength and deformation characteristics of plain concrete subjected to high repeated and sustained loads, in: Fatigue of Concrete, ACI Special Publication, SP-41, 1974, pp. 1–13. [5] M. Behera, S.K. Bhattacharyya, A.K. Minocha, R. Deoliya, S. Maiti, Recycled aggregates from C & D waste and its use in concrete – a breakthrough towards sustainability in construction sector: a review, Constr. Build. Mater. 68 (2014) 501–516. [6] H.J. Chen, T. Yen, K.H. Chen, Use of building rubbles as coarse aggregates, Cem. Concr. Res. 33 (2003) 125–132. [7] Z.H. Duan, C.S. Poon, Properties of recycled aggregate concrete made with recycled aggregates with different amounts of old adhered mortars, Mater. Design 58 (2014) 19–29. [8] EPA, Environmental Protection Agency. Available in:
(accessed January 2014). [9] Eurostat, Waste statistics in Europe. Available in: (accessed in July 2014). [10] M. Exteberria, E. Vasquez, A.R. Mari, Influence of amount of recycled coarse aggregates and production process on properties of recycled aggregate concrete, Cem. Concr. Res. 37 (2007) 735–742. [11] Freedonia, World construction aggregates, Industry Study No. 2838: The Freedonia Group, 2012, p. 334. [12] J.W. Galloway, K.D. Raithby, Effects of rate of loading on flexural strength and fatigue performance of concrete, Department of the Environment, TRRL Report LR 547, Crowthorne, Transport and Road Research Laboratory, 1974.
[13] R.K. Ghosh, K.L. Sethi, V.P. Arora, Laboratory studies on lean cement-fly ash concrete as construction material, IRC Highway Research Board, No. 19, 1982, pp. 13–26. [14] S. Goel, S.P. Singh, P. Singh, Flexural fatigue strength and failure probability of self compacting fibre reinforced concrete beams, Eng. Struct. 40 (2012) 131– 140. [15] S. Goel, S.P. Singh, Fatigue performance of plain and steel fibre reinforced self compacting concrete using S–N relationship, Eng. Struct. 74 (2014) 65–73. [16] E.J. Gumble, Parameters in distribution of fatigue life, J. Eng. Mech. ASCE (1963) 45–63. [17] T.C. Hanson (Ed.), Recycling of demolished concrete and masonary, Taylor and Francis, Oxfordshire, UK, 1992. [18] M. Heeralal, P.K. Rathish, Y.V. Rao, Flexural fatigue characteristics of steel fiber reinforced recycled aggregate concrete (SFRRAC), Arch. Civil Eng. 7 (1) (2009) 19–33. [19] H.K. Hilsdorf, C.E. Kesler, Fatigue strength of concrete under varying flexural stresses, ACI J. Proc. 63 (10) (1966) 1059–1076. [20] G. Kaur, S.P. Singh, S.K. Kaushik, Flexural fatigue strength of steel fibre reinforced concrete containing blends of limestone powder and silica fume, Int. J. Emerg. Technol. Adv. Eng. 2 (6) (2013) 2250–2259. [21] C.E. Kesler, Effect of speed of testing on flexural fatigue strength of plain concrete, Highway Res. Board 32 (1953) 251–258. [22] R. Kumutha, K. Vijai, Strength of concrete incorporating aggregates recycled from demolition waste, ARPN J. Eng. Appl. Sci. 5 (5) (2010) 64–71. [23] S.T. Lee, R.N. Swamy, S.S. Kim, Y.G. Park, Durability of mortars made with recycled fine aggregates exposed to sulfate solutions, J. Mater. Civ. Eng. 20 (1) (2008) 63–70. [24] M.C. Limbachiya, T. Leelawat, R.K. Dhir, Use of recycled concrete aggregate in high-strength concrete, Mater. Struct. 33 (2000) 574–580. [25] K. McNeil, T.H.K. Kang, Recycled concrete aggregates: a review, Int. J. Concrete Struct. Mater. 7 (2013) 61–69. [26] P.K. Mehta, Reducing the environmental impact of concrete, Concr. Int. 23 (2001) 61–66. [27] Y. Mohammadi, S.K. Kaushik, Flexural fatigue-life distributions of plain and fibrous concrete at various stress levels, J. Mater. Civ. Eng. ASCE 17 (6) (2005) 650–658. [28] J.W. Murdock, C.E. Kesler, Effect of range of stress on fatigue strength of plain concrete beams, ACI J. Proc. 55 (2) (1958) 221–231. [29] T.R. Naik, S.S. Singh, Fatigue property of concrete with and without mineral admixtures (Ph. D thesis), 1994. [30] B.H. Oh, Fatigue analyses of plain concrete in flexure, J. Struct. Eng. ASCE 112 (2) (1986) 273–288. [31] B.H. Oh, Fatigue life distribution of concrete for various stress levels, ACI Mater. J. 88 (2) (1991) 122–128. [32] S. Ozaki, N. Sugata, Fatigue of concrete composed of blast furnace slag or silica fume under submerged conditions, in: Proceedings of the Fourth International Conference, Istanbul, Turkey, vol. II, ACI Special Publication No. SP-132, 1992, pp. 1509–1524. [33] S. Pimplikar, Use of recycled aggregate in concrete, Int. J. Eng. Res. Technol. 2 (2013) 1–9. [34] M.J.R. Prince, B. Singh, Bond behaviour between recycled aggregate concrete and deformed steel bars, Mater. Struct. (2013). [35] K.D. Raithby, J.W. Galloway, Effects of moisture condition, age, and rate of loading on fatigue of plain concrete, in: Fatigue of Concrete, ACI Special Publication No. SP-41, 1974, pp. 15–34. [36] RILEM Committee 36-RDL, Long term random dynamic loading of concrete structures, Materials and Structures, Research and Testing (RILEM, Paris), vol. 17, No. 97, pp. 1–28.
S. Arora, S.P. Singh / Construction and Building Materials 102 (2016) 782–791 [37] K.K. Sagoe-Crentsil, T. Brown, A.H. Taylor, Performance of concrete made with commercially produced coarse recycled concrete aggregate, Cem. Concr. Res. 31 (2001) 701–712. [38] A. Shayan, A. Xu, Performance and properties of structural concrete made with recycled concrete aggregate, ACI Mater. J. 100 (2003) 371–380. [39] X.P. Shi, T.F. Fwa, S.A. Tan, Flexural fatigue strength of plain concrete, ACI Mater. J. 90 (5) (1993) 435–440. [40] S.P. Singh, S.K. Kaushik, Flexural fatigue life distributions and failure probability of steel fibrous concrete, ACI Mater. J. 97 (2000) 658–667. [41] S.P. Singh, S.K. Kaushik, Fatigue strength of steel fibre reinforced concrete in flexure, Cement Concr. Compos. 25 (2003) 779–786. [42] S.P. Singh, Y. Mohammadi, S.K. Kaushik, Flexural fatigue analysis of steel fibrous concrete containing mixed fibres, ACI Mater. J. 102 (2005) (2005) 438– 444. [43] P.R. Sparks, J.B. Menzies, The effect of the rate of loading upon the static and fatigue strengths of plain concrete in compression, Mag. Concr. Res. 75 (83) (1973) 73–80. [44] C. Thomas et al., Fatigue limit of recycled aggregate concrete, Constr. Build. Mater. 52 (2014) 146–154. [45] E.W. Tse, D.Y. Lee, F.W. Klaiber, Fatigue behavior of concrete containing fly ash, in: V.M. Malhotra (Ed.), Proceedings of the Second International Conference on the Use of Fly Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, Madrid, Spain, vol. 1, ACI Special Publication No. SP-91, 1986, pp. 273–289.
791
[46] P.H. Wirsching, J.T.P. Yao, Statistical methods in structural fatigue, in: Proceedings, ASCE100, No. ST6, 1970, pp. 1201–1219. [47] H.Q. Yan, Q.Y. Wang, Experimental research on fatigue behavior of recycled aggregate reinforcement concrete from earthquake-stricken area, Adv. Mater. Res. 906 (2010) 160–162. [48] H.Q. Yan, Q.Y. Wang, Y. Ning, Experimental research on fatigue behavior of recycled aggregate reinforcement concrete made from building scrap, Adv. Mater. Res. 339 (2011) 448–451. [49] J. Yang, Q. Du, Y. Bao, Concrete with recycled concrete aggregate and crushed clay bricks, Constr. Build. Mater. 25 (2011) 1935–1945. [50] M. Tavakoli, P. Soroushian, Strengths of aggregate concrete made using fielddemolished concrete as aggregate, ACI Mater. J. 93 (2) (1995) 182–190. [51] T.H.K. Kang, W. Kim, Y.K. Kwak, S.G. Hong, The choice of recycled concrete aggregates for flexural members. In: Proceedings of 18th international association for bridge and structural engineering congress on innovative infrastructures, Seoul, Korea, 2012. [52] K.H. Yang, H.S. Chung, A.F. Ashour, Influence of type and replacement level of recycled aggregates on concrete properties, ACI Mater. J. 105 (3) (2008) 289– 296. [53] J.B. Kennedy, A. Neville, Basic statistical methods for engineers and scientists, Longman Higher Education, 1986.