Fatigue capacity of plain concrete under fatigue loading with constant confined stress

H. L. Wang1 and Y. P. Song2

(1) Civil and Architectural Engineering College, Dalian University, Dalian, 116622, China (2) State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology, Dalian, Liaoning Province, 116024, China H. L. Wang Email: whl2003_2002@tom.com Received: 7 February 2009Accepted: 26 April 2010Published online: 9 May 2010

Abstract The fatigue tests of concrete have been conducted with constant confined stress, including compression, tension and alternate tensioncompression fatigue loading. Based on the results of experiment, the equations of SN curves under certain lateral stress were established and fatigue failure criterions for different loading cases were proposed. Thus according to the SN curves, the fatigue reduction coefficients for different fatigue loading were obtained. At last, the necessity of the proposed fatigue failure criterion was shown by checking an existing concrete structures computation under fatigue loading. The investigation of this paper can provide some useful information for the design of concrete structures such as reinforced concrete bridge crane beams, offshore platforms, concrete sleepers, nuclear power plants and pressure vessels. Keywords Concrete Constant confinement Biaxial Fatigue Failure criterion

1 Introduction
Many structures are often subject to repetitive cyclic loads. Examples of such cyclic loads include machine vibration, sea waves, wind action and automobile traffic. The exposure to repeated loading results in a steady decrease in the stiffness of the structure, which may eventually lead to fatigue failure. The earliest research on fatigue properties of concrete materials is traced back to the end of the nineteenth century [ 1]. In recent years, many investigations [2, 3] concerning plain concrete under uniaxial cyclic tension, uniaxial alternate tensioncompression and biaxial fatigue loads have been carried out, but the uniaxial compressive fatigue tests so far have been most investigated [4]. Fatigue behavior of plain concrete under biaxial compression [5], high-strength concrete subjected to proportional biaxial-cyclic compression [6] and steel fiber reinforced concrete subjected to biaxial compressive fatigue loading [ 7] were investigated. Hollow concrete specimens were tested to investigate the fatigue response of concrete subjected to biaxial stresses in the compressiontension region and tension tension region by Subramaniam et al. [8] and Subramaniam and Shah [9]. A review of the available literature indicates that most of the experimental results pertain to proportional fatigue loading. In practical structural applications, however, concrete structures may be subjected to nonproportional fatigue loading, where the load along the one axis are fixed with the fatigue loading imposed in the orthogonal

direction. Structures such as airport/highway pavements and bridge decks are subjected to repetitive live loads of high-stress amplitude due to passing vehicles, and the dead loads of these structures can also have some effects on the he fatigue performance of concrete. Consequently, it is necessary to research on the fatigue behavior of concrete subjected to nonproportional biaxial stresses. The limited research of plain concrete on nonproportional biaxial fatigue compressive loading was only reported by Lv [ 10]. Lv [10] founded the compressive deformation and fatigue strength of concrete under biaxial nonproportional fatigue loading is higher than that under uniaxial compression for any given number of load cycles. The conclusion was same with that of proportional biaxial-cyclic compression [57]: the fatigue failure envelopes in compressive regions have shapes similar to the envelope for static strength. But many concrete structures suffer from tension or alternate tensioncompression fatigue loading with constant lateral stress, very little information is available on the fatigue strength under these biaxial loading conditions. So study on the fatigue strength and failure modes of plain concrete under biaxial loading is necessary for evaluating the safety of concrete structures subjected to fatigue loading. Accordingly, the main purpose of this paper was to investigate the fatigue strength of plain concrete under cyclic compression, cyclic tension and alternate tensioncompression cyclic loading with constant lateral pressures, thus deduce the failure criterion of concrete under multiaxial fatigue loading.

2 Experimental investigation
2.1 Specimen and mix proportions
Plain concrete cubes with a size of 100 100 100 mm were subjected to cyclic compression loads and special dog-bone specimens with 350 mm long, 100 mm wide; 100 mm thick were subjected to tension or alternate tensioncompression fatigue loading. A detailed diagram for dog-bone specimens used in this paper was given in Fig. 1. All the dog-bone specimens were set by eight pieces of steel bar embedded in their two ends for connection between specimen and testing machine.

Fig. 1Dog-bone specimen (unit: mm)

The concrete used in the study had the following proportions, per cubic meter: cement 383 kg; fine aggregate, 663 kg; coarse aggregate, 1,154 kg; and water, 200 kg. A Chinese standard (GB175-99) [11] 42.5# Portland cement was used. Crushed stones were used as coarse aggregate with particle size ranging from 5 to 20 mm. River sand was used as the fine aggregates. The cubic specimens were cast in steel molds and the dog-bone specimens were cast in wood molds. The molds were removed after 24 h from casting and the specimens were placed in a curing room at a relative humidity of 95% and at a temperature of 20C for 4 weeks. The 28-day compressive strength, f c, and the average modulus of elasticity of concrete, E 0, obtained by testing standard prism specimens (150 mm 150 mm 300 mm) were 24.64 MPa and 24.60 GPa respectively. The average uniaxial tensile strength of the concrete, f t obtained by testing dog-bone specimens was 2.43 MPa. The age of the specimens at the time of testing was more than half a year because the difference in age of strength was not deemed to affect the test results significantly.

2.2 Test facilities

A 1,000 kN capacity Servo-controlled Material Testing System (MTS2810NEW) with a control computer was employed to apply longitudinal cyclic loading (see Fig. 2). An independent loading frame was constructed for applying lateral confinement in the horizontal direction. Lateral pressures was applied with manually operated 250 kN capacity screw jacks and calibrated by the load transducers. At the end of every jack a load cell and spherical hinge were used.

Fig. 2 Specimen and setup. a Setup of cubic specimen, b setup of dog-bone specimen

2.3 Experimental program

In the nonproportional biaxial fatigue tests, the lateral confinement stress level |l|/f c and the maximum stress level S max were two primary variables. During the test, all specimens were subjected to two loading phases: (1) Initial confining phase, where specimen was installed exactly and then loaded horizontally to a predetermined value. For all the biaxial tests in this paper, three lateral confinement stress levels were concluded, i.e., |l|/f c = 0, 0.25 and 0.50. (2) Application of longitudinal loading. The sinusoidal wave form cyclic loading with the predetermined minimum and maximum stresses levels was applied until the specimen failure occurred or the prescribed maximum number (200 million cycles) was reached. For compressive fatigue with constant tests, the maximum stress level of the vertical fatigue compression S max = max/f c varied from 0.65 to 1.05, but the

minimum stress level min/f c was always 0.1, here negative values indicate compression. For tensile fatigue with constant tests, the maximum stress level of the vertical fatigue tension S 1max = 1max/f t varied from 0.3 to 0.85, and the minimum stress level 1min/f t was always 0.1. For alternate tensile-compressive fatigue tests, the maximum stress level of the vertical fatigue tension S 1max = 1max/f t varied from 0.3 to 0.85, and the minimum stress level 1min/f c was always 0.2. To avoid the frictional confinement of the test specimens, friction-reducing materials consisting of three layers of plastic membrane with grease between each layer were placed between the specimens and the compressive loading platens in all tests. Three to five specimens were tested at each case, as seen in the scatter degree of the results. In this paper, the three principal stresses are expressed as 1 2 3 (tension denoted as positive).

3 Test results
3.1 Failure modes
Two main typical failure modes of the cube specimens are shown in Fig. 3. For specimens tested in uniaxial compressive fatigue, due to the effective measures to reduce friction, the cubes became column-type fragments, and the cracks were basically parallel to the applied load. For specimens tested in biaxial compression, it is apparent that the shear-type failure was formed by the introduction of the constant confined stress. This finding on the failure modes obtained for plain concrete under biaxial compression conditions was in agreement with a previous study [10].

Fig. 3 Failure modes of concrete specimens under compressive fatigue loading. a Uniaxial tests, b biaxial tests

As for the tensile fatigue and the alternate tensile-compressive fatigue tests, all the dogbone specimens have the same failure modes. Figure 4 shows some typical dog-bone specimens after testing. In the present study, stress concentration at the ends of the embedded bars is substantially reduced. Most of the specimens were ruptured by only one apparent tensile crack perpendicular to the direction of tensile load. Subramaniam [8] and Subramaniam et al. [9] found that the damage in concrete subjected to biaxial fatigue loading in the compressiontension region via a torsion test was localized to a single crack, and the crack growth governed the observed load-deformation responses. This finding was also in agreement with this study.

Fig. 4 Failure mode of specimens under fatigue tensile and alternate tensile compressive loading

3.2 SN diagrams
3.2.1 Biaxial compressive fatigue tests According to the above experimental program, the fatigue life of concrete under biaxial compressive cyclic loading with lateral pressure is obtained. The results are listed in Table 1. Because fatigue life is considered to follow a normal logarithmic distribution generally, the antilogarithm of the average of fatigue life logarithms is used as the mean value for the given loading case in this paper. By a regression analysis, the relationship of the average fatigue life and the maximum stress level is expressed as Eqs. 13. The correlative coefficients for them were 0.96, 0.92 and 0.88 respectively.
Table 1 Test results for compressive fatigue loading with various confined stresses

Fatigue life N f |2/f c| S

max

Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

0 (uniaxia l)

0.9 0.8 5

57 248

72 283

98 390

157 617

177 624

Fatigue life N f |2/f c| S

max

Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

0.8 0.7 5 0.7 0.6 5 1.0 5 0.9 5 0.25 0.8 5 0.7 5 0.6 5 1.0 5 0.9 5 0.5 0.8 5 0.7 5 0.6

455 12,728 63,638 2.5 1 06 () 156

574 23,795 798,362 2.5 1 06 () 224

739 47,261 826,899

1,489 49,386

2,951

578

1,563

3,078

372

2,166

31,650

70,811 293,67 9

17,624 660,55 1

1,801

1,834

98,011

654,317 2.5 1 06 () 889

695,406 2.5 1 06 () 1,969

890,318

5,387

8,529

25,397

5,623

12,108

15,390

21,923 251,18 9

36,401 325,83 6

6,823

19,409

39,810 1,565,57 7

792,239 2.5 1

893,743 2.5 1

Fatigue life N f |2/f c| S

max

Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

06 ()

06 ()

Note: The lateral confinement stress level: |2/f c|; the maximum stress level: S max = max/f c; the symbol of denotes that failure was not detected

\left| {S_{\max } } \right| = 0.9885 0.0618\,\log \, N_{\text{f}} \left( {\left| {{{\sigma_{2} } \mathord{\left / {\vphantom {{\sigma_{2} } {f_{\text{c}} }}} \right. \kern\nulldelimiterspace} {f_{\text{c}} }}} \right| = 0} \right)
(1)

\left| {S_{\max } } \right| = 1.1615 0.0762\,\log \, N_{\text{f}} \left( {\left| {{{\sigma_{2} } \mathord{\lef t/ {\vphantom {{\sigma_{2} } {f_{\text{c}} }}} \right. \kern\nulldelimiterspace} {f_{\text{c}} }}} \right| = 0.25} \right)
(2)

\left| {S_{\max } } \right| = 1.4412 0.1022\,\log \, N_{\text{f}} \left( {\left| {{{\sigma_{2} } \mathord{\lef t/ {\vphantom {{\sigma_{2} } {f_{\text{c}} }}} \right. \kern\nulldelimiterspace} {f_{\text{c}} }}} \right| = 0.50} \right)
(3) Aas-Jakobsen [12] showed that the relationship between max/f c and min/f c was linear for compressive fatigue failure, and proposed the following equation:

|Smax|=1(1R)logNf
(4) where R = min/max is the ratio of the minimum to the maximum stress; is the slope of the SN curve when R = 0. Using a large database of uniaxial fatigue test results for 475 specimens, Tepfers and Kutti [13] found that is 0.0685 and Eq. 4 is applicable to both ordinary concrete and lightweight concrete. To account for the effect of lateral confinement, a modified Aas-Jakobsen relationship between max/f c and min/f c is proposed in the present study through a linear regression analysis, i.e.

\left\{ {\begin{array}{*{20}c} {\left| {S_{\max } } \right| = \alpha \beta (1 R)\;{ \log }\;N_{\text{f}} } \hfill \\ {\alpha = 1 + 0.8304\left( {{{\left| { \sigma_{2} } \right|} \mathord{\left/ {\vphantom {{\left| {\sigma_{2} } \ right|} {f_{\text{c}} }}} \right. \kern\nulldelimiterspace} {f_{\text{c}} }}} \right)} \hfill \\ {\beta = 0.0638 + 0.115\left( {{{\left| {\sigma_{2} } \right|} \mathord{\left/ {\vphantom {{

\left| {\sigma_{2} } \right|} {f_{\text{c}} }}} \right. \kern\nulldelimiterspace} {f_{\text{c}} }}} \right)} \hfill \\ \end{array} } \rig ht.\;\;\;\;\left( {0 \le \left| {{\frac{{\sigma_{2} }}{{f_{\text{c}} }}}} \rig ht| \le 0.5} \right)
(5) where and are material constants correlated with the lateral confined stress level |2|/f c, the coefficients of correlation of which are 0.96 and 0.99, respectively. When 2/f c = 0, = 0.0638, then Eq. 5 became Aas-Jakobsen Equation which is applicable to uniaxial compressive fatigue. The SN diagrams for compressive fatigue with different confined stress levels and the ratio of stress can be determined by Eq. 5. The typical SN curves and the test results are shown in Fig. 5. In these curves the logarithm of the fatigue life is plotted versus the maximum stress level (because the minimum stress level is usually constant, it is not taken account to the SN curve generally). It can be seen that, at a given maximum stress level, the compressive fatigue life will be longer with increasing lateral pressure level.

Fig. 5 SN curves of compressive fatigue with various confined stresses

3.2.2 Tensile fatigue with constant lateral stress Results of the tension fatigue tests are shown in Table 2. These scattered results were statistically analyzed to obtain SN curves for different confined stress:
Table 2 Test results for tensile fatigue loading with various confined stresses

(|2|/ f c)

Fatigue life N f S
1max

Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

0.8

199

357

633

854

2,512

(|2|/ f c)

Fatigue life N f S
1max

Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

5 0.7 5 0.7 0.7 5 0.25 0.6 0 0.4 5 0.6 0 0.50 0.4 5 0.3 0 1,778 16,817 22,140 25,119 32,613 1,258,92 6 755

136,254

210,146

498,701

565,668

46

104

128

321

49,326 2,301,29 7 306

62,653 2,300,41 5 713

247,431

304,064

444,162

1,107

1,830 1,053,46 5 2,583,93 0

3,071 1,131,10 9

531,411 1,753,60 3

697,928 2,256,69 4

979,278 2,418,79 6

Note: The lateral confinement stress level: |2/f c |; the maximum stress level: S 1max = 1max/f t

S1max=0.0476logNf+0.9659(|2/fc|=0)
(6)

S1max=0.0596logNf+0.8855(|2/fc|=0.25)
(7)

S1max=0.0722logNf+0.8226(|2/fc|=0.5)
(8) Furthermore, the relationship of fatigue life N f, maximum stress level S 1max and confined stress level 2 was derived through a multiple linear regression analysis:

logN=12.87S1max5.69|2/fc|+13.79

(9) The SN curves denoted by Eq. 9 are plotted in Fig. 6. The experimental data with different confined stress were also showed in Fig. 6, and they indicated a good

correlation with the three best-fit SN curves. It can be seen that at a given maximum stress level, the tensile fatigue life will decrease with increasing lateral pressure level.

Fig. 6 SN curves of tensile fatigue with various confined stresses

3.2.3 Alternate tensilecompressive fatigue with constant lateral stress Results of the alternate tensioncompression fatigue tests are shown in Table 3. It was also proposed that the relationship of the average fatigue life and the maximum stress level can be represented by an expression given below

S1max=1maxft=ABlogNf
(10) where A and B are material constants correlated with the lateral confined stress level |2|/f c. They could be determined from the experimental data through a multiple linear regression analysis:

\left\{ {\begin{array}{*{20}c} {A = 0.5636{{|\sigma_{2} |} \mathord{\left/ {\vphantom {{|\sigma_{2} |} {f _{\text{c}} + 0.9818}}} \right. \kern\nulldelimiterspace} {f_{\text{c}} + 0.9818}}} \\ {B = 0.0458{{|\sigma_{2} |} \mathord{\left/ {\vphantom {{|\sigma_{2} |} {f _{\text{c}} + 0.0764}}} \right. \kern\nulldelimiterspace} {f_{\text{c}} + 0.0764}}} \\ \end{array} \;\;\;\;} \ri ght.\left( {0 \le {{|\sigma_{2} |} \mathord{\left/ {\vphantom {{|\sigma_{ 2} |} {f_{\text{c}} }}} \right. \kern\nulldelimiterspace} {f_{\text{c}} }} \le 0.5} \right)
(11) where the coefficients of correlation of A and B are 0.98 and 0.79, respectively.
Table 3 Test results for tensioncompression fatigue loading with various confined stresses

Lateral

Nf

stress ratio (|2|/f c )

1max

Sample 1

Sample 2

Sample 3

Sample 4

Sampl e5

0.8 5 0.7 5 0 (uniaxia l) 0.6 5 0.5 5 0.5 0 0.7 0 0.6 5 0.6 0 0.25 0.5 5 0.5 0 0.4 0 0.5 5 0.50 0.5 0 0.4

79

178

196

398

1,702

3,981

5,623

2,673

8,810

18,323

57,544

27,416 2,500,0 00 () 54

45,186 2,500,0 00 () 92

95,499

885,116

125

318

341

158

545

1,312

6,326

7,736

2,374

5,427

13,529

20,637

71,847 294,39 8

10,552

11,274

50,910

130,815 1,284,9 98

194,342 2,500,0 00 () 365

645,790 2,500,0 00 () 467

913,011

571

765

3,815

3,155 8,988

4,357 20,859

4,760 28,416

7,590 55,231

84,019

146,45

Lateral stress ratio (|2|/f c )

Nf S
1max

Sample 1

Sample 2

Sample 3

Sample 4

Sampl e5

5 0.4 0 0.3 5 0.3 0 44,191 490,675 610,247 1,175,4 92 682,157 1,879,5 84

8 974,15 3

275,463 2,500,0 00 ()

891,725 2,500,0 00 ()

Note: The lateral confinement stress level: |2/f c|; the maximum stress level: S 1max = 1max/f c; the symbol of denotes that failure was not detected. The minimum stress level 1min/f c was always 0.2

The SN diagrams for concrete under alternate tensilecompressive fatigue with different confined stress levels and the ratio of stress was shown in Fig. 7, and the experimental data were also showed in Fig. 7. It can be seen that at a given maximum stress level, the fatigue life of alternate tensioncompression will decrease with increasing lateral pressure level.

Fig. 7 SN curves of alternate tensioncompression fatigue with various confined stresses

3.3 Fatigue strength

Fatigue strength is commonly defined as a fraction of the static strength that can be supported repeatedly for a given number of cycles. In practice, the upper stress level at N = 2 106 cycles on the SN curve is usually taken as the fatigue strength. For analysis and comparison of the results, the ratio of fatigue strength to static strength f c or f t is defined as the fatigue strength factor. Equations 5, 9 and 10 were the failure criterions

of concrete under lateral constant compressive stress. Using these failure criterions, the fatigue strength factors can be calculated for different lateral confined stress corresponding to the fatigue life of 2 million cycles. Table 4 lists the fatigue strength factors for compression fatigue with confined stress from Eq. 5 and Refs. [10, 14] at different minimum stress levels. It can be seen that, at given minimum stress level, increasing the lateral pressure results in larger factors. The conclusion that the lateral pressure has much effect on compressive fatigue strength can be drawn. Thus, it can also be concluded that material will be unnecessarily wasted if concrete structures are designed on the basis of uniaxial compressive fatigue strength when they are in fact subjected to multiaxial compressive fatigue loading. It is found that the fatigue strength of concrete under biaxial compression loading obtained in this study is somewhat lower than that obtained by Tan [14], but they have the same trend, and this study tends to be more safe in engineering.
Table 4 Compressive fatigue strength factors of plain concrete with various confined stresses

|2|/f c 0 0.1 0.2 0.3 0.4 0.5

Eq. 5, R = 0.1 Ref. [14], S min = 0.1 Ref. [10], R = 0.1

0.64 0.65 0.55

0.66 0.70 0.66

0.68 0.75 0.78

0.69 0.80 0.89

0.71 0.85 1.01

0.73 0.90 1.13

As for tension and alternate tensioncompression fatigue with confined compressive stress, according to the S/N Eqs. 9 and 10 model, the multiaxial fatigue strength can be calculated easily. Table 5 lists the fatigue strength factors for tension and alternate tensioncompression fatigue with confined stress at different minimum stress levels under 2 106 cycles and 1 106 cycles.
Table 5 Tensile and tensilecompressive fatigue strength factors of plain concrete with various confined stresses

Fatigue loading form

|2|/f
c

f1

f2

Minimum stress level

Uniaxial tension compression Tensioncompression with confined stress

0 0.25 0.50 0 0.25

0.5081 0.4820 0.3637 0.665 0.5279

0.4835 0.4657 0.3460 0.633 S min = 0.10 0.5100 S min = 0.20

Uniaxial tension Pure tension with confined

Fatigue loading form

|2|/f
c

f1

f2

Minimum stress level

stress
1

0.50

0.3894
6

0.3677
2

Note: f is the fatigue strength factor under 2 10 cycles and f is the fatigue strength 6 factor under 1 10 cycles

It can be seen that the fatigue strength of concrete under pure tension with the same confined stress is higher than that of alternate tensioncompression fatigue for any given number of load cycles. A higher lateral stress level result in a less fatigue strength factor for these two fatigue loading cases. Thus, it will be dangerous if the concrete structures at given lateral pressure are designed on the basis of uniaxial tensile fatigue strength when they are in fact subjected to multiaxial tension or alternate tension compression fatigue loading.

4 Application of failure criterion

Fatigue strength is very important for fatigue design of concrete structures in engineering practice. For example, many concrete ocean platforms are subjected to tensioncompression cyclic loading with lateral pressure caused by water pressure and other loadings. For lack of sufficient experimental data, the design value of concrete tensile fatigue strength was only adapted by uniaxial tension and the fatigue strength factor was 0.74 in the previous study [15]. Figure 8 shows one leg of the ocean platform. The fatigue strength of concrete at section I-I was recalculated according to the conclusions of this paper as followings.

Fig. 8 Schematic drawing of the leg of ocean platform

In this figure, r 1 = 3,000 mm, r 2 = 2,650 mm, r g = 2,850 mm, area of reinforcing steel bar A y = 93,200 mm2, strength of reinforcing steel bar R y = 750 MPa, modulus of reinforcing steel bar E y = 1.8 105 MPa, compressive strength of concrete R 4 N a = 23 MPa, tensile strength of concrete R f = 2.55 MPa, N = 1,014 10 (compressive force), modulus of concrete E h = 3.3 104 MPa, fatigue modulus of concrete E h p = 1.5 104 MPa, M = 3,937 107 N mm. Prestressed reinforcement is placed along the circle equably, the prestressed force is y = 133.25 MPa, the concrete prestressing stress is h = 2 MPa.

Checking the computation of fatigue strength of concrete, the transformation area is,

A0=(r21r22)+EyEhAy=67117.14102mm2
Consider the prestressed reinforcement as a steel circle, where the

Outsideradius,rg1Insideradius,rg2=Ay2rg12+rg=2827.6mm\hfill=2822.4 mm\hfill
The section modulus:

W0=4(r41r42)1r1+4(r4g1r4g2)1r1\hfill=841348.24104mm2\hfill
The maximum compressive stress caused by fatigue loading is,

=NA0+MW0=6.19MPa(herecompressivestressispositive)
and considering the prestressing stress,

pmax=+h=8.19MPa
The minimum compressive stress caused by fatigue loading is,

pmin=h+NA0MW0=1.17MPa(heretensilestressisnegative) =pminpmax=0.14
Based on Eq. 6, we consider the lateral compressive stress |2/f c| = 0.25, thus we have the fatigue strength factor from Table 5, r p = 0.4657, where

Rp=rpR=0.465723=10.711MPa Rpf=rpRf=0.46572.55=1.19MPa,
and max p = 8.19 < R p = 10.711, | min p | = 1.17 < R f p = 1.19. When the lateral stress is less than or equal to 0.25f c, the fatigue strength of the concrete meets demand, and the concrete does not craze. But obviously, when the lateral stress is 0.5f c, for example, crack will emerge in the concrete. It should be pointed out here that the level of constant confined stress is not calculated exactly.

5 Conclusions
Based on the results presented in this paper, the following conclusions are drawn: 1.The failure mode of concrete under biaxial static stress and the failure mode under biaxial cyclic loading with constant lateral pressure are similar. 2.The present investigation indicates that the biaxial compressive fatigue strength with confined stress is higher than the uniaxial value and a higher lateral pressure results in a larger value. 3.The fatigue life and fatigue strength under tension cyclic loading with lateral pressure are both lower than those under uniaxial tensioncompression cyclic loading and the values decrease with increasing lateral pressure level. 4.The fatigue life and fatigue strength under biaxial alternate tension compression cyclic loading with lateral pressure are both lower than those under uniaxial alternate tensioncompression cyclic loading and the values decrease with increasing lateral pressure level. It is noted that the fatigue strength of alternate tension compression loading is lower than that of tension loading at the same lateral stress level. 5SN diagrams are derived for plain concrete subjected to fatigue loading with constant confine stress. These expressions are based on much experimental data. Whats more, these failure criterions considering the effect of constant confined stress; it can make the checking of concrete structure more accurate.

6.The fatigue strength factors derived from the failure criterion of concrete can provide information for the fatigue design of concrete members. At last, the computation of the fatigue strength of concrete legs of ocean platform was checked to prove the necessary of this research. Acknowledgement
This work was performed under grants from the National Science Foundation of China (grants No. 50908026) and from China Postdoctoral Foundation (grant No. 20080441117, 200902541).

References 1.Joly D (1898) La resistance et l, elasticite des ciments Portland. Ann Pons Chaussees 16(7):198244 2.Kim J-K, Kim Y-Y (1996) Experimental study of the fatigue behavior of high strength concrete. Cem Concr Res 26(10):15131523CrossRef 3.RILEM Committee 36-RDL (1984) Long term random dynamic loading of concrete structures. RILEM Mater Struct 17(97):128 4.Hsu TTC (1981) Fatigue of plain concrete. ACI J 78(27):292305 5.Su ECM, Hsu TTC (1988) Biaxial compression fatigue and discontinuity of concrete. ACI Mater J 85(3):178188 6.Nelson EL, Carrasquillo RL, Fowler DW (1988) Behavior and failure of high-strength concrete subjected to biaxial-cyclic compression loading. ACI Mater J 85(4):248253 7.Yin WS, Hsu TC (1995) Fatigue behavior of steel fiber reinforced concrete in uniaxial and biaxial compression. ACI Mater J 92(1):7181 8.Subramaniam KV (1999) Fatigue of concrete subjected to biaxial loading in the tension region. PhD dissertation, Northwestern University, Evanston, IL, pp 125 9.Subramaniam KV, Popovics JS, Shah SP (2002) Fatigue fracture of concrete subjected to biaxial stress in the tensile C-T region. J Eng Mech ASCE 128(6):668 676CrossRef 10.Lv PY, Li QB, Song YP (2007) Behavior of concrete under nonproportional biaxial fatigue stresses with one constant. ACI Mater J 104(1):312 11.GB175-99 (1999) Portland cement and ordinary Portland cement. National Standard of the Peoples Republic of China (in Chinese) 12.Aas-Jakobsen K (1970) Fatigue of concrete beams and columns. Bulletin no 70-1. Division of Concrete Structures, The Norwegian Institute of Technology, The University of Trondheim, pp 142148 13.Tepfers R, Kutti T (1979) Fatigue strength of plain and ordinary and lightweight concrete. ACI J 76(5):635652

14.Hooi TT (2000) Effects of passive confinement on fatigue properties of concrete. Mag Concr Res 52(1):715CrossRef 15.Lv PY (2000) Experimental study on dynamic strength and deformation of concrete under uniaxial and biaxial action. Ph.D. Dissertation, Dalian University of Technology (in Chinese)7.637.143 scientific documents at your fingertips