Cyclic Behavior of Reinforced Concrete Beams with

3 downloads 0 Views 4MB Size Report
corroded beams by using the ACI 318 (2008) shear strength equation. ...... 340 residual shear strength of concrete and the hoops were sufficient to develop ...
Manuscript Click here to download Manuscript: JSCE_ASCE_final_2013_8_24.doc

Cyclic Behavior of Reinforced Concrete Beams with Corroded Transverse

1

Steel Reinforcement

2 3

Yu-Chen Ou 1, and Hou-Heng Chen

4

2

5 6

Abstract

7

This study examined the seismic performance of reinforced concrete beams with corrosion only

8

induced in the transverse steel reinforcement by using cyclic loading. The beams were designed

9

with closely spaced steel hoops as transverse reinforcement conforming to ACI 318 seismic

10

design provisions. Seven beams were constructed. One beam was used as a control without

11

corrosion. The other six beams were subjected to six levels of corrosion in the potential plastic

12

hinge region by using an electrochemical method. Corrosion test results indicate that pitting

13

corrosion increased with an increasing corrosion level. The hoops fractured at a corrosion weight

14

loss of 35%. Cyclic test results indicated that the beams could sustain a corrosion weight loss of

15

6% in the hoops and still maintain a ductile flexural behavior. Corrosion of hoops adversely

16

affected the deformation capacity of the beams significantly, yet did not significantly influence

17

the load-carrying capacity of the beams. The residual shear strength provided by concrete and

18

steel was still sufficient to develop flexural yielding, even for the beams with hoops corroded to

19

fracturing. Methods were developed to estimate the residual shear strength and ductility of

20

reinforced concrete beams with corroded hoops. The amount of corrosion substances filled in the

21

cracks volume was approximately 25% of the reduced volume of the steel due to corrosion.

22

Although shear strength estimation based on the average corrosion weight loss was not

23

conservative, that based on the minimum residual cross sectional area was too conservative at a

24

high corrosion level due to severe pitting corrosion. Shear strength estimation based on average

25

weight loss and minimum residual cross sectional area produced results that describe the

26

experimental behavior more reasonably.

27

Keywords: Reinforced concrete beams; corrosion; transverse reinforcement; seismic;

28

cyclic loading; flexural-shear failure. 1

2

Associate Professor, Department of Construction Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan. E-mail: [email protected] Formerly Master student, Department of Construction Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan.

29 30

INTRODUCTION

31 32

Corrosion of steel reinforcement in reinforced concrete structures is a common deterioration

33

issue. Corrosion reduces the area of steel reinforcement (i.e. uniform corrosion) and causes a

34

non-uniform distribution of cross sectional area along the reinforcement (i.e. pitting corrosion).

35

Uniform corrosion reduces the load-carrying capacity of the reinforcement, while pitting

36

corrosion causes localized yielding, reducing the strength and ductility of the reinforcement. Steel

37

expands after corrosion, resulting in cracking, delamination, and spalling of cover concrete,

38

ultimately degrading steel-concrete bond and load-carrying capacity of the structure.

39

Earlier studies have demonstrated that transverse reinforcement, due to a smaller diameter,

40

corrodes faster than the longitudinal reinforcement in terms of volume reduction. Consequently, a

41

beam designed to fail in flexure gradually shifts the failure mode from flexural to flexural-shear

42

or shear failure with an increasing corrosion level, transforming a ductile failure to a brittle one

43

(Rodriguez et al. 1997, and Ou et al. 2012). Based on reliability analysis, Val (2007) investigated

44

how uniform and pitting corrosion affects failure probability of beams designed to fail in flexure.

45

According to their results, the probability of shear failure increases faster than that of flexural

46

failure, especially when pitting corrosion occurs. The probability of shear failure may exceed that

47

of flexural failure within a 50 year life span of a typical structure.

48

Corrosion of transverse reinforcement generally involves uniform corrosion along the length

49

of steel bars. At moderate and severe corrosion levels, pitting corrosion becomes significant and

50

typically occurs near the corners of the bents (Higgins and Farrow 2006, Juarez et al. 2011, and

51

Wang et al. 2011). Rodriguez et al. (1997) indicated that shear strength estimated using the

52

Eurocode (1992) shear strength equation with a reduced section of the transverse reinforcement at

53

the corrosion pit and original concrete cross section overestimates the experimental results. A

54

conservative estimation can be obtained if cover concrete is removed completely when

55

calculating the shear strength. Juarez et al. (2011) estimated the residual shear strength of

56

corroded beams by using the ACI 318 (2008) shear strength equation. According to their results,

57

shear strength calculated based on the average residual diameter of the transverse reinforcement

58

overestimates the measured shear strength; meanwhile, that based on the minimum residual

59

diameter can provide conservative estimation of shear strength. The shear strength of concrete is 1

60

assumed in Juarez’s study to remain constant after corrosion. Xia et al. (2011) proposed

61

empirical equations to predict the ratio of the shear strength of a corroded beam to that of the

62

corresponding uncorroded beam based on the average cross-sectional loss of corroded transverse

63

reinforcement, average corrosion crack width, or maximum corrosion crack width. Note that the

64

above studies on shear behavior used monotonic loading to test specimens.

65

Under a seismic condition, the plastic hinge regions of a beam are expected to undergo

66

inelastic deformation reversals. As is well known, shear strength of concrete degrades under

67

inelastic reversals with an increasing ductility (Aschheim and Moehle, 1992). Therefore, the

68

ACI-318 code requires more transverse reinforcement for shear reinforcement in the seismic

69

design of beam plastic hinge regions. Moreover, transverse reinforcement plays an important role

70

in providing confinement to concrete and delaying buckling of longitudinal reinforcement,

71

ensuring a satisfactory ductility capacity of the plastic hinge region under seismic loading

72

reversal. Consequently, corrosion of transverse reinforcement is more detrimental to the

73

structural performance under seismic condition than under monotonic loading condition.

74

Figure 1 illustrates examples of transverse reinforcement corrosion in reinforced concrete

75

buildings. Exactly how transverse reinforcement corrosion affects the inelastic shear behavior of

76

beams has received limited attention in the literature. Kato (2006) and Ou et al. (2012)

77

investigated the cyclic behavior of beams with corroded transverse reinforcement, indicating that

78

corrosion of transverse reinforcement influences the beam ductility. However, most of the

79

specimens investigated incurred corrosion both in the longitudinal and transvers reinforcement,

80

making it difficult to differentiate between the effects of transverse and longitudinal

81

reinforcement corrosions on the structural performance. This study conducted cyclic tests on

82

beams with corrosion induced only in the transverse reinforcement. Six corrosion levels were

83

examined. Results of this study shed further light on the seismic behavior of beams with corroded

84

transverse reinforcement and develop residual shear strength and ductility capacity models.

85 86

EXPERIMENTAL PROGRAM

87 88 89 90

Specimen Design The specimen was designed in compliance with the ACI 318 code (2008). Figure 2(a) 2

91

illustrates the side view of the specimen. The specimen contains a beam connected to an

92

anchorage block. The end of the beam connected to the anchorage block is referred to as fixed

93

end while the other end referred to as free end. Figure 2(b) illustrates the cross-sectional view of

94

the beam. The specified and actual concrete compressive strengths f c' are 28 and 38 MPa,

95

respectively. The beam was designed with three #9 bars in the top and bottom sides of the beam

96

and with #4 hoops as transverse reinforcement having a horizontal spacing of 10 cm. The

97

specified and actual: yield strengths of longitudinal reinforcement f y are 412 and 444 MPa,

98

respectively; in addition, those for transverse reinforcement are 412 and 432 MPa, respectively.

99

The beam has a tension longitudinal reinforcement ratio of 1.5% and a volumetric ratio of

100

transverse reinforcement to the concrete core of 1.8%. Shear strength of the beam was designed

101

based on shear demand, which corresponds to moment strength assuming 1.25 specified yield

102

strength in the longitudinal reinforcement. Concrete shear strength was not considered in the

103

shear strength calculation based on the seismic provisions of the ACI 318 code. Figure 2(c)

104

illustrates the corrosion observation specimen. The specimen is a replication of the portion of the

105

beam from the fixed end extending 60 cm towards the free end. Figure 2(d) shows the

106

construction of the beam and corrosion observation specimens. There were a total of seven sets of

107

specimens. One set was used as the control group without corrosion, while the other six sets were

108

subjected to different corrosion levels.

109 110

Test Setup

111

Accelerated Corrosion

112

Corrosion in the transverse reinforcement was accelerated using an electrochemical method

113

by imposing an electrical current to the reinforcement. The current density was approximately

114

600 µ A/cm 2 . Corrosion was controlled to occur only in the potential plastic hinge region of the

115

beam, from the fixed end extending 60 cm to the free end of the beam. Six hoops from the fixed

116

end of the beam were imposed with an electrical current by connecting the ends of the hoops to

117

the anodes of DC power supplies. The 60-cm region was enclosed with a water tank infilled with

118

a NaCl solution of 5%. Four copper plates were placed on the four sides of the beam in the water

119

tank. The cathodes of the power supplies were connected to the copper plates. Figures 3(a) and

120

3(b) illustrate the elevation and cross-sectional views of the accelerated corrosion setup,

121

respectively. Figure 3(c) illustrates the electric wires connecting the hoops to the power supplies. 3

122

Figure 3(d) illustrates the overall view of the test setup. The beam specimen and the corrosion

123

observation specimen in each set were subjected to the corrosion process starting from the same

124

time and lasted for the same period.

125 126

Cyclic Testing

127

Following the corrosion, the corrosion observation specimen was demolished and the

128

reinforcement was removed for corrosion measurement. The beam specimen was subjected to

129

cyclic loading to evaluate the seismic performance. Figure 4 shows the test setup for cyclic

130

loading. The anchorage block of the specimen was fixed to the strong floor. Cyclic load was

131

applied using a hydraulic actuator attached to the beam at 120 cm from the fixed end. The beam

132

was cyclically loaded up to drift levels of 0.25%, 0.375%, 0.5%, 0.75%, 1.0%, 1.5%, 2.0%, 3.0%,

133

4.0%, 5%, and 6%. Each drift level was repeated twice. The actuator was controlled by using the

134

relative displacement measured during testing between the loading point and the fixed end of the

135

beam to achieve the prescribed drift levels precisely.

136 137

RESULTS AND DISCUSSION

138 139 140

Corrosion Test Results

141

Table 1 lists, for all of the specimens, the corrosion weight loss, ∆w , average and minimum

142

residual cross-sectional areas, Aavg and Amin , maximum pit depth, pmax , and total crack width,

143

Wcr . The values of ∆w , Aavg , Amin , and pmax represent the average of the values from the six

144

hoops subjected to corrosion. The average residual cross-sectional area was calculated based on

145

the residual weight. The minimum residual cross-sectional area was calculated based on the

146

average of three diameters measured at the smallest cross section. The corrosion pit was

147

measured by a caliper with respect to a nearby point with negligible corrosion. Note that

148

corrosion measurements on steel reinforcement were obtained from the corrosion observation

149

specimens. To calculate the total crack width, crack patterns were first recorded with the width of

150

each crack measured for both the beam specimens and corrosion observation specimens. Then,

151

the total crack width was determined by summing all of the crack widths along the perimeter of a

152

selected cross section. The 60-cm corroded portions of the beams and corrosion observation 4

153

specimens were divided equally into five regions, and the total crack width was determined for

154

each region. Table 1 presents the average value of the total crack widths from the five regions,

155

indicating that the beam specimen and the corresponding corrosion observation specimen have

156

similar corrosion crack widths for medium to high corrosion levels.

157

Table 1 indicates that an increasing corrosion levels increases the difference between the

158

average and minimum residual cross-sectional area. This finding suggests that pitting corrosion

159

becomes more significant with an increasing corrosion level. This phenomenon is owing to that

160

severe pitting corrosion typically occurred near corrosion cracks. Corrosion cracks facilitated the

161

reaction of ferrous iron with hydroxide to form corrosion substances. Due to the volumetric

162

expansion, the corrosion substances subsequently widened the cracks, exacerbating the corrosion

163

further. For the specimen with 35% corrosion, every hoop was corroded in the accelerated

164

corrosion process until it broke. This was achieved by continuing the accelerated corrosion

165

process for each hoop until a sudden increase of electrical resistance, which is a sign of the hoop

166

breakage, as confirmed upon the retrieval of the corroded hoops (Fig. 5). The electric-chemical

167

process was then terminated by turning off the electrical current to the broken hoop. Note that

168

severe pitting corrosion is evident in Fig. 5. Figure 6 illustrates the distribution of cracks on the

169

four faces of the 60-cm region of the beam specimen subjected to accelerated corrosion. Note that

170

the direction of most of the cracks is along the longitudinal direction of the beam, rather than

171

along the transverse direction. This is likely because the longitudinal reinforcement, which did

172

not corrode, restrained the development of cracks along the transverse direction of the beam.

173 174

Cyclic Test Results

175

Figure 7 illustrates the damage distribution of the specimens at selected drift levels. Figure 8

176

illustrates the hysteretic behavior of all specimens with the instances of spalling of cover concrete

177

and fracture of hoops indicated. Table 2 lists various performance indicators for the specimens.

178

The symbols used in Table 2 are defined as follows: ∆ y denotes idealized yield drift; ∆ u

179

denotes ultimate drift, defined when the applied load declines more than 20% from the peak

180

value; Ppeak denotes peak applied load; µ denotes ductility, defined by the ratio of ∆ u to

181

∆ y ; and ∆ p denotes plastic rotation, defined as the difference between ∆ u and ∆ y (FEMA

182

267). The idealized yield drift refers to the drift at the intersection of two lines in the hysteretic

5

183

behavior plot (Fig. 8): one line starts from the origin and intersects the envelope of the hysteretic

184

behavior at approximately 60% of the force at the idealized yield drift; the other line ends at the

185

ultimate drift with zero stiffness. The intersection of the two lines was chosen so that the area

186

beneath the envelop response equals that beneath the two lines. The values listed in Table 2 are

187

the average value of those from positive and negative drift responses.

188

The specimen without corrosion (Bt-0) exhibited a typical flexural-dominant behavior. At the

189

first cycle of the 5% drift (Fig. 7(a)), flexural-shear cracks were well distributed within the region

190

approximately 70 cm from the fixed end. Significant spalling of concrete started to occur. During

191

the second cycle of the 5% drfit loading, the core concrete near the fixed end of the beam started

192

to spall off. Consequently, the beam lost its load-carying capacity, showing a significant

193

distortion in the 40-cm region of the beam from the fixed end. The applied load dropped

194

significantly. This specimen failed in a flexural failure mode. For specimen Bt-3, during the

195

second cycle of 4% drfit loading, significant spalling of cover concrete occurred. At 5% drift,

196

most of the cover concrete of the 40-cm region spalled off (Fig. 7(b)). A significant portion of the

197

core concrete also spalled off. The beam failed with a significant distortion in the 40-cm region

198

from the fixed end. The failure mode is a flexural type same as specimen Bt-0. Specimen Bt-6

199

exhibited a flexural failure pattern similar to the previous two specimens, except for that the hoop

200

closest the fixed end of the beam fractured during the second cycle of the 4% drift loading.

201

Figures 7(a), 7(b) and 7(c) compare specimens Bt-0, Bt-3 and Bt-6 at the first cycle of 5% dirft,

202

respectively, revealing that an increasing corrosion level increased the extent of spalling of cover

203

concrete. Additionally, the peak applied load and ultimate drift decreased with an increasing

204

corrosion level (Table 2), due to the reduced load-carrying capacity of cover concrete and

205

confinement capacity of the corroded hoops. However, the corroded beams (Bt-3 and Bt-6) still

206

possess a drift capacity higher than 4% and a plastic rotation capacity exceeding 3% with a fairly

207

ductile behavior (Figs. 8(b) and 8(c)) similar to the control specimen. This finding suggests that a

208

beam designed based on the current design code can sustain a corrosion weight loss of 6% in the

209

hoops without significantly impairing the seismic capacity. Note that this finding is based on a

210

beam with a moderate shear stress level of 0.46

211

approximately 55% the maximum average shear stress allowed by the ACI 318 Code (2008) and

212

an equal amount of tension and compression longitudinal reinforcement.

213

f c' (MPa) at the peak applied load, i.e.

Specimens Bt-11, Bt-12 and Bt-36 showed a flexural-shear failure mode. The failure of these 6

214

specimens was preceded by a major diagonal shear crack, as shown in Figs. 7(d), 7(e) and 7(f),

215

respectively. Subsequently, a large portion of cover concrete spalled and then fracture of transvers

216

reinforcement occurred, as shown in Figs. 8(d), 8(e) and 8(f), respectively. Note that negative

217

drift loading was applied before positive drift loading in each cycle of loading. With a corrosion

218

weight loss between 11% and 16%, the corroded hoops appeared to be unable to restrain the

219

diagonal shear cracking. The beams failed immediately after the formation of a major diagonal

220

crack. The peak applied loads of these specimens resembled that of specimen Bt-6 (Table 2).

221

However, the drift capacities were reduced significantly with ultimate drifts ranging from 2% to

222

3% and plastic rotations from 1.7% to 2.2%. The higher the corrosion level, the lower the drift

223

capacity. These three beams exhibited a hysteretic behavior substantially less ductile than the

224

previous three specimens (Bt-0 to Bt-6). The hoops of specimen Bt-35 were corroded to fracture

225

before cyclic testing. Several major diagonal cracks formed at the first negative 2% drift loading

226

(Fig. 7g). When the load was reversed to a positive 2% drift, a large portion of cover concrete

227

spalled, exposing several fractured hoops and ultimately leading to failure of the beam (Fig.

228

8(g)).

229

All specimens developed yielding of longitudinal reinforcement, as evidenced by the clear

230

softening of hysteretic behavior (Fig. 8h) before failure. Even for specimen Bt-35, which had

231

hoops already broken before cyclic testing, the specimen still appeared to possess a significant

232

amount of shear strength. This amount is largely owing to the shear strength contribution from

233

concrete and the fact that critical shear cracks may cut the hoops at locations a distance from the

234

broken locations, in which a certain amount of shear strength of the hoops remained. For instance,

235

according to Fig 5(f), one leg of the hoop was broken, while a significant cross-sectional area

236

remained for the other leg. It is likely that before the spalling of cover concrete, the unbroken leg

237

could still provide a significant shear resistance. When the cover concrete started to spall, the

238

shear strength dropped drastically, as mentioned earlier. Since the longitudinal reinforcement of

239

all of the specimens yielded before failure, the peak applied load did not decrease significantly

240

with an increasing corrosion level (Fig. 8h). The reduction of the peak applied load in specimen

241

Bt-35 is only 13% compared to the control specimen. However, the ultimate drift, ductility and

242

plastic rotation capacity significantly decreased with an increasing corrosion level. This

243

phenomenon is because corrosion decreased the ability of the hoops to confine the core concrete

244

(specimens Bt-3 and Bt-6) and restrain shear cracks (specimens Bt-11, Bt-12, Bt-16, and Bt-35), 7

245

thus accelerating failure of the beams and ultimately reducing deformation capacity.

246 247

Residual Shear Strength

248

The shear strength of beam Vn consists of concrete contribution Vc and shear

249

reinforcement contribution Vs (Eq. 1). Vc decreases after corrosion due to the softening of

250

concrete caused by corrosion cracks. Vs decreases due to the reduction of cross sectional area.

251

The effect of bond reduction on Vs is not considered since hoops rely mainly on anchorages to

252

develop stress and anchorages (hooks) of the hoops are better protected by a thicker cover than

253

the other region of the hoops (Fig. 5).

Vn = Vc + Vs

(1)

254

where Vn is shear strength; Vc is shear strength provided by concrete; and Vs is shear strength

255

provided by shear reinforcement

256

The softening effect of concrete due to corrosion cracks is assumed to occur only in the cover.

257

The expansion of corrosion substances exerts an outward pressure to the cover concrete, leading

258

to splitting cracks in the cover concrete. These cracks decrease the ability of cover concrete to

259

withstand the compression loading. This effect is simulated in this study based on the softening

260

theory proposed by Hsu (1992), as shown in Eq. (2). According to the softening theory, the

261

concrete compressive strength f c' decreases to ζ f c' after corrosion.

ζ =

εr =

0.9 1 + 600ε r

Wcr = pcp

∑W

cri

pcp

(2)

(3)

262

where ε r is tensile strain; Wcr is total crack width; Wcri is width of crack i as illustrated in Fig.

263

9(a); and pcp is outside perimeter of beam cross section.

264

As illustrated in Fig. 9(a) by the shaded regions, a portion of the corrosion substances fills in

265

the space surrounding the bars released by corrosion volume reduction of the hoop; in addition, a

266

portion of the corrosion substances fills in the corrosion cracks. Eq. (4) calculates the ratio of the

267

amount of corrosion substance filled in the cracks to the volumetric loss of a hoop due to

268

corrosion, α . α was 0.20, 0.24, 0.26, 0.30, 0.27, and 0.25 for specimens Bt-3, Bt-6, Bt-11, 8

269

Bt-12, Bt-16, and Bt-35, respectively, using the measured values of Wcr as listed in Table 1. The

270

value of α remains relatively constant with an average value of 0.25. Figure 9(b) reveals

271

another portion of the corrosion substances escaping from the concrete. Wcr = α

∆w γ steel sCc

(4)

272

where s is horizontal spacing the hoops; Cc is clear cover of the hoops; and γ steel is specific

273

weight of steel.

274 275

Vc is estimated using the ACI 318 code shear strength equation (ACI 2008) provided by concrete, as defined in the following equation.

 Vd Vc =  0.16 f c' + 17 ρ w u  bw d (MPa) Mu  

(4)

276

where f c' is concrete compressive strength; ρ w is tension reinforcement ratio; d is beam

277

effective depth; Vu is factored shear; M u is factored moment; and bw is beam web width. In

278

this study M u Vu is assumed to be the shear span, 120 cm. For corroded specimens, f c' in Eq.

279

(4) is replaced with ζ f c' for cover concrete area. For the core concrete area, f c' remains the

280

same. Vc degrades with an increasing ductility of the specimen during the cyclic loading. The

281

degradation parameter k , as defined by Eq. (5) and proposed by Aschheim and Moehle (1992)

282

is used to estimate the degraded concrete shear strength, kVc . 1≥ k =

4−µ ≥0 3

(5)

283

where µ is the ductility. Table 3 lists the residual values of Vc of all specimens. Corrosion

284

generally does not significantly decrease Vc . The maximum loss (specimen Bt-35) is

285

approximately 7.5%. However, ductility significantly reduces Vc . Since specimens Bt-0 to Bt-11

286

have a ductility capacity larger than 4, Vc of these specimens drops to zero before failure.

287

Vs is estimated by the following equation. Vs =

Av f y d s

9

(6)

288

where Av is cross-sectional area of shear reinforcement; and f y is yield stress of shear

289

reinforcement. Since a hoop contains two legs, Av represents the total cross sectional area of the

290

two legs. This study considers three cases of Av . In the first case, average residual

291

cross-sectional area Aavg is assumed for two legs. In the second case, the minimum residual

292

cross-sectional area Amin is assumed for two legs to consider an extreme corrosion condition. In

293

the third case, one leg of the hoop is assumed to have Aavg while the other leg has Amin . This

294

third case reflects the fact that the critical shear crack unlikely cuts both of the two legs at or near

295

the minimum residual cross section (Fig. 5). The values of Vs , as estimated based on the first,

296

second, and third cases, are referred to as Vs _ avg , Vs _ min , and Vs _ min_ avg , respectively. Table 4

297

lists the value of Vs calculated based on the three cases.

298

The shear strength of beam Vn is obtained by summing up Vc and Vs . Figure 10(a)

299

summarizes those results. This figure also reveals the experimental peak applied loads. According

300

to this figure, shear strengths calculated based on Vs _ avg always exceed the experimental peak

301

loads, even when considering reduction due to ductility, implying a flexural failure mode of all

302

specimens. This does not correspond to the shear failure mode of specimens Bt-11, to Bt-35. The

303

values of Vn based on Vs _ min or Vs _ min_ avg and reduced by duality are lower than the test

304

results for specimens Bt-11 to Bt-35. This finding suggests that these two shear calculation

305

methods can provide a conservative estimation of shear strength. Vn based on Vs _ min may be

306

too conservative, especially for high corrosion levels such as specimen Bt-35. This specimen had

307

a severe pitting corrosion and had a large difference between the minimum cross sectional area

308

and the cross sectional areas of the other portions of the hoop, which possibly provided a certain

309

amount of shear resistance, as mentioned earlier. Vn based on Vs _ min_ avg is closer to

310

experimental values both under low and high corrosion levels.

311

Figure 10(b) illustrates the relationship between parameter (Vc + Vs − Vy ) (Vinc + Vins ) and

312

the ductility, µ , of the specimens. Vinc and Vins are initial uncorroded shear strengths by

313

concrete and by shear reinforcement, respectively. The ductility capacity decreases with a

314

decreasing residual shear strength. The figure also shows the regression results, which are

10

315

expressed by Eq. (7). These regression models can be used for ductility evaluation of beams with

316

corroded transverse reinforcement.

µ=β

Vc + Vs − Vy Vinc + Vins



(7)

317

where β is 15.92, 9.18, and 6.34 and γ is -1.30, 2.16, and 3.58 for Vn based on Vs _ avg ,

318

Vs _ min_ avg , and Vs _ min , respectively; Vy is shear corresponding to the first yielding of

319

longitudinal reinforcement of the beam. R 2 for the three regressions are 0.89, 0.79, and 0.74,

320

respectively.

321 322 323

CONCLUSIONS

324 325

The effects of transverse steel reinforcement corrosion on the seismic behavior of reinforced

326

concrete beams designed conforming to the ACI 318 seismic design provisions, and with a

327

moderate shear stress level of 0.46

328

longitudinal reinforcement were examined using cyclic loading. Important conclusions are

329

summarized as follows.

330

(1) As the increase of the corrosion level, the difference increased between the average

331

cross-sectional area of the corroded hoop and the minimum residual cross-sectional area. This

332

suggests that pitting corrosion increased as the corrosion level increased. The hoops fractured

333

at a corrosion weight loss of approximately 35%.

f c' (MPa) and equal amount of tension and compression

334

(2) The specimens was able to sustain approximately a 6% corrosion level while still maintaining

335

a satisfactory ductile flexural behavior with plastic rotation capacities larger than 3%. Further

336

increase of corrosion level changed the failure mode of the beam from flexural failure due to

337

crushing of core concrete to flexural-shear failure due to diagonal tension failure. As the

338

corrosion level increased, the deformation capacities reduced significantly including the

339

ultimate drift, ductility, and plastic rotation capacities. However, the peak applied load did not

340

show a significant reduction even for the specimens with hoops broken by corrosion. The

341

residual shear strength of concrete and the hoops were sufficient to develop flexural yielding.

342

For the specimen with broken hoops, the remaining portions of the hoops still preserved a

343

significant cross-sectional area due to severe pitting corrosion. It appears that the hoops were 11

344

still able provide a certain amount of shear resistance before spalling of cover concrete.

345

(3) Methods to estimate the residual shear strength and residual ductility of the corroded beams

346

are proposed. In the estimation of the residual shear strength of concrete, the core concrete is

347

assumed not affected by corrosion. The shear strength of cover concrete decreases due to the

348

splitting cracks by corrosion. This is modeled using softening of concrete in compression due

349

to transverse tensile strain. Experimental results showed that amount of corrosion substances

350

filled in the cracks volume was approximately 25% of the reduced volume of the steel due to

351

corrosion. In the estimation of the residual shear strength by steel reinforcement, three

352

residual steel cross sectional areas are used. Comparison with the experimental results show

353

that residual shear strength based on the average weight loss is not conservative. On the other

354

hand, the residual shear strength estimated based on minimum residual cross-sectional area

355

provides conservative results. However, the estimated shear strength becomes too

356

conservative at severe corrosion levels due to significant pitting corrosion. Shear strength

357

estimated based on both average weight loss and minimum residual cross-sectional area yield

358

results that reasonably reflect experimental behavior both under low and high corrosion

359

levels.

360 361

ACKNOWLEDGMENTS

362 363

The authors would like to thank the National Science Council of Taiwan under Contract No. NSC

364

100-2628-E-011-019-MY3 and Architecture and Building Research Institute of Taiwan for

365

financially supporting this research.

366 367

REFERENCES

368 369 370

ACI Committee 318. (2008). Building Code Requirements for Structural Concrete (ACI 318-08)

and Commentary (ACI 318R-08), American Concrete Institute, Farmington Hills, MI.

371

Aschheim, M., and Moehle, J. P. (1992). “Shear strength and deformability of RC bridge columns

372

subjected to inelastic displacements,” Report No. UCB/EERC-92/04, Earthquake Engineering

373

Research Center, University of California, Berkeley, CA.

374

Eurocode 2 (1992). Design of concrete structures. General rules and rules for buildings. 12

375 376 377 378 379

European Committee for Standardization. FEMA-267b. (1999). Interim Guidelines, Advisory No. 2, Supplement to FEMA 267. Federal

Emergency Management Agency, Washington, DC. Higgins, C., and Farrow, III W.C. (2006). “Tests of reinforced concrete beams with corrosion-damaged stirrups,” ACI Structural Journal, 103(1), 133-141.

380

Hsu, T. T. C. (1992). Unified Theory of Reinforced Concrete, CRC Press, Boca Raton, FL.

381

Juarez, C. A., Guevara, B., Fajardo, G., and Castro-Borges P. (2011). "Ultimate and nominal

382

shear strength in reinforced concrete beams deteriorated by corrosion," Engineering Structures,

383

33(12), 3189-3196.

384

Kato, E., Iwanami, M., and Yokota, H. (2006). “Deterioration in ductility of RC beams with

385

corroded reinforcement,” Proceedings of the 2nd fib International Congress, Naples, Italy,

386

1-8.

387

Ou, Y. C., Tsai, L. L., and Chen, H. H. (2012). “Cyclic performance of large-scale corroded

388

reinforced concrete beams,” Earthquake Engineering and Structural Dynamics, 41(4),

389

593-604.

390 391 392 393

Rodriguez, J., Ortega, L. M., and Casal, J. (1997). “Load carrying capacity of concrete structures with corroded reinforcement,” Construction and Building Materials, 11(4), 239-248. Val, D. V. (2007). "Deterioration of strength of RC beams due to corrosion and its influence on beam reliability," Journal of Structural Engineering, ASCE, 133(9), 1297-1306.

394

Wang, X. H., Gao, X. H., Li, B., and Deng, B. R. (2011). “Effect of bond and corrosion within

395

partial length of shear behavior and load capacity of RC beam,” Construction and Building

396

Materials, 25, 1812-1823.

397 398

Xia, J., Jin, W. L., and Li, L. Y. (2011). “Shear performance of reinforced concrete beams with corroded stirrups in chloride environment,” Corrosion Science, 53, 1794-1805.

13

Table 1 Click here to download Table: Table. 1.doc

Table 1: Corrosion measurement Specimen

∆w (%)

Bt-0 0.00 Bt-3 2.90 Bt-6 5.87 Bt-11 11.73 Bt-12 12.40 Bt-16 15.67 Bt-35 35.06 ∆w is weight loss; Aavg is

2

Aavg (mm

126.67 121.57 116.44 106.19 105.15 99.33 65.58 average

Wcr (mm) ) Amin (mm ) pmax (mm) Observation Beam specimen specimen 126.67 0.00 0.00 0.00 98.85 1.47 0.17 0.27 83.91 2.56 1.28 0.65 70.45 2.62 0.97 1.38 77.98 2.97 1.74 1.70 59.29 4.73 1.82 1.93 0.00 6.35 4.56 4.06 residual cross-sectional area; Amin is minimum residual 2

cross-sectional area; pmax is maximum pit depth; and Wcr is total crack width.

Table 2 Click here to download Table: Table. 2.doc

Table 2. Performance indicators Specimen Bt-0 Bt-3 Bt-6 Bt-11 Bt-12 Bt-16 Bt-35  y is idealized

 y (%)

Ppeak (kN)

 u (%)



 p (%)

0.80 365.40 4.97 6.21 4.17 0.70 354.57 4.67 6.66 3.97 0.70 333.38 4.27 6.10 3.57 0.70 340.49 2.87 4.09 2.17 0.68 336.88 2.67 3.97 1.99 0.70 331.68 2.44 3.49 1.74 0.65 318.80 1.79 2.77 1.14 yield drift;  u is ultimate drift; Ppeak is peak applied load;  is

ductility; and  p is plastic rotation.

Table 3 Click here to download Table: Table. 3.doc

Table 3 Residual shear strength by concrete Specimen

Initial Vc (kN)

Vc at ultimate (kN)

Bt-0 Bt-3 Bt-6 Bt-11 Bt-12 Bt-16 Bt-35

137.03 133.94 132.76 130.95 130.29 129.84 126.69

0.00 0.00 0.00 0.00 1.31 22.26 52.14

Vc loss due to

Vc loss due to

corrosion (%) 0.00 2.25 3.11 4.43 4.92 5.24 7.54

ductility (%) 100.00 100.00 100.00 100.00 98.99 82.86 58.85

Table 4 Click here to download Table: Table. 4.doc

Table 4 Residual shear strength by shear reinforcement Vs _ avg (kN) Vs _ min (kN) Vs _ min_ avg (kN) Specimen Bt-0 473.45 473.45 473.45 Bt-3 454.37 411.91 369.45 Bt-6 435.19 374.39 313.60 Bt-11 396.88 330.09 263.29 Bt-12 392.99 342.23 291.47 Bt-16 371.25 296.42 221.59 Bt-35 245.12 122.56 0.00 Vs _ avg is based on average residual cross-sectional area Aavg ; Vs _ min is based on the

minimum residual cross-sectional area Amin ; and Vs _ min_ avg is based on Aavg for one leg and Amin for the other leg.

Figure 1 Click here to download Figure: Fig. 1.pdf

Figure 1. Examples of transverse reinforcement corrosion in reinforced concrete buildings

Figure 2 Click here to download Figure: Fig. 2.pdf

50

50

#4@100

50

1850

40 #4@100

#4@100

900

50

500

500

490 490

800

#4@75

#9

300 #9

Unit : mm

Unit : mm

(b)

200

(a)

600

#9

#4@100

500

Unit : mm

(c)

(d)

Figure 2. Specimen design: (a) side view; (b) cross-sectional view; (c) corrosion observation specimen; and (d) construction of specimens.

Figure 3 Click here to download Figure: Fig. 3.pdf

Cupper plate hoop longitudiual bars

Cupper plate longitudiual bars hoop

5% NaCl Solution

5% NaCl Solution

+ D.C Power -

+ D.C Power -

Source

Source

(a)

(b)

(c)

(d)

Figure 3. Accelerated corrosion process: (a) elevation view; (b) top view; (c) electric wires; and (d) overall view of test setup

Figure 4 Click here to download Figure: Fig. 4.pdf

-

P

0 4 0 5

0 4 r o o l f g d n n o r e t ds en ee o r t F d de ex xi i F F 5 6 m c : t i n U

0 5 2 7 1 2

0 9

Figure 4. Cyclic loading setup

Figure 5 Click here to download Figure: Fig. 5.pdf

(a)

(b)

(c)

(d)

(e)

(f)

Figure 5.Corroded hoops for specimen Bt-35 from the fixed end of the beam: (a) the first hoop; (b) the second hoop; (c) the third hoop; (d) the fourth hoop; (e) the fifth hoop; (f) the sixth hoop.

Figure 6 Click here to download Figure: Fig. 6.pdf

0.15 0.04

0.04

0.5

0.15

0.1 0.05

0.15

0.14

0.08

0.1

1.3

0.3

0.2

0.08

0.1

0.04

0.06

0.25 0.04

0.1

0.08

0.15

0.35

0.45 0.35

0.35

0.3

0.08 0.2

0.1

F

T

R

0.2

B

F

T

R

(a)

B

(b)

0.2

0.08

0.2

0.25

0.04

0.75

0.04

0.08

0.08

0.04

0.04

0.02

0.04

0.08 0.6

0.25

0.85

0.04

0.1 0.1

0.1

1.42

0.04

0.25

0.06

0.04

0.04

0.04

1.6

0.06

0.04

0.08

0.4

0.2

0.04

F

T

R

0.1

B

F

T

R

(c) 0.04

0.1 0.04

0.1

0.04

0.1

0.15

0.2

0.04 0.04

0.04

0.04

0.35

0.25

0.15

0.5

0.5

1.6 0.06

0.5 0.04

0.04

0.7 0.04

0.5

0.1

0.35

0.04

0.04

0.04

0.8

0.04 0.3

0.1

B

(d) 0.04

0.2

0.6

0.04

0.75

0.04 0.1

0.08

0.35

0.55 0.2 0.2

F

0.2

T

0.1

(e)

R

0.1

0.35 0.06

0.35

0.04

0.45

0.5 0.3

0.2

0.04 0.4

B

3.5 1.0

F

0.04

0.2

T

0.3

0.04

R

0.35

B

(f)

Figure 6. Corrosion crack pattern: (a) Bt-3; (b) Bt-6; (c) Bt-11; (d) Bt-12; (e) Bt-16; and (f) Bt-35 (the number shown in the figure denotes crack width (mm); F, T, R, and B denote front, top, rear, and bottom faces of the beam, respectively; the front face is the face shown in Fig. 7)

Figure 7 Click here to download Figure: Fig. 7.pdf

(a)

(b)

(c)

(d)

(e)

(f)

(g) Figure 7. Damage distribution: (a) Bt-0 at -5% drift (1st cycle); (b) Bt-3 at -5% drift (1st cycle); (c) Bt-6 at -5% drift (1st cycle); (d) Bt-11 at -3% drift (1st cycle); (e) Bt-12 at -3% drift (1st cycle); (f) Bt-16 at -3% drift (1st cycle); and (g) Bt-35 at -2% drift (1st cycle) (Negative drift denotes the beam is displaced downward at the loading end)

Figure 8 Click here to download Figure: Fig. 8.pdf

400

400

300

300

Idealized response

200 Applied load (KN)

Applied load (KN)

Test result

Test result

100 0 -100 -200

100 0 -100 -200

:cover spalling :hoop fracture :yield point :Ultimate point

-300

Idealized response

200

:cover spalling :hoop fracture :yield point :Ultimate point

-300

-400

-400

-6

-4

-2

0 Drift (%)

2

4

6

-6

-4

-2

(a)

6

Test result

300

Idealized response

200

Applied load (KN)

Applied load (KN)

4

400 Test result

300

100 0 -100 -200

Idealized response

200 100 0 -100 -200

:cover spalling :hoop fracture :yield point :Ultimate point

-300

:cover spalling :hoop fracture :yield point :Ultimate point

-300

-400

-400 -6

-4

-2

0 Drift (%)

2

4

6

-6

-4

-2

(c) 400

0 Drift (%)

2

4

6

(d) 400

Test result

300

Test result

300

Idealized response

200

Applied load (KN)

Applied load (KN)

2

(b)

400

100 0 -100 -200

Idealized response

200 100 0 -100 -200

:cover spalling :hoop fracture :yield point :Ultimate point

-300

:cover spalling :hoop fracture :yield point :Ultimate point

-300

-400

-400 -6

-4

-2

0 Drift (%)

2

4

6

-6

-4

-2

(e)

2

4

6

2

4

6

400

Test result

300

0 Drift (%)

(f)

400

Bt-0 Bt-3 Bt-6 Bt-11 Bt-12 Bt-16 Bt-35

300

Idealized response

200 Applied load (KN)

Applied load (KN)

0 Drift (%)

100 0 -100 -200

100 0 -100 -200

:cover spalling :hoop fracture :yield point :Ultimate point

-300

200

-300 -400

-400 -6

-4

-2

0 Drift (%)

2

4

6

-6

-4

(g)

-2

0 Drift (%)

(h)

Figure 8. Hysteretic behavior: (a) Bt-0; (b) Bt-3; (c) Bt-6; (d) Bt-11; (e) Bt-12; (f) Bt-16; (g) Bt-35; and (h) envelope responses of all specimens

Figure 9 Click here to download Figure: Fig. 9.pdf

Rust Uncorroded d bars ‫ݓ‬௖௥௜

‫ܥ‬௖

(a)

(b)

Figure 9. Corrosion substances: (a) filled in the space surrounding the bar and cover cracks; and (b) escaping out of concrete

Figure 10 Click here to download Figure: Fig. 10.pdf

700

7

600

6

Shear (kN)

500

5

400



300

4 Ppeak Vn(Vs _avg) Vn(Vs _min) Vn(Vs _avg_min)

200 100 0

Vn(Vs _avg) Vn(Vs _avg_min) Vn(Vs _min) Vn(Vs _avg)_regression Vn(Vs _avg_min)_regression Vn(Vs _min)_regression

0

10

3

20 Weight loss (%)

30

40

2 -0.3

-0.15

(a)

0 0.15 0.3 (Vc+Vs-Vy)/(Vinc+Vins)

0.45

0.6

(b)

Figure 10. (a) Residual shear strength estimation; and (b) relationship between residual shear strength and ductility

Figure Caption List

1

FIGURE CAPTIONS

2 3

Figure 1. Examples of transverse reinforcement corrosion in reinforced concrete buildings

4 5

Figure 2. Specimen design: (a) side view; (b) cross-sectional view; (c) corrosion observation specimen; and (d) construction of specimens.

6 7

Figure 3. Accelerated corrosion process: (a) elevation view; (b) top view; (c) electric wires; and (d) overall view of test setup

8

Figure 4. Cyclic loading setup

9 10 11

Figure 5.Corroded hoops for specimen Bt-35 from the fixed end of the beam: (a) the first hoop; (b) the second hoop; (c) the third hoop; (d) the fourth hoop; (e) the fifth hoop; (f) the sixth hoop.

12 13 14

Figure 6. Corrosion crack pattern: (a) Bt-3; (b) Bt-6; (c) Bt-11; (d) Bt-12; (e) Bt-16; and (f) Bt-35 (the number shown in the figure denotes crack width (mm); F, T, R, and B denote front, top, rear, and bottom faces of the beam, respectively; the front face is the face shown in Fig. 7)

15 16 17 18

Figure 7. Damage distribution: (a) Bt-0 at -5% drift (1st cycle); (b) Bt-3 at -5% drift (1st cycle); (c) Bt-6 at -5% drift (1st cycle); (d) Bt-11 at -3% drift (1st cycle); (e) Bt-12 at -3% drift (1st cycle); (f) Bt-16 at -3% drift (1st cycle); and (g) Bt-35 at -2% drift (1st cycle) (Negative drift denotes the beam is displaced downward at the loading end)

19 20

Figure 8. Hysteretic behavior: (a) Bt-0; (b) Bt-3; (c) Bt-6; (d) Bt-11; (e) Bt-12; (f) Bt-16; (g) Bt-35; and (h) envelope responses of all specimens

21 22

Figure 9. Corrosion substances: (a) filled in the space surrounding the bar and cover cracks; and (b) escaping out of concrete

23 24

Figure 10. (a) Residual shear strength estimation; and (b) relationship between residual shear strength and ductility

25