1 Choi et al.

DEVELOPMENT OF TUBE FEEDING EQUIPMENT FOR 1 CONSTRUCTION OF CONTINUOUSLY REINFORCED CONCRETE 2 PAVEMENT 3 4

5 6

Pangil Choi, Ph.D. Corresponding Author 7 Senior Research Associate, Center for Multidisciplinary Research in Transportation 8 Texas Tech University, Lubbock, TX 79409 9 Tel: 806-742-3523 Fax: 806-742-2644; E-mail: pangil@hotmail.com 10 11 Yongsoo Kil, Ph.D. 12 Research Engineer, Technology Research Center, Samwoo IMC Co., Ltd, Seoul, South Korea 13 E-mail: kil0911@daum.net 14 15 Beom Jun Chon, Ph.D. 16 Deputy Manager, Technology Research Center, Samwoo IMC Co., Ltd, Seoul, South Korea 17 E-mail: najunya@samwooimc.com 18 19 Moon C. Won, Ph.D., P.E. 20 Professor, Department of Civil, Environmental, and Construction Engineering 21 Texas Tech University, Lubbock, TX 79409 22 Tel: 806-834-2248 Fax: 806-742-2644; E-mail: moon.won@ttu.edu 23 24

25 Word count: 4,531 words text + 7 figures x 250 words (each) = 6,281 words 26 27 28 29 30 31 32 Submission Date: July 31, 2015 33

2 Choi et al.

ABSTRACT 34 In South Korea, the performance of continuously reinforced concrete pavement (CRCP) has been 35 better than that of jointed concrete pavement (JCP). However, most of the Portland cement 36 concrete (PCC) pavements in South Korea have been JCP, primarily due to the difficulty in 37 providing extra space along the edge of the pavement needed for concrete delivery in CRCP 38 construction. About 80 % of Korea is covered by mountains, and it is quite expensive to acquire 39 and prepare land to provide space for concrete delivery. To build CRCP at a reasonable cost in 40 South Korea, a CRCP construction method is needed that does not require extra space for 41 concrete delivery. Several different methods have been tried that allowed the delivery of concrete 42 to the front of a paver, including a pre-assembly method. Although the pre-assembly method did 43 not require extra space for concrete delivery, it did require a high level of labor and its 44 productivity was quite low. A more efficient method was needed. In a tube-feeding method 45 (TFM), concrete can be supplied from the front of a paver with better productivity. Efforts were 46 made to improve the capability of TFM in placing longitudinal steel at correct elevations. The 47 ability and productivity of the newly developed TFM in placing reinforcing steel in correct 48 elevations were quite satisfactory, and CRCP was constructed with the new TFM. Early age 49 performance of CRCP built with TFM in terms of characteristics of transverse cracks has been 50 satisfactory. 51 52 53 Keywords: Continuously reinforced concrete pavement: CRCP construction: Tube-feeding 54 method 55 56

3 Choi et al.

INTRODUCTION 57 In South Korea, about 60 percent of expressway lane miles of are Portland cement 58

concrete (PCC) pavement, and the remaining 40 percent is asphalt pavement. The reason for a 59 larger portion of PCC pavement in expressways is the ability of PCC pavements to provide better 60 performance for heavy truck traffic with lower maintenance cost. At the early ages of PCC 61 pavement usage in Korea, many miles of continuously reinforced concrete pavement (CRCP) 62 were built; however, a higher initial cost and concerns about numerous transverse cracks resulted 63 in discontinuation of the use of CRCP and more use of jointed concrete pavement (JCP). Another 64 reason for the discontinued use of CRCP is the constructability issue. In South Korea, more than 65 80 percent of land is mountains, and extra space is needed in CRCP construction for concrete 66 delivery; therefore, CRCP construction in mountainous areas when steel is placed with chairs is 67 cost prohibitive. Meanwhile, the performance of CRCP built at early ages in Korea has been 68 better than that of JCP built even at later ages (1, 2). Also, CRCP and JCP test sections built in 69 1985 by the same contractor - with the same pavement structure (slab thickness and base 70 type/thickness), concrete materials, traffic, and environmental loading - clearly indicated better 71 performance of CRCP than JCP. This finding has lately prompted a second look at the potential 72 use of CRCP in Korea. However, the issue of extra space needed for CRCP construction to 73 provide concrete still remained. Extra space is not needed for JCP construction, since concrete 74 can be supplied in front of a paver onto a prepared base. For CRCP construction to be feasible in 75 South Korea, concrete has to be supplied in front of a paver so that extra space is not needed. 76

77 There are several methods available that allow the delivery of concrete from the front of a paver, 78 eliminating the need for extra space. Methods include the pre-assembly method and the tube 79 feeding method (TFM). In the pre-assembly method, longitudinal and transverse steel are 80 assembled at the jobsite and moved to the front of a paver. Concrete is supplied from the front of 81 a paver, as shown in Figure 1(a). However, due to the weight of the steel assembly as well as the 82 limited extension of the concrete delivery machine, the length of the steel assembly in the 83 longitudinal direction is quite limited, which requires extensive moving and tying longitudinal 84 steel operations, and accordingly, the productivity of this method is quite low. 85 86 TFM, which allows the concrete delivery in front of a paver without needing extra side space, 87 has been utilized in the US and Europe, with varying degrees of performance. It has been 88 reported that the primary cause for poor performance of CRCP with TFM was the inability to 89 ensure the steel is placed at a proper depth, and the use of TFM in CRCP construction is not 90 recommended (3, 4). On the other hand, a detailed study conducted to compare the long-term 91 performance of CRCP built with the traditional chair method and TFM indicated, from a 92 practical standpoint, no difference between them, as shown in Figure 1(b) (5). It appears that, as 93 long as longitudinal steel is placed at a proper depth within a given tolerance, CRCP construction 94 with TFM can be a viable option in regions where providing extra space for concrete delivery is 95 quite difficult and expensive, such as in South Korea. Compared to the pre-assembly method, the 96 productivity of TFM will be much higher since no moving operations of assembled steel bars are 97 needed and longitudinal steel can be tied together in advance. Efforts were made to develop a 98 TFM that ensures placement of longitudinal steel at a proper depth within allowed tolerances. 99 100

4 Choi et al.

101 (a) 102

103

104 (b) 105

106 FIGURE 1(a) CRCP construction by pre-assembly method and (b) Performance of CRCP 107

built with chairs & TFM (5) 108 109

CRCP PERFORMANCE BUILT WITH TFM IN THE PAST 110 CRCP used to be constructed with TFM in several states in the US and in Europe. Figure 111

2(a) illustrates CRCP construction with TFM in Virginia (6). It is noted that the space between a 112 tube feeder and a paver is rather large, and concrete is placed between the tube feeder and a 113 paver, which could make it quite difficult to control the elevations of steel. It was reported that 114 distresses developed due to the variations in steel depth beyond a tolerance level. If the steel was 115 placed too high in the concrete slab, corrosion of steel and spalling of concrete resulted, as 116 shown in Figure 2(b). On the other hand, if steel is placed too low, crack widths at the top of the 117 slab could be excessive (3). The same type of issue has been reported in South Carolina, where 118 variations in steel depth were quite large in CRCP built with TFM, resulting in poor performance. 119 On the other hand, an in-depth investigation conducted in Illinois showed comparable 120 performances of CRCPs built with a chair method or TFM, as shown in Figure 1(b) (5). In Texas, 121

5 Choi et al.

CRCP sections built with TFM performed satisfactorily, and the Texas Department of 122 Transportation (TxDOT) specifications allow both a chair method and TFM for steel placement 123 (7), even though the last time TFM was used in Texas was in the middle of 1980s. In Europe, 124 CRCP sections built with TFM have provided satisfactory performance (8). Based on the past 125 experience, it appears that the performance of CRCP built with TFM could be comparable to that 126 of CRCP built with a chair method, as long as the elevations of longitudinal steel are maintained 127 within a given tolerance. 128

129

130 (a) 131

132

133 (b) 134

135 FIGURE 2 (a) CRCP construction with TFM in Virginia and (b) Distress due to steel 136

placed near the surface (6) 137 138

6 Choi et al.

IMPROVEMENTS OF TFM 139 This section describes the efforts made to improve TFM in placing longitudinal steel at a 140

correct depth with minimal variations. First, potential causes for a large variance in steel depths 141 with TFM in the past were identified. Next, means to minimize variations in steel depth were 142 devised, and tubes and other parts with proper characteristics were made and assembled. 143

144 Factors Affecting Steel Elevation in TFM 145

In manual steel placement method with chairs, the elevations of steel in areas between 146 adjacent chairs are secured by transverse steel. On the other hand, in TFM, there are no chairs, 147 and therefore no vertical support, so the steel elevations should be secured by other means. Since 148 there is no vertical support for steel in TFM, it is essential that longitudinal steel elevation should 149 be guided by tube elevation, and adequate tension is maintained in longitudinal steel during 150 construction. Since concrete is supplied from the front of the tube feeding machine at a certain 151 height, there will be concrete weight loading on longitudinal steel, which could impact the final 152 elevation of longitudinal steel. The maximum steel deflection due to self-weight of steel and 153 over-burden concrete weight is computed by the following equations: 154

155 𝑑𝑑2𝑦𝑦𝑑𝑑𝑥𝑥2

−𝑃𝑃𝐸𝐸𝐸𝐸𝑦𝑦 =

𝑤𝑤2𝐸𝐸𝐸𝐸

𝑥𝑥2 −𝑤𝑤𝑤𝑤2𝐸𝐸𝐸𝐸

𝑥𝑥 (1)

𝑦𝑦 =𝐸𝐸𝐸𝐸𝑤𝑤𝑃𝑃2

𝑒𝑒(𝐿𝐿−𝑥𝑥)�𝑃𝑃𝐸𝐸𝐸𝐸 + 𝑒𝑒𝑥𝑥

�𝑃𝑃𝐸𝐸𝐸𝐸 − 𝑒𝑒𝐿𝐿

�𝑃𝑃𝐸𝐸𝐸𝐸 − 1

𝑒𝑒𝐿𝐿�𝑃𝑃𝐸𝐸𝐸𝐸 + 1

−𝑤𝑤2𝑃𝑃

𝑥𝑥2 +𝑤𝑤𝑤𝑤2𝑃𝑃

𝑥𝑥 (2)

where, 156 y = deflection of steel 157 x = distance from the end of the tube 158 P = tensile force in steel 159 w = weights of steel and over-burden concrete 160 L = distance between the end of tube and the front of the mold 161 E = modulus of elasticity of steel 162 I = moment of inertia 163

164 Equation 1 is the governing equation and, with the boundary condition of y=0 at x=0 and x=L/2, 165 the Equation 2 is obtained. In the modeling of tube feeding equipment, it was assumed that the 166 steel is fixed at the front of the mold and supported by roller at the end of the tube, which may 167 not represent the real situation perfectly; however, it was considered that the assumptions were 168 adequate for this work. Figure 3(a) illustrates the relationship between exposed steel length, 169 defined here as the distance between the end of a tube and the start of the mold, and deflection of 170 the steel for various tensile forces in the steel. The figure shows beneficial effects of tensile 171 forces in the steel on the deflections while, with no tensile forces, the deflections increase 172 exponentially with exposed length. For example, with the exposed length of 1,000 mm, the 173 deflections could exceed the tolerance limit of 12.5 mm. It also shows that tensile forces in the 174 range of 196 N and 589 N reduce steel deflections substantially, while no appreciable difference 175 is observed in steel deflections for the range of tensile forces analyzed here. Based on this 176 analysis, it was decided to provide tensile forces in steel of the values in the range analyzed here 177

7 Choi et al.

in TFM. Several methods to induce tensile forces in the steel were considered, and it was 178 determined that frictional forces by the use of curved tubes would be the most efficient and 179 practical. To determine the optimum radius of the curved tubes, experiments were conducted. 180 Tubes with six radii were made – 5,000, 6,000, 7,000, 8,000, 9,000, and 10,000 mm – and pull-181 out testing conducted. The inside and outside diameters of the tubes were 50 mm and 60 mm, 182 respectively. Figure 3(b) shows the testing results: the frictional force decreased rather quickly as 183 the radius increased from 5,000 mm to 8,000 mm, while it changed little for the radius above 184 8,000 mm. Based on the analysis and test results, a radius of 9,500 mm for tubes and 1,300 mm 185 for the distance between the end of tubes and the front of the mold were selected for tube feeding 186 equipment. 187 188

189 (a) 190

191

192 (b) 193

194 FIGURE 3 (a) Exposed rebar length vs deflection and (b) Radius of tube vs pull-out force 195

0.0

2.5

5.0

7.5

10.0

12.5

15.0

0 30 60 90 120 150

Defle

ctio

n [m

m]

Exposed Rebar Length [cm]

P = 0P = 196 N (44 lbf)P = 392 N (88 lbf)P = 589 N (132 lbf)

y = 0.000015x2 - 0.292874x + 1,632.524143R² = 0.969169

-

100

200

300

400

500

600

5,000 6,000 7,000 8,000 9,000 10,000

Pull-

out F

orce

[N]

Radius of Tube [mm]

8 Choi et al.

196 Another factor influencing steel placement depths could be concrete consolidation under the 197 paver. In general, a paver progresses at a rate of one meter per minute, with vibrators on at about 198 8,000 vibrations per minute. The concrete consolidation operation might affect the elevations of 199 steel. To investigate the effects of concrete consolidation on the settlement of steel, an 200 experiment was conducted. Forms for three concrete blocks, each 500 mm in width, 300 mm in 201 height, and 10,000 mm in length, were fabricated and two #6 reinforcing steel (D=19 mm) with 202 9,000 mm long were placed at the mid-depth with tie wires at two blocks, as shown in Figure 203 4(a). Concrete was placed into all three forms, tie wires that were supporting the steel in two 204 blocks were removed, and two #6 reinforcing bars (D=19 mm) were placed on top of the 205 concrete in the third block, as shown in Figure 4(b). A slip-form paver proceeded with a speed of 206 one meter per minute, with one vibrator in each concrete block running at about 8,000 vibrations 207 per minute, which simulated actual concrete consolidation occurring during concrete paving with 208 a slip-form paver. The depths of steel bars were evaluated after the slip-form paver completed 209 consolidation of concrete. The settlements of steel bars placed at the mid-depth were in the range 210 of 5 to 20 mm, while those of steel bars placed at the concrete surface varied from 3 mm to 18 211 mm. Most states limit the vertical variation of steel depth to ± 12.5 mm. The settlements 212 observed at some locations in this experiment exceeded the maximum tolerance; however, the 213 maximum settlements occurred at the end parts of the bars. The settlements in other parts of the 214 bars were quite small, within 8 mm. This experiment indicated that, if steel bars are continuous, 215 vibration operation by a slip-form paver would not cause settlements of steel bars greater than 216 the afore-mentioned tolerance limits. 217 218

219

220 (a) 221

222

9 Choi et al.

223 (b) 224

225 FIGURE 4 (a) Forms for steel settlement experiment and (b) Vibration of concrete 226

227 This experiment did not simulate the compression of concrete that occurs during concrete 228 placement due to the weight of a slip-form paver. However, it is postulated that the compressive 229 stresses on the steel due to the compression of concrete are equal in all directions, as concrete is 230 liquefied by vibration, and the settlements of steel bars due to compressive stresses might be 231 negligible. Based on the findings from analysis and testing results, it was decided that tubes with 232 a radius of 9,500 mm would be used for the manufacture of a tube feeding equipment. 233 234 Structure of Tube Feeding Equipment (TFE) 235

TFE is self-propelled equipment with its own power system and consists of three 236 components – concrete receiving module, tube module, and vibration module. The paving 237 system using TFE developed in this study is shown in Figure 5(a). It shows that longitudinal 238 steel was placed on the edges of the base layer, which provides a space in the middle portion for 239 concrete trucks to have access to the concrete delivery equipment. Concrete is delivered to the 240 TFE through a conveyor belt. Figure 5(b) illustrates the TFE, with concrete receiving and tube 241 modules shown. Figure 5(c) shows the side view of the TFE. Tubes have flared ends, with 105 242 mm diameter. Also, tubes with various lengths were used, from 1,800 mm to 2,250 mm, to 243 accommodate the various radii of longitudinal steel in front of the TFE. Figure 5(d) illustrates 244 underside of the TFE, with tubes, vibrators and auger shown. Rubber plates were installed to 245 protect the ends of tubes, so that concrete cannot get into the tubes. The vibrator module 246 consolidates the concrete in order to reduce any stresses imposed on steel by concrete, with the 247 goal of minimizing the variations of steel elevations. 248

249

10 Choi et al.

250 (a) 251

252

253 (b) 254

255

11 Choi et al.

256 (c) 257

258

259 (d) 260

261 FIGURE 5 (a) Paving system with TFE, (b) Front view of TFE, (c) Side view of TFE, and 262

(d) Underside view of TFE 263 264

CRCP CONSTRUCTION WITH TFE 265 Once the TFE was assembled, its performance in placing steel in the right elevations 266

within a given tolerance level was evaluated. The testing was conducted on September 26, 2013 267 in a section of freeway that was not in use due to re-alignment of the freeway. At the beginning 268 of the test section, longitudinal steel was placed with chairs and transverse steel, quite similar to 269 normal manual placement of steel. A paver was brought to the beginning of the test section, TFE 270 was placed to the front of the paver and longitudinal steel was fed into the TFE. Concrete was 271 supplied from the front of the TFE. While the TFE progressed, the elevations of longitudinal 272

12 Choi et al.

steel were periodically checked by removing concrete and exposing longitudinal steel. They 273 were all within the tolerance limits of ±12.5 mm of a target elevation, except at a pavement edge, 274 where excessive mortar was accumulated at the edge due to super-elevation of the section. 275 Figure 6(a) shows the concrete being placed by TFE and the tension in the steel being adequate 276 to support about 70 kgf (686 N) weight. Once the concrete was hardened, two full-depth concrete 277 saw cuts were made in a transverse direction at 300 mm apart. Concrete block was removed and 278 steel depths were evaluated as shown in Figure 6(b). There were a total of 49 longitudinal bars, 279 and all the bars were within ±12.5 mm tolerance limits, except at the pavement edge. 280

281 The ability of TFE in placing longitudinal steel at a correct elevation was confirmed in a trial 282 with a test section, and TFE was used to construct CRCP as part of an expressway system in 283 Korea. There was a jointed concrete pavement construction project in Goryeong, in the southern 284 part of Korea, and a field change was made to place 400 meters of CRCP. The construction 285 began on December 5, 2013 and was completed in the same day. Total width of the pavement 286 was 8.4 meters. The construction of CRCP with TFE in this project is shown in Figure 5(a). 287 Longitudinal steel bars of 9 m length were tied with couplers, and placed along the edge of the 288 base one day prior to the concrete placement. 289 290

291 (a) 292

293

13 Choi et al.

294 (b) 295

296 FIGURE 6 (a) Concrete placement by TFE and (b) Steel depth evaluation 297

298 During the construction, the elevations of steel were checked at six different locations just after 299 the TFE. The locations were identified as a distance from a string line as well as a marking at 300 the string line. Concrete was removed to expose the steel, steel depth measured from the bottom 301 of the slab, and concrete was put back. Once the paver passed, the depth of the steel was 302 measured at the same location. These comparisons were made at six locations, and the difference 303 in steel elevations at these six locations was almost negligible, ranging from 1 mm to a 304 maximum of 6 mm, indicating that concrete consolidation operations by a paver do not alter the 305 steel depth appreciably with the tension provided in TFE. 306 307 Performance of CRCP constructed with TFE 308

The CRCP section built with TFE is less than two years old, and it is too early to evaluate 309 its performance. Instead, early age behavior in terms of transverse crack characteristics was 310 evaluated. Three traits of transverse cracks are reported to be related to long-term performance: 311 crack spacing, crack width, and meandering of cracks. Crack spacing and crack width are 312 determined primarily by concrete placement temperature, the amount of longitudinal steel, and 313 concrete properties such as coefficient of thermal expansion (9, 10). These variables are 314 independent of construction methods – either traditional chair method or TFE. Accordingly, field 315 evaluation of CRCP built with TFE focused on the meandering characteristics of transverse 316 cracks. One document recommends the placement of transverse steel primarily in order to 317 achieve straighter transverse cracks (11). Since no transverse steel was used in this test section, 318 the probability of meandering cracks might increase. One method proposed to quantify the 319 degree of meandering of cracks is to quantify the skewness of transverse cracks, called 320 randomness index (RI) (12). RI is a mathematical model for predicting the randomness rating 321 using certain physical measurements of a crack. When a crack is straight and perpendicular to the 322 pavement centerline, RI of the crack would be 5.5, while that with 45 degrees to the pavement 323 centerline has an RI of 2.1. Cracks more deviant from straight and perpendicular to the pavement 324 centerline will have a larger RI value of cracks. RI values evaluated in CRCP constructed with 325 the chair method were in the range of 1.6 to 5.0 (12). RI values were estimated on typical cracks 326

14 Choi et al.

observed in the section built with TFE and the results, shown in Figure 7, show RI values in the 327 range of 2.4 to 5.5, which is well within the range of values reported for cracks in CRCP built 328 with the chair method. 329

330

331 332

FIGURE 7 Randomness index (RI) distribution 333 334

Potential Issues Related to CRCP Constructed with TFE 335 Transverse steel design has been based on subgrade drag theory, which requires more 336

transverse steel (smaller spacing between transverse steel) as more lanes are tied together. The 337 use of subgrade drag theory for the design of transverse steel implies keeping longitudinal cracks 338 if they occur, reasonably close. However, longitudinal joints are provided at every 3.6 m spacing, 339 and curling stresses for slabs with 3.6 m width are maintained quite small (13). Accordingly, 340 longitudinal cracks are not likely, if longitudinal joint saw cuts are made early enough with a 341 proper depth. Also, in practice, subgrade drag theory is not used, and transverse steel spacing is 342 normally fixed regardless of how many lanes are tied together. For example, in Illinois DOT, #4 343 bars (D=12.7 mm) are placed at 1.2 m spacing (14), while in Texas DOT, #5 bars (D=16 mm) 344 are placed at 1.2 m spacing (15). The amount of transverse steel used in practice when more than 345 4 lanes are tied together is smaller than the value from subgrade drag theory, which implies that, 346 in current practice, the role of transverse steel is to provide support for longitudinal steel. In 347 jointed concrete pavement, transverse steel is not placed. Since TFE developed in this study 348 places and is able to keep longitudinal steel at correct elevation within a tolerance level, 349 transverse steel may not be needed when placing CRCP with TFE. 350

351 CONCLUSIONS AND RECOMMENDATIONS 352

In South Korea, the performance of CRCP has been better than that of JCP. However, 353 most of the Portland cement concrete pavements in South Korea have been JCP, primarily due to 354 the difficulty in providing extra space needed for concrete delivery in CRCP construction, 355 because 80% of South Korea is mountainous. To build CRCP at a reasonable cost in South 356 Korea, a CRCP construction method is needed that does not require extra space for concrete 357 delivery – a method that would allow the supply of concrete from the front of the paver. In a tube 358

-

1.0

2.0

3.0

4.0

5.0

6.0

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37

Rand

omne

ss In

dex

(RI)

Crack ID

15 Choi et al.

feeding method (TFM) that was used in the past for longitudinal steel placement, concrete can be 359 supplied from the front of a paver, eliminating the need for extra space on the side of the 360 pavement for the delivery of concrete. 361

362 The use of TFM in CRCP construction has been banned by some states due to the difficulty in 363 placing longitudinal steel at the right elevations, which resulted in performance issues. Even 364 where TFM is allowed, its use in CRCP construction has almost disappeared due to the advent of 365 more efficient and inexpensive chairs and contractors’ increasing familiarity with steel 366 placement with chairs. However, there are places where TFM could be quite useful for CRCP 367 construction if TFM could place longitudinal steel at correct elevations within a tolerance level. 368 Efforts were made to improve the capability of TFM in placing longitudinal steel at correct 369 elevations, and tube feeding equipment (TFE) was developed that enhanced the capability of 370 placing longitudinal steel at correct elevations. CRCP was constructed with the newly developed 371 TFE and early age performance was monitored. Based on the findings made so far, the following 372 conclusions could be made: 373 374

1. TFE with a capability of placing longitudinal steel within ±12.5 mm of a specified 375 elevation was developed. 376 A. The key to ensuring steel placement at a correct elevation was to keep longitudinal 377

steel in adequate tension. 378 B. Structural analysis showed that 196 N of tensile force in steel was sufficient to limit 379

the settlement of steel within 12.5 mm. The distance from the back of the tubes and 380 the front of the mold, up to 1,500 mm, did not have substantial effects on the steel 381 settlement. 382

C. Several ways were identified for providing tensile force in steel while TFE 383 progressed, and utilizing tubes with an adequate curvature was the simplest way. 384 TFE developed in this study used tubes with 9,500 mm radius. 385

2. To evaluate the viability of CRCP construction with TFE, a 400 meter long CRCP 386 section with 8.2 m wide and 300 mm slab thickness was constructed. 387 A. Splicing of longitudinal steel, which is 9 m long, was made with couplers. 388

Longitudinal steel bars thus pre-assembled at the jobsite were placed at a maximum 389 about 30 degrees from the inset of the tubes to the edges of the pavement, at which 390 point they were placed parallel to the centerline of the pavement at pavement edges. 391 This setup allowed concrete placer in the middle of construction area and in front of 392 the paver. 393

B. The productivity of CRCP construction with a TFE developed in this study was 394 quite satisfactory, only limited by the rate of concrete supply. 395

C. To evaluate elevations of longitudinal steel in hardened concrete, a concrete slab 396 was cut full-depth in a transverse direction at two locations with 300 mm spacing. 397 Locations of longitudinal steel thus evaluated from the slice of the concrete slab 398 were all within the tolerance limits (±12.5 mm from a target elevation) except at a 399 pavement edge due to super-elevation. 400

D. The elevations of longitudinal steel were evaluated during the construction, by 401 removing fresh concrete at random locations. At almost all the locations, they were 402 within the tolerance limits. 403

16 Choi et al.

3. The characteristics of transverse cracks – meandering of cracks – were evaluated. 404 Meandering of cracks in terms of randomness index was comparable to that of CRCP 405 built by traditional manual steel placement method with chairs. 406

407 TFE allows CRCP construction in areas where securing extra space needed for concrete delivery 408 from sides of the roadway is quite expensive. Also, TFE does not require transverse steel, which 409 lowers initial construction cost slightly. Even though the early age behavior of CRCP built with 410 TFE is comparable to that of CRCP built with manual steel placement method, long-term 411 performance needs to be further monitored. 412 413 ACKNOWLEDGMENTS 414 This research was supported by a grant (14TBIP-C073609-01) from Technology Business 415 Innovation Program (TBIP) funded by Ministry of Construction & Transportation of Korean 416 government. 417

17 Choi et al.

REFERENCES 418 1. Seong Sun An. A Comparative Study of Crcp Abd Jcp Performance with 20-Year Field 419

Data in Jung-Bu Expressway. Hanyang University, 2006. 420 2. Kim, H. B., C. K. Ock, I. S. Kim and S. Y. In. A Study on the Field Performance Analysis 421

for Ltpp Section. South Korea, 2006. 422 3. Rasmussen, R. O., R. Rogers and T. R. Rerragut. Continuously Reinforced Concrete 423

Pavement Design & Construction Guidelines. 2011. 424 4. Roesler, J. R. and J. G. Huntley. Performance of I-57 Recycled Concrete Pavements. 425

Urbana, 2009. 426 5. Gharaibeh, N., M. Darter and L. Heckel. Field Performance of Continuously Reinforced 427

Concrete Pavement in Illinois. Transportation Research Record: Journal of the 428 Transportation Research Board, No. 1684, 1999, pp. 44-50. 429

6. Elfino, M., C. Ozyildirim and H. Nair. I-295 CRCP Performance Updates. Virginia 430 Concrete Conference, 2010. 431

7. Texas Department of Transportation. Standard Specifications for Construction and 432 Maintenance of Highways, Streets, and Bridges. Item 360 Concrete Pavement. 433 ftp://ftp.dot.state.tx.us/pub/txdot-info/des/spec-book-1114.pdf Accessed July 10, 2015. 434

8. Carlos Jofré, Joaquín Romero and R. Rueda. Contribution of Concrete Pavements to the 435 Safety of Tunnels in Case of Fire. 2010. 436

9. Choi, P., S. Ryu, W. Zhou, S. Saraf, M. Won, J. Yeon and S. Ha. Project Level 437 Performance Database for Rigid Pavements in Texas, Ii. Lubbock, TX, 2013. 438

10. Suh, Y. C., K. D. Hankins and B. F. McCullough. Early-Age Behavior of Continuously 439 Reinforced Concrete Pavement and Calibration of the Failure Prediction Model in the 440 Crcp-7 Program. 1992. 441

11. Hassan, K. E., J. Chandler, H. Harding and R. Dudgeon New Continuously Reinforced 442 Concrete Pavement Designs. 2005. 443

12. Suh, Y.-C., B. F. McCullough and K. D. Hankins. Development and Application of 444 Randomness Index for Continuously Reinforced Concrete Pavement. Transportation 445 Research Record, No. 1307, 1991, pp. 136-142. 446

13. Choi, S. and M. Won. Design of Tie Bars in Portland Cement Concrete Pavement 447 Considering Nonlinear Temperature Variations. Transportation Research Record: 448 Journal of the Transportation Research Board, No. 2095, 2009, pp. 24-33. 449

14. Illinois Department of Transportation. Highway Standards. 450 http://www.idot.illinois.gov/Assets/uploads/files/Doing-Business/Standards/Highway-451 Standards/216%20Highway%20Standards%20Complete%20Set.pdf Accessed Jul. 10, 452 2015. 453

15. Texas Department of Transportation. Continuously Reinforced Concrete Pavement One 454 Layer Steel Bar Placement. ftp://ftp.dot.state.tx.us/pub/txdot-455 info/cmd/cserve/standard/roadway/crcp113.pdf Accessed Jul. 10, 2015. 456