Self-compacting recycled aggregate concrete using out of service railway 1 superstructures wastes 2

Abstract


Introduction 24
For approximately 130 years, rail traffic has been carried out, mostly on ballasted track. The 25 most demanding requirements, due to the appearance of high-speed networks, are higher axle 26 loads, higher train frequencies and greater environmental awareness, all of which has led to a 27 demand for track systems without ballast, also called slab track systems. The slab track involves 28 an important economic saving in maintenance and a better mechanical behaviour, although it 29 requires a higher initial investment with lower constructive performances [1]. 30 There are comparative studies between the traditional track, formed by sleepers and ballast, 31 and the slab track, analysing economic aspects such as the reduction in maintenance costs or 32 the life cycle cost and technical aspects like weight reduction or higher lateral track resistance 33 [2,3]. In the present work, it is proposed that, when the superstructure is obsolete, a second life 34 should be given to any element of the superstructure that can be recycled. In addition, this 35 solution is a better solution from the mechanical point of view and therefore involves lower 36 maintenance costs. In this way, an economic benefit is generated, on the one hand, from the 37 saving in the purchase of the new materials and, on the other hand, from the savings in the 38 transport to the landfill of the removed material. The environmental benefits are derived also, 39 from the same points: both the reduction in the use of natural resources and the reduction of 40 the volume of waste generated after the withdrawal of the obsolete ballasted track are avoided. 41 The use of RA for the manufacture of concrete is relatively widespread nowadays, since it is used 42 in different concrete construction projects [4,5]. Although the properties of the recycled 43 aggregates (RA) must be analysed before dosing a concrete because these have some important 44 differences with respect to natural aggregates, such as the contaminants or higher absorption 45 coefficient [6,7], been these differences more important in the fine particles. Consequently, the 46 recycled concretes (RC) made with RA have some disadvantages such as a higher porosity, worse 47 frozen-thaw behaviour [8] or less compressive strength than the traditional concrete [9]. The 48 use of fine particles has also been studied and shows that the concretes with RA have a worse 49 behaviour to durability and worse mechanical properties [10,11]. For these reasons, the 50 standards impose some limitations on the use of coarse aggregates and ban the use of fine 51 recycled aggregates, for example, the EHE-08 or EN 13242 [12,13] or provide catalogues such as 52 the Construction and demolition waste catalogue [14], which classifies each waste, specifying 53 its possible uses, although these regulations are rather conservative. 54 With the aim of optimizing the construction process, the use of a self-compacting concrete is 55 proposed, which avoids the process of vibrating the concrete, improving the construction 56 performances modification that a part from the 200 ml which must be tested, also, 500, 750 and 1000 ml were 77 analysed in order to obtain a more robust and measurable results. 78

Recycled aggregates 79
To produce the recycled aggregate (RA), on one hand, out-of-use sleepers and, in the other 80 hand, out-of-use ballast aggregates were crushed and sieved. From these materials, six different 81 types of aggregates were obtained, Table 1. With these six aggregate fractions, three different 82 types of concretes were mixed. The first of them is made only with RA from crushed ballast (RC-83 B), the second with RA from crushed sleepers (RC-S) and the third with a mix between 84 aggregates from crushed ballast and aggregates from crushed sleepers in the track proportion 85 (RC-M). 86 The crushing process of the ballast and sleepers out-of-use for use as RA for the manufacture of 87 slab track, are presented in Figure 1 and Figure  To obtain RA from the ballast and from the out-of-use sleepers, the first step followed was to 94 find the out-of-service ballast and the sleepers. This material was given by the ADIF (Spanish 95 state-owned railway infrastructure manager under the responsibility of the Ministry of Public 96 Works and Transport) [25]. This procedure was done in the waste treatment plant VALORIA 97 RESIDUOS S.L. (Spain) using a portable jaw crusher which was able to separate the steel from 98 the concrete. Once the material was crushed, 3 sieves were used to separate each aggregate in 99 three different sizes: 0-2 mm, 2-5 mm and 5-15 mm, in order to ease the mix proportion of these 100 concretes. A sample of each aggregate size is shown in Figure 3. 101  These RA were subjected to mechanical and geometrical tests in order to be able to mix a self-106 compacting concrete. These tests were divided into two main groups, aggregate properties and 107 impurity characterization. 108 The measured aggregate properties were the densities according to EN  The main issue when RA are used is the presence of the impurities. In order to quantify these 113 impurities, a visual characterization was carried out. 1 kg of RA-BCA and RA-SCA were separated 114 according to the origin of each particle. At the end of the process, all the groups were weighted, 115 and the percentage of each material were calculated. 116 The main properties of the RA are shown in Table 2, and the grading curves are represented in 117 Figure 4. 118  is less than 1% wt. of impurities, Figure 5, and in the crushed sleepers, there are no impurities. 124 This is because the ballast aggregates had materials coming from the ground. Meanwhile, the 125 sleepers were taken one by one and, for this reason, they have no impurities. A ballast sample 126 was analysed and no hydrocarbons were detected. 127

Mix proportions 130
The mix proportions were defined in several sub steps. Firstly, the quantity of superplasticizer 131 additive, this process has been defined in 2.2. Secondly, the relation between the different sands 132 and the cement quantity were fixed, through 40 mortar tests per material, in which the mini-133 slump test and the compressive strength were determined. The final mix proportions are shown 134 in Table 3. 135 RC-B and RC-S were designed using 100% of incorporation of RA and the RC-M using a proportion 136 of 1/7 RA from sleeper and 6/7 RA form ballast). The mix proportions are shown in Table 3. 137  Table 4 shows the workability results. Figure 6 shows details of the workability of the concrete. 195 In the Figure 6, it can be seen that there are no segregation. On the one hand, in all the photos, 201 you can see that there are aggregates even in the border of the flow spread, and on the other 202 hand, no one presents a water/paste/mortar ring beyond the coarse aggregate. The RC-S has a 203 higher quantity of aggregates on the border of the flow spread, which is cause of the higher 204 quantity of aggregates and the higher viscosity of the mortar cause by the lower w/c ratio. Also, 205 it can be appreciated that the perimeter of the flow spread is much regular in the RC-B, its mean 206 a higher thixotropic behaviour reflected in a lower value in the V funnel test. 207 As has been demonstrated by other authors, it is possible to mix a self-compacting concrete 208 using RA [53]. While the increase in RA from crushed concrete normally reduces the workability 209 of the concretes [54], in this case, due to the geometry of the particles, observed in the flakiness 210 index, the RC-S is the concrete which obtains the best results in the slump flow and the L-box 211 test. 212 Once it was proved that the workability of the concretes was adequate, several tests specimens 214 were manufactured in order to analyse its hardened properties. The physical properties, 215 mechanical properties and durability behaviours were analysed. 216

Physical properties 217
The physical properties are shown in Table 5. 218   Table 6 show the results of the compressive strength test. 228 The most resistant is the RC-S one, due to the lower effective water/cement ratio of this dosage. 232

Mechanical properties 227
The RC-B obtained the lower results although the exceptional tribological properties of these 233 aggregates. The RC-M is between the RC-B and the RC-S as can be expected. All the mix 234 proportions had being always above the 37 MPa at the age of 28 days, which is the requirement 235 for

Wear resistance 247
The results are shown in Figure 7. 248 The permeability is a great parameter to measure the durability of a concrete because it is the 253 property responsible of the penetration of any agent that could damage the concrete. The 254 results of both the oxygen permeability and the water permeability are shown in Table 8. The  255 lower values of oxygen permeability are in the RC-S, and the higher ones are in the RC-B. It is 256 well known that the presence of RA does not really affect the permeability of the concrete [59], 257 and the lower water/cement ratio provides a lower permeability to RC-S. Anyway, it is well 258 known that the use of fly ash will help to reduce the permeability of a concrete [43], so all the 259 mix proportions obtained good results. These values of permeability are lower, so a good 260 durability behaviour can be expected. 261 The water penetration in the RC-B is the higher one, actually, higher than expected. The reason 265 of this high penetration is the presence of impurities and the higher effective water cement 266 ratio. 267 The variation of the length of the concrete is plotted in . Usually, the presence of aggregates from recycled concretes aggregates increases the shrinkage 276 [60], but as can be noted from the above chart, the RC-B samples suffer more strain due to the 277 higher water/cement ratio of these mixes [61]. 278 The damage suffered by the samples after 56 frozen-thaw cycles is shown in Figure 9. In addition, 279 as a quantitative parameter, the loss of mass is plotted in Figure 10, where it can be appreciated 280 that the RC-S are the more resistant to these kinds of cycles. The RC-S have a better frozen-thaw 281 behaviour due to the lower water/cement ratio [51]. 282  The damage suffered by the samples after 100 drying-wetting cycles can be appreciated in 287 Figure 11. Also, the evolution of the mass variation is plotted in Figure 12. At the end of the 100 288 cycles, the samples were tested and the loss in compressive strength was analysed. The results 289 are shown in Table 9. Visually, there are no big difference after these 100 cycles in any of the 290 samples, just a superficial deterioration. 291 Figure 11. Drying-wetting test evolution. 293 294 295 Figure 12. Drying-wetting test mass evolution. 296 The recycled aggregate from the crushing of ballast and sleepers meet the requirements for the 301 manufacture of structural concrete. Specifically, the evolution of the uniaxial resistance and the 302 elastic modulus as a function of time was analysed. It was found that in the three cases analysed, 303 mechanical properties were higher to the properties provided by the manufacturers of the main 304 types of track. 305 It has been possible to manufacture self-compacting concretes that meet the mechanical criteria 306 for the construction of slab track using, exclusively, recycled aggregate; that is, without the need 307 to add any type of non recycled aggregate. This is due to the use of a type IV cement. 308 From the above results, the possibility of manufacturing a concrete with low CO2 emissions is 309 demonstrated. 310 The 3 characterized concretes correctly fulfil the durability requirements. 311 The importance of adjusting the water/cement ratio in the concrete has been proven. This may 312 become more influential than the quality of the aggregate. This is clear when comparing the 313 results of both the durability and the mechanical properties of recycled concrete from crushed 314 ballast and recycled concrete from crushed sleepers. 315

Acknowledgments 316
The authors would like to thank: 317 The Spanish Ministry of Economy and Competitiveness for financing the project MAT2014-318 57544-R. 319 To the companies Adif, for the sleepers and the ballast out of use which are the starting point 320 of this research, and Cementos Alfa, who provided the Cement which have been use to 321 manufacture the concrete. 322 323 Biliography 324