Mechanical assessment of the induction heating as a method to accelerate the drying process of cold porous asphalt mixtures

NOTICE: this is the author’s version of a work that was accepted for publication in Construction and Building Materials. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Construction and Building Materials, 208, (2019) DOI: 10.1016/j.conbuildmat.2019.03.053

Author's post print: Lastra González, P., Indacoechea Vega, I., Calzada Pérez, M. A., Castro Fresno, D., & Vega Zamanillo, A. (2019). Mechanical assessment of the induction heating as a method to accelerate the drying process of cold porous asphalt mixtures. Construction and Building Materials, 208, 646 650. doi:10.1016/j.conbuildmat.2019.03.053 Cold mix asphalts present clear advantages such as the fact that they do not need to be heated, what results in lower energy consumptions and emissions, or the possibility to be transported long distances and manufactured on an offsite. However, their use is highly limited due to the long curing times that are needed to reach their final strength and the lower mechanical performance achieved comparing to hot mix asphalts. This paper studies induction heating as a process to accelerate the drying time of the emulsion and compares it, in terms of the mixture mechanical performance, with a more conventional method in which the cold sample is heated up in an oven. Different tests, as Cantabro, stiffness and Indirect tensile strength have been carried out. The mechanical results have shown that the induction heating could be a feasible alternative to increase the initial strength and reduce the opening time for this type of layers, although more research is necessary concerning the optimization of the mixture and the improvement of the induction device configuration.

Introduction
Cold mix asphalts have some clear environmental advantages because the aggregates and binder do water included in the emulsion and therefore to achieve their full strength 2 energy efficient technology mainly because the bitumen and metal particles within the asphalt mixture are heated 6 . This technology is currently assessed with reclaimed asphalt (RA) 7 , by products as alternative aggregates and heating inductors 8 , even as method to maintain the roads by melting binder pellets 9 . However, its use with cold asphalt mixtures has not been deeply analysed yet. In this paper, the induction heating is proposed as a potential method to selectively heat the emulsion and accelerate the evaporation of the water contained in a cold porous asphalt mixture, trying to decrease the required curing time of this type of mixtures.
not have to be heated up, so the energy consumption and the greenhouse gas emissions are greatly reduced 1 . However, these type of mixtures have also significant disadvantages, such as the lowest mechanical performance comparing to hot mix asphalts or the need of several weeks to evaporate the . In this sense, the porosity of the mixture is one of the main parameters affecting the drying process, because of its influence on the trapped water 3 .
Induction heating is a technology that has been explored in the last years to speed up the self healing of bituminous mixtures 4 . It consists on asphalt mixtures incorporating metallic particles in their composition that can be heated via induction. The bitumen around the particles is heated and due to the thermal expansion and viscosity reduction, flows and fills in the existing fissures 5 . This is a high

Materials
Apart for the metallic particles, conventional materials were used to design the cold porous asphalt mixture. Thus, limestone and ophite (porphyry igneous rock) were used as fine (including filler) and coarse aggregates, respectively. Their properties are presented in Table 1 and Table 2 In addition, a slow breaking cationic emulsion with 60 % of residual asphalt content was employed to design the porous asphalt, whose properties are included in Table 3. It should be highlighted that this emulsion does not contain any flammable diluent (i.e. kerosene), which could ignite due to the high temperatures reached during the induction heating process.  Finally, steel grit ( Figure 1) was added as the metallic particle to be heated by induction. Normally used for blasting, it is a granular material with 100 % broken surfaces and uniform gradation between 2 mm and 1 mm grain size. The particle size distribution was designed by volume of total aggregate due to the high density of the steel grit.

Methodology
Two porous asphalts were designed, one without metallic particles, used as a reference, and the other one incorporating the metallic particles; both with the same grading expressed in percentage by volume of the total aggregate. The samples of the reference mixture (without metal particles) were cured in an oven following a procedure divided in two steps 10,11 . In the first step, they were kept at 75°C for two days and as the samples did not present binder drainage, in a second step, they stayed at 90°C in the oven for other five additional days. On the other hand, the samples with the metal particles (experimental PA) were dried using the induction machine and varying the time and temperature of the heating process. As cold mixtures do not have enough consistency after compaction, the sample was contained in a silicon ring in which holes were made to allow the water vapour to get away ( Figure 2). It should be noted that the material used to keep the consistency of the mixture should withstand high temperatures without melting and that metallic meshes or grids should be avoided because they are heated by the magnetic field and this could affect the test. Initially, the high temperatures reached when heating the experimental PA by induction produced the drainage of the residual binder ( Figure 3). In order to solve the problem, both mixes (reference and experimental) were redesigned increasing the percentage of fine aggregate and filler. For the re designed mixtures, the induction heating was applied for 30 minutes at different intensities. The heating process was carried out in two phases. Firstly, in order to reach an average temperature of the samples of 120°C, 300 A intensity was applied for approximately 15 20 minutes. In the second phase, and in order to keep the temperature in the range of 120 130°C, the intensity was reduced to 200 A until the end of the test (around 10 15 minutes). No drainage of the bitumen was observed this time. Figure 4 presents the configuration of the induction machine.

Design of a cold PA mixture
The aggregate grading of the experimental and reference mixes designed are shown in Figure 5. As it can be observed, the grading was adjusted to the higher limit to avoid drainage problems during the testing. This is because the main objective of this preliminary research was evaluating the technical feasibility of using induction heating to reduce the time for evaporating the water in cold mixes. In a future continuation of this research, the time and temperature parameters for the induction heating

Results and discussion
The results were analysed with the software Minitab. When the data fulfilled a normal distribution and there was homogeneity of variances the Student t test was performed. Otherwise, the U of Mann-Whitney test was used. The confidence interval was always 95 %, so a statistical significance of 0.05 states the threshold level of acceptance or rejection.
Following, the main results are presented.
need be optimized to avoid bitumen drainage.
Regarding the design of the mixtures, in the case of the experimental PA, the only difference compared to the reference mixture was the addition of 1% of steel grit by volume of total aggregate. The percentage of bitumen was kept the same in both mixes (4.5% by weight of mixture).

4.2.Mechanical performance of cold PA mixture
To evaluate the feasibility of the induction heating technology for this application, different tests have been carried out and their results compared with the oven heated samples.
Thus, in the first place, the effect of the heating procedure on the bitumen properties was analysed. To do so, two samples of bitumen were extracted for both the experimental and reference mixes after the heating process (oven and induction) and the penetration and softening point of the recovered samples were analysed (Table 4). In both cases, the bitumen is clearly aged, but the aging is more pronounced when the sample is oven heated.  Afterwards, the density and void content of both mixes were compared (Table 5). Despite the different heating method used, the density and void content are practically the same and the difference was found to be insignificant (see p values in Table 8). However, in both mixes, experimental and reference, the void content was quite high compared to traditional PA mixes (around 20%).

Reference Experimental
Density (g/cm 3  Concerning the Cantabro test, it was performed at 25°C by measuring the loss of particles occurred every 50 drum revolutions until reaching 300 revolutions ( Figure 6). Figure 7 shows the particle loss ratio every 50 drum revolutions. Important differences were found concerning the particles loss of the experimental and reference mixtures after induction and oven heating, respectively. The particle loss was higher in the case of the mixture heated in the oven. Since the percentage of voids was similar in both mixtures, the differences could be related to, as observed before, the higher aging level of the bitumen of the reference mixture compared to the bitumen of the experimental mixture. In addition, the main loss in the fragmentation resistance for the reference mixture occurred in the first 100 drum revolutions, so likely the experimental mixture will present a higher service life. On the other hand, despite the great differences observed between the mixes, the results were not statistically significant due to this test showed a high variability ( Table 8).
The indirect tensile strength (I.T.S) was calculated according to EN 12697 12 with dry samples. Based on the results (Table 6), the reference mixture showed a higher strength, probably also related to the higher stiffness of the binder. However, in consonance with the statistical analysis, the differences were not significant (Table 8).
heating, even though the latter reached higher temperatures.  (Table 7), although, only in this case, the result was statistically different (Table 8). It should be highlighted that as these mixtures are destined to surface layers and their function is not bearing loads, the difference found in the stiffness is not critical.

Conclusions
Two cold PA mixtures have been designed using the same percentage of residual binder and a similar grading, with the only difference of the addition of 1% steel grit by volume of aggregate in one of them. The evaporation of the water included in the emulsion has been carried out using two different methods: the traditional method used to design cold mixes at the laboratory consisting in heating the mixture in an oven and a new innovative method consisting in heating the mixture via induction. In order to evaluate the feasibility of the latter, once the water is evaporated, the mixes were subjected to different mechanical tests. The conclusions can be summarised as follows: The oven heating method seems to age the residual binder of the emulsion more than the induction In terms of mechanical performance, the mixture heated via induction presented, on the one hand, a better behaviour on the Cantabro test with a less particle loss and, on the other hand, a lower stiffness. Based on the statistical analysis, only the latter result was significant.
The induction heating seems a feasible technology to shorten the time needed for curing the emulsion of the cold porous asphalt mixture. Although preliminary, the results are promising since similar mechanical performance has been obtained reducing the time to evaporate the water from 7 days to 30 minutes.