Bituminous concrete mixes were prepared to fabricate 30 specimens, divided into two categories of test specimens, were prepared, using two different types of geogrids and two types of emulsions different percentages. All these materials were chosen according to the test conditions and the equipment and instrumentation available in accordance with the standards in force (19).

The experimental study consists of evaluating, the flexural strength and the modulus of rupture, using seven different prismatic beams (305×90×75 mm): standard beam (control), beam with emulsion1 and beam with emulsion2, beam reinforced by geogrid1 with emulsion1, beam reinforced by geogrid1 with emulsion2, beam reinforced by geogrid2 with emulsion1 and beam reinforced by geogrid2 with emulsion2. For each test, two specimens were tested and the average values strength and modulus of ruptura are reported. The deflection was measured with a linear variable displacement transducer (LVDT), and the load-displacement curves were plotted.

Three point vending tests were also conducted on eight pre-cracked asphalt concrete slabs of 500×180×100 mm, with the insertion of emulsions and geogrids: standard slab not pre-cracked (reference), pre-cracked reference slab, pre-cracked slab with emulsion1, pre-cracked slab with emulsion2, pre-cracked slab with emulsion1 and geogrid1, pre-cracked slab with emulsion2 and geogrid1, pre-cracked slab with emulsion2 and geogrid1 and pre-cracked slab with emulsion2 and geogrid2. The objective of this second serie of tests is to evaluate the crack propagation on the behaviour of each emulsion and geogrid reinforced bituminous slab (30, 31).

A total of 30 (thirty) specimens were made with the same composition of bituminous concrete. These specimens are divided into two categories: the first category, with a number of 14 (fourteen) prismatic beams in the form of rectangular shaped slabs (305×90×75) mm, composed of two layers of asphalt concrete, cationic emulsion (E1 or E2) and geogrids (G1 or G2). Table 2 shows the details and identification of the manufactured slabs tested under three-point bending test with a span of 240 mm where the flexural strength and modulus of ruptura were determined. Deflection was measured using a linear variable displacement transducer (LVDT), in accordance with ASTM E2309 (32). All tests were were conducted at 20°C, under displacement control at a constant speed of 50.8 mm/min.

The slabs of this first category represent a simulation of the repair of a deteriorated asphalt layer and recharged by a top layer with emulsion bonding and geogrid insertion. In effect, the lower layer simulated an existing deteriorated asphalt layer and the upper layer represented the reinforcement layer (overlay). The first step in the manufacture of the slabs was to produce a 50 mm high asphalt concrete layer in an appropriately sized mould and to compact it using the rolling compaction procedure in accordance with NF-EN 12697-33 (33).

The second category is composed of 16 pre-cracked slab specimens of dimensions (500×180×100 mm), made with a concrete-asphalt mixture, prepared in accordance with European standards (EN 12697-35) (34). These slabs are composed of layers of asphalt concrete, cationic emulsion (E1 or E2) and geogrids (G1 or G2). Table 3 shows the details and identification of these slabs.

The calculated amount of mixture was poured into preheated moulds of dimensions (500×180×100 mm) and compacted in the laboratory using the roller plate compactor in accordance with the above-mentioned standard (EN 12697-33).

The production of the double-layer slabs was carried out in several stages. In the first stage, the bottom layer was compacted to a thickness of 50 mm, which resulted in an air void content of approximately 5%. After cooling the slab for 4 hours in the laboratory, the bituminous emulsion was spread at the interface and left for 1 hour in the air to evaporate. Then the geogrid was carefully placed and a 50 mm top bituminous layer was applied and compacted (Figure 4a). Compaction was applied until an air void content of approximately 5% was achieved. The determination of the latter was done according to NF EN 12697-8 (35). Finally, the direction of compaction was marked on the surface of the slab in order to carry out the tests in the correct direction of traffic.

In order to simulate a flexible pavement (road or airfield), with matching cracks, a 40 mm notch was made at the base of the manufactured slab (10 mm below the geogrid) by sawing to impose the location of the crack initiation (Figure 4a). Thus, during the test, the crack propagation starts at the location of the pre-crack. In order to perfectly visualise the crack path, a layer of plaster was spread over the central area of the samples. The objective of this second category of specimens is to study the effect of crack propagation on the behaviour of a slab reinforced (Figure 4b), simultaneously, with a geogrid layer and the type of emulsion chosen (36).

This is a non-destructive test device designed to reproduce, by means of an impact on a disc in contact with the road surface, the load corresponding to half an axle of a truck travelling at approximately 80 km/h and to measure, at the same time, the deflections generated on the surface. This test offers the possibility to vary the intensity of the applied load according to the structural stiffness observed in situ (37). HWD loads, designed for roadways and airports, are generally between 20 and 75 kN. This device could be used to make relative comparisons of pavements based on deflection indices and to determine the structural capacity of pavements and the modulus of elasticity of material layers by back-calculation. HWD can also be used to calculate deformations and stresses in pavements, detection of voids under slabs and as a tool for pavement quality control during construction.

HWD is used for deflection testing on flexible or rigid pavements on airport runways using additional loads where the total load can be up to 250 kN (38).

The case study concerns the main runway (02/20) of an airfield that was commissioned in 1951 and consists of two runways of 3000 × 45 m each. Its pavement is composed of several sub-layers (Figure 5) and this infrastructure has been the subject of several reinforcement and modernisation works since it was commissioned. This runway is operated by various aircraft and the critical aircraft criticized for the study is a Boeing 737-800 (39).

The procedure of reinforcement by insertion of the emulsion layer and geogrid sheet was as follows:

The geogrid used for the reinforcement of the main runway 02/20 and ancillary structures is a non-woven geocomposite composed of continuous polypropylene filaments combined with glass fiber cables, known as PGM-G 50x50 (Figure 6). This non-woven geogrid composed of continuous glass filaments and polypropylene fibers type of geogrid was used because of its superior properties. These geogrids were different than those used in the laboratory tests because of their unavailability from the manufacturer during the laboratory tests (40).

Deflection measurements in the longitudinal direction were carried out by preforming six measurement profiles located at 3.5 m, 6.5 m and 12 m on either side of the main runway axis (Figure 7). The measurements were carried out before the reinforcement of the runway with geogrids (in 2009) and then after the reinforcement, i.e. seven years later (in 2017), in order to monitor its behaviour over time. A comparison between the two measurements is made to assess the effect of the geogrid reinforcement on the values of the deflections and stresses of the running surface (bituminous concrete). The comparison between the state of the runway, before and after its reinforcement, allows the contribution of the asphalt layers reinforced by the geogrid to the bearing capacity of the lower layers to be estimated.

The first results obtained during the experimental investigation, in the laboratory, are presented in Figure 8, observing the evolution of the flexural strength. These are -the load-displacement curves of the different prismatic beams tested are given in Figure 8. The maximum load is determined at full crack propagation, prior to total failure of the specimen. In all the tests carried out, the specimens failed ductilely, without tearing or disintegration of the applied geogrid. For the control beam, referred to as “R”, the maximum load value was 15.37 kN, for a mid-span displacement of 0.223 mm. The use of the emulsion layer type 2 (E2) is more effective than the one used for the emulsion E1, due to the higher percentage of 69% compared to 65%. The maximum load value is 16.83 kN (gain of 9.50%) for E2, compared to 16.27 kN (gain of 5.86%) for E1. The use of an emulsion layer seems to give a non-negligible gain in flexural strength (41). The insertion of the geogrid, with the emulsion layer at the interface, avoids the provocation of the disbonding effect between the layers (42) and ensures an appreciable cohesion and gain for the pavement (the beam in our case). Type 1 of the geogrid with E1 emulsion (E1G1 beam), gives a value of 21.17 kN (a gain of 37.74%), compared to the value of 22.38 (G1 geogrid + E2 emulsion), therefore a gain of 45.61%. The E2 emulsion, with the G1 geogrid (E2G1), gave a value of 22.53 kN (gain of 46.58%), compared to the value of 23.08 kN (a gain of 50.16%). These results show that it is more favorable to apply geogrid type 2 (or any equivalent type), with the tack coat (E2 emulsion) for any reinforcement as the gain is more than 50%.

For the deflections (displacements at mid-span of the tested beam), the above curves shows that the maximum displacement of the control beam (R) is of the order of 0.353 mm, whereas for the G2 reinforcement and E1 emulsion, it is of the order of 0.415 mm, i.e. an increase of 19.26%, opposing only 0.415 mm (gain of 17.56%) for the E2G2 beam and 0.398 mm (gain of 12.75%), for the E2G1 beam. Concerning the emulsions applied alone at the interface, which ensure good bonding (adhesion), reductions in deflections were observed (0.270 mm for E1 and 0.245 mm for E2), i.e. significant gains of the order of 23.51% (emulsion E1) and 30.59% for emulsion E2 respectively. These tests also showed the importance of the emulsion for the bonding of asphalt concrete layers, where the mechanical performance is improved. The gain in mechanical performance is more important when geogrids are used as reinforcement by the use of emulsion, where the gain can reach 50% for the applied force and 31% for the displacements. Consequently, the increase in mechanical performance is closely related to the increase in adhesion between the two bituminous layers and between the bitumen and the geogrid. The comparison between the two types of geogrids used shows that the type 1 geogrid gave lower performance due to its insufficient resistance to lateral movements. For the evolution of the modulus of rupture (MOR), which is defined as the maximum stress that any rectangular prismatic beam can withstand when subjected to bending and which allows the evaluation of progressive damage, generally related to cracking (43), Table 4, gives the values found during the different tests of the specimens used. These results shows that the insertion of the geogrid sheet improves the modulus of rupture (MOR) by 8.74% for the beam (E2G2), 7.32% for the beam (E1G2), 4.47% for the beam (E2G1) and 4.07 for the beam (E1G1). The beams with emulsions E1 or E2, do not seem to give any gain in modulus of rupture (MOR), which proves that the emulsions, in spite of the adhesion they provided, do not influence the stress at rupture.

The three points bending test was chosen to evaluate the perfromance of the pre-cracked slabs because it produces the maximum bending moment at the middle of the slabs. Figure 9 shows the different load-displacement curves obtained for the slabs tested.

The load-deflection curves show that for each slab tested, the maximum load (Pmax) is the failure one, with a clear advantage for geogrid-reinforced slabs bonded with emulsion layers. The area under each curve is composed of two zones: the zone under the curve up to the flexural strength (Pmax), which represents the crack initiation energy (Ei), while the zone under the curve from Pmax to the failure of the specimen is the pre-crack propagation energy (called Eup) and the curve between Pmax and the failure of the slab, which is the crack propagation energy (called Ep). It is noted that no cracks occurred or appeared (neither ordinary, nor of contiuation of the pre-crack) in the specimens, before the maximum load Pmax (44).

As all our specimens (slabs) are made of two layers, the energy (Ep) is the sum of the energy necessary for propagation in the lower layer (Einf) and the energy necessary for propagation in the upper layer (Esup). The values of Ei and Ep are calculated by the areas under each curve by discriminating the curves. It was observed that the cracks started to propagate from the notch tip (pre-crack), location of the stress’s highest concentration and then upwards in the direction of the applied load, perpendicular to the maximum principal tension (45). It has been reported that in the case of reinforced pavements, the reinforcement mainly affects the crack propagation in the top layer (asphalt concrete wearing course) (46). Based on the results obtained, it can be seen from Figure 9, that the unreinforced and non-pre-cracked slab (RR) has a higher value of flexural strength (Pmax) (27.45 kN) than the unreinforced and pre-cracked slab (PRR) (24 kN) and this is due to the higher failure energy, as a result of not facilitating crack initiation (no presence of notch). The introduction of the emulsion layers ensured an adequate bond between the interface of the slab and the double layer (47). Values of the order of 28.5 kN for the pre-cracked slab with emulsion E1 (PE1) and 29.2 kN for the pre-cracked slab with emulsion E2 (PE2) were obtained, confirming the absence of disabonding at the interface.

The insertion of the geogrid sheet seems to give a very appreciable gain, with Pmax values for the slabs reinforced respectively by the type 1 geogrid and the type 2 geogrid, namely 30.2 kN (slab PEIGI), 31.4 kN (slab PE2G1), against the values of 32.5 kN (slab) PE1G2 and 33.4 kN (slab PE2G2 . The gain values are 21.70% (slab PE2G2), 18.40% (PE1G2 slab), 14.39% (slab PE2G1), 10.02% (slab PE1G1). The emulsions gave insignificant gains for Pmax, with a value of 6.38% (slab PE2) and 3.83% (slab PE1). The quantification of the contribution of the geogrid in the crack propagation phase is given, first, by the value of the crack propagation energy (EP), summing its value before and after the failure phase (Table 5). The coefficient of performance (k), defined as the ratio of the propagation energy of the unreinforced and pre-cracked slab (PRR) after failure, to the propagation energy of the reinforced and pre-cracked slab, is detailed in the Table below.

The results found, shows that the coefficient of performance (k), for the specimens tested in the laboratory and calculated for maximum deflection at mid-span, is influenced by the mode and type of reinforcement. It can be observed that, in addition to the bonding emulsions, all geogrids significantly increase the crack propagation (coefficient k) and significantly increase the crack propagation energy (Eup). Indeed, if the emulsions increase by almost 83% (E1 emulsion) and 143% (E2 emulsion), the geogrids surpass these Figures by far, with a percentage varying from 149% (PE1G1 slab), 157% (PE2G1 slab), 173% (PE1G2 slab) and 237% (PE2G2 slab).

These results show that the geogrid reinforcement (G2) with emulsion bonding (E2) performs best and delays the crack propagation in the notch above the reinforcement. The coefficient of performance (k) varies from a value of 2.26 (E1 emulsion) to a value of 4.16 (G2 reinforced slab and E2 emulsion), i.e. a difference of 84%, clearly indicates, that the service life of the reinforced asphalt pavement increases, by delaying the upward cracking and the reflection crack.