Rehabilitation and reinforcement of cracked pavements using crack sealers and composite patches

Abstract

Cracking is a significant concern for pavements and should be appropriately treated during road, highway, and runway rehabilitation. This study investigates the behavior of asphaltic materials under tensile and shear loading modes in intact, fractured, and repaired conditions. With this aim, several methods and materials are utilized for repairs, such as poring adhesive into the crack (using bitumen, neat epoxy resin, and polymer concrete adhesives) and patching the crack with textile (by glass fiber and epoxy resin or bitumen). These tests were conducted at +10 °C, with a three-point bending loading configuration, the same as the actual loading configuration of pavements. Criteria such as failure load, failure work, and post-failure work, as well as failure patterns, were assessed to assess the effectiveness of repairs. Numerical analysis was also performed, and a constitutive model was presented. The ultimate tensile capacity of the cracked specimen is measured at 63 % lower than the intact condition (778 N). The ultimate tensile load of the bitumen-repaired specimen is higher than that of the cracked specimen, but it is still 11 % lower than that of the intact condition. The ultimate tensile capacity of epoxy resin repaired and polymer concrete repaired specimens are 88 % and 79 % higher than the intact specimen (about 1400 N). The ultimate tensile load of the fabric patch reinforced specimen that used bitumen as the adhesive is 38 % higher than the intact specimen (1075 N), while for the case of using the epoxy resin adhesive, this value is 258 % (2788 N). Observations of tensile failure patterns show that, because bitumen is viscoelastic, failure in bitumen-repaired specimens happens in bitumen necking mode and starts at the repaired crack tip. In other cases, the failure occurred far from the pre-crack plane.

1. Introduction

Asphalt is among the common materials for paving roads and runways. With proper design, construction, and maintenance, pavements can last many years. However, like other materials, pavements are susceptible to various forms of damage [1,2]. Cracking is a typical form of asphalt damage. Cracks can form due to mechanical causes like overloading, impact, wear, or fatigue, and environmental causes like freezing and thawing, water seepage, or ground movement [[3], [4], [5]].

The crack configuration depends on the pavement deformation mode. Assuming a section of pavement as a plate, stresses act in three main modes: tensile, in-plane shear, and out-of-plane shear, which appear as opening, sliding, and tearing deformations. As an example, in pavements, an even traffic load distribution generates opening mode, an uneven traffic load distribution generates in-plane shear stress (slide), and an opposite-direction traffic load generates out-of-plane shear stress (tear) [[6], [7], [8], [9], [10]]. Although longitudinal and transverse cracks are two important cracks that mainly emerge in the pavements due to tensile stresses, edge cracking is typical damage due to in-plane shear stresses [[11], [12], [13]]. Fig. 1 illustrates traffic load, loading condition, and related deformation based on cracking mode.

Fig. 1.

Mode of failure in pavements.

The process of repairing and improving pavements is commonly known as rehabilitation. According to the American society of Civil Engineers, over fifty percent of roadways in the United States of America need rehabilitation. Driving on these roads annually costs each driver about $1000 in wasted time and fuel. On the other side, it is estimated that each dollar invested in road, highway, and bridge rehabilitation returns around five dollars in the form of lower vehicle maintenance costs, reduced delays, lower fuel consumption, increased safety, and reduced emissions [14]. In addition to roadways, asphaltic pavements are used in most runways across the world, however, they are susceptible to damage from severe weights during landings and takeoffs. Airport pavements are constructed stringently to satisfy criteria for passenger and personnel safety. However, compared to conventional highways, runways require more repair to meet performance and safety standards.

A variety of pavement rehabilitation techniques are being tested all over the world. The extent of the damage and the pavement's performance restrictions usually determine which method is the best. For small amounts of damage, it is routine to pour a binder (often bitumen or emulsion) into the crack, whereas, for extensive damage (or excessive amounts of damage), removal of the pavement and reconstruction is required. In the restoration project for the Meridian Regional Airport runway (a key field in Mississippi, United States), the asphaltic runway was milled after 20 years of service. Knowing the existence of cracks in deeper layers and to fix potential cracks that may return to the surface, in addition to a new surface, a stress relief layer was constructed. In this project, in addition to the challenging schedule of working between airport operating hours, the rehabilitation program was interrupted by the weather during long rainy hours.

Rebuilding asphalt pavement is the most fundamental technique for repairing damaged pavement. However, previous projects have demonstrated that removing and reconstructing the pavement is expensive, and in many situations, it is nearly impossible to stop or limit passing traffic. In addition, this approach raises considerable concerns because it generates large quantities of environmentally hazardous waste with limited recycling potential. Worryingly, the transportation sector consumes around 25 % of fossil fuel globally, and approximately 7 % of this is attributed to the pavement industry [15,16]. On this basis, it is unquestionable that the approach for pavement repair should be changed to one that is quicker and cheaper; however, this method must be efficient enough.

Looking at previous projects shows that additional techniques can be utilized for pavement rehabilitation, each with its advantages and disadvantages. Pouring a binder into the crack is the most typical rehabilitation procedure for fast repair of limited cracks; however, choosing an appropriate binder is critical to its efficacy [17]. Bitumen is the most used material for pouring in cracks. Although it behaves similarly to asphalt, its application is difficult and time-consuming, and the temperature rise significantly reduces its mechanical performance. Asphalt emulsions are an alternative substance for filling pavement cracks; their application is quicker and more straightforward than bitumen, but they have disadvantages, such as poor mechanical performance and lack of durability [18].

To counter these shortcomings, superior adhesives, such as epoxy resin, have lately been used. It is simple to execute and possesses excellent mechanical and durability features. However, the high cost of pure epoxy resin adhesive makes resin epoxy-based materials such as polymer concrete more desirable (which can be prepared with 20 percent pure epoxy resin adhesive and 80 percent aggregates). Also, some other methods, such as using adhesive-contained capsules (mainly epoxy resin) [[19], [20], [21]] or heat healing method [22], have been studied by accepting disadvantages such as low mechanical efficiency, high costs, and difficulty of execution.

The process of asphalt crack repair is straightforward. After the crack occurrence, the crack gets widened using a rotary cutting blade to ensure proper filling; then the crack fills with adhesive (although in some cases, such as wide cracks or when the time is limited, the widening process may be ignored) [23].

Using the experience from similar issues, the repair of crack damage in concrete pavements and components is studied more comprehensively [[24], [25], [26]]. In these cases, repairing using adhesives is also a common method of rehabilitation. In these cases, cracks were repaired by injecting adhesive into them (mainly epoxy resin) [[27], [28], [29]], or for more strength, fabric patches were applied (primarily carbon or glass fabrics) [[30], [31], [32]]. Using fiber patches and fabrics is an excellent strength in the presence of excessive loads and impacts. This method uses a binder to attach a fabric on a surface (or embed it between layers). A composite system was developed during this method, and the properties of used materials (substrate, binder, and fabric) dictated its behavior. It is worth mentioning that a similar concept also stabilizes the soil embankments, where a fabric such as geogrids was used to increase the tensile or shear strength of the soil.

The use of fabrics in the pavement industry is new. In 2020, at Pula Airport (Croatia), a geosynthetic composite reinforcement was installed to rehabilitate the asphalt pavement of the runways. The aim was to improve the elastic modulus for better load distribution against wheel rutting and reduce reflective cracks for better long-term performance. The chosen composite reinforcement was a double-layer glass geogrid composite, featured by a bitumen coating and a nonwoven polypropylene substrate impregnated by bitumen. Alternatively, in another project, Chandler Regional Airport runway rehabilitating, large cracks were saw cut, cleaned, and filled with Portland cement concrete, and a mastic sealer was used for narrower cracks (1–3 inches). Then, some low-elongation paving fabric was installed on some cracks, and finally, a 3-inch overlay was executed on the whole runway. Also, the same method was used to rehabilitate the Stillwater Regional Airport runway.

Literature review shows that few studies have been conducted on repairing cracked materials and structures. Most studies focused on strengthening the structural components (i.e., beam or column) made of cement concretes, or some other studies propose using adhesive-filled capsules (to fill the crack after the damage). However, assessing the mechanical behavior of damaged and repaired materials is rare. Some of the comprehensive research in this field of study was done by Issa and Debs [33], who studied the effect of concrete crack repair using resin epoxy. In their research, three concrete cubes with no cracks, six cracked without repair, and six cracked bonded with gravity-filled epoxy were crushed, and their compressive strengths were obtained. It was found that the cracks caused a reduction in compressive strength up to 41 %, whereas the epoxy system restored the compressive strength by decreasing the reduction to 8.23 %.

Han et al. [34] studied the effect of repairing on the behavior of beam-shaped specimens made of sandstone and granite rocks. They used three scenarios: intact and two cracks repaired conditions (epoxy resin repair and epoxy resin along with a resin pre-coating process) and a three-point loading fixture for bending tests. Results show that for sandstone and granite, the load-bearing capacity of specimens in the repaired condition is improved by about 12 and 37 percent compared to intact conditions. The crack path observations show that the initial crack in intact material is vertical (parallel to the loading plane), while the crack propagation in the repaired specimen occurs adjacent to the repaired zone.

Ekenel and Myers [35] investigate the durability performance of rebar-reinforced concrete beams strengthened with epoxy injection, and carbon-fiber reinforced polymer (CFRP) fabrics. They built twenty-three beam-shaped specimens to evaluate their behavior with and without epoxy resin injection and carbon-fiber reinforced polymer (CFRP) reinforcement. To prepare the specimens, they were first loaded to 40 % of their ultimate moment capacity. Then, after sealing the crack, epoxy resin was injected into the crack; in the following, the CFRP process was conducted. The result shows that the average ultimate load capacity of intact beams is about 55 KN, the cracked beam is about 22 KN, the crack injected is about 27 KN, and the crack injected plus CFRP is about 44 KN. However, the initial slope of load-displacement curves exhibited significant differences between test samples. So, the epoxy-injected samples exhibited a slope value of 1.25 times and 3.5 times higher than the cracked samples. The crack in epoxy-injected samples did not show a visual opening during loading, and two new cracks formed next to the injected zone. This could be explained by the bond-strength-to-concrete of the low-viscosity injection material (between 3.4 and 4.1 MPa), which was higher than the tensile strength of concrete (roughly estimated between 2.1 and 2.9 MPa); hence, new cracks were formed next to the injected ones at weakened locations.

Karimi et al. [36] investigate the efficiency of using bitumen and polymer concrete to repair cracked asphalt pavement at temperatures ranging from −20 °C to 30 °C. Their study uses semi-circular bend (SCB) tests to evaluate the behavior of intact, cracked, and repaired samples under tensile loading. The study finds that failure patterns differ significantly between the two materials: bitumen repairs fail at the repair zone, while polymer concrete repairs fail outside the repair zone. Bitumen is found to be effective only at sub-zero temperatures, while polymer concrete consistently outperforms bitumen, providing higher strength and efficiency across all tested temperatures.

Wan et al. [37] present an innovative combined healing system for asphalt concrete using calcium alginate/Fe³O⁴ capsules, which respond to cyclic loading and microwave irradiation. The study characterizes the capsules' morphology, mechanical performance, and thermal resistance. Results show a mechanical strength of 11.8 N and a mass loss of 3.8 % at 200 °C. The capsules released more rejuvenator with increased loading cycles, achieving a 91.7 % healing level after 64,000 loading cycles and 30 s of microwave irradiation. The rejuvenator softened the asphalt binder, enhancing its flowability. This dual-responsive system offers a sustainable, low-carbon pavement maintenance strategy by effectively restoring strength and improving the rheological properties of asphalt concrete under varying service conditions.

Awuah and Garcia-Hernández [38] introduce a novel machine process for repairing asphalt pavement cracks using hot bitumen, employing a 3D printer for its simplicity. The study evaluates how filling speed, temperature, bitumen type, crack width, crack irregularity, and bitumen flow impact the quality of crack repair. Key quality metrics include porosity, shear strength, and tensile strength of the repaired cracks. Findings indicate that precise control of bitumen flow, filling speed, and crack dimensions is crucial for optimal repair quality. The research highlights the need for further investigation into the interactions among these parameters to develop effective autonomous crack-filling devices.

Chen and Wang [39] investigated an innovative asphalt pavement pothole repair method using recycled asphalt pavement (RAP) and preheating. Their study evaluated the mechanical performance and environmental impact of this method, focusing on moisture resistance, bonding strength, and abrasion resistance under various conditions. Laboratory tests showed that 30 % RAP content provided satisfactory moisture resistance and durability, making it suitable for field applications. Preheating was found to enhance the durability of the repaired patches. The study concluded that using RAP and preheating can match or exceed the performance of traditional hot-mix asphalt while reducing energy use and emissions.

Zhang et al. [40] developed a hybrid toughened unsaturated polyester resin (RMUP) for asphalt crack repair, utilizing liquid nitrile rubber (LNR) and nano-montmorillonite (nano-MMT). The combination improves the material's ductility and maintains high mechanical strength. The RMUP formula showed superior workability, quick strength gain, and temperature resistance, with durability in water immersion and extreme heat. Its tensile strength reaches up to 19 MPa, performing similarly to epoxy sealants but at half the cost. FT-IR analysis reveals that LNR cross-links with UPR, while nano-MMT forms an intercalated nanocomposite, enhancing toughness.

1.1. Aims and scopes

Rehabilitation projects of roads and runways are costly, time-consuming, and environmentally dangerous; except for shutting down roads, highways, or airports, the generated bituminous wastes can pollute the environment [41]. Rehabilitation can be accomplished using various techniques and materials, such as removing and reconstructing the pavement, reinforcing the pavement structure with textiles, or sealing the cracks with adhesives and sealants. Literature review reveals that studies on the rehabilitation effectiveness of pavements are scarce, particularly for pavements strengthened with textiles or repaired by pouring adhesive into cracks.

This study's primary objective is to evaluate the efficacy of pavement rehabilitation through a mechanical testing program of intact, cracked, and repaired specimens. In order to achieve this, after specimens were prepared, asphalt's behavior was first assessed in both normal and damaged conditions, then repaired using various techniques. Initially, the crack was fixed with three adhesives, including bitumen, neat epoxy resin, and epoxy resin-based polymer concrete. Then, the pavement is reinforced with fabric patches (woven glass fiber), and applied with two adhesives (epoxy resin and bitumen). These scenarios result in 14 test cases (each test with three repeats). At an ambient temperature of +10 °C, the behavior of specimens under tensile and shear loading conditions was investigated. Hot-mixed asphalt (HMA) was used as the most popular pavement material in the rehabilitation procedure [[42], [43], [44]]. The strength characteristics and performance, including ultimate loads, energies, and failure patterns, are compared and discussed based on the results. A series of numerical models were studied to provide a deeper insight, and a constitutive model was developed to represent the anticipated behavior of repaired components.

2. Materials

Several materials are employed to prepare and repair the samples, including HMA for making the base specimens, bitumen, epoxy resin adhesive, silica aggregate for preparing polymer concrete, and woven glass fiber fabric for repairing the crack.

This investigation used crushed limestone aggregate to prepare HMA material, and the ASTM D-3515 standard recommended granulation was utilized. Virgin aggregates were obtained from Asbcheran quarries in Tehran, Iran, and separated as fine aggregates (particle sizes of 0–6 mm) and coarse aggregates (particle sizes of 6–19 mm). Table 1 represents the aggregate characteristics provided by the supplier's datasheet. Aggregates were of a dolomitic limestone type with some quartz [45]. The aggregate gradation was selected according to AASHTO M323 with a nominal maximum aggregate size of 12.5 mm.

Table 1.

The characteristics and properties of aggregates.

Property	Value	Requirement	Test standard
Specific gravity of coarse aggregate	2.641	–	ASTM C127
Specific gravity of fine aggregate	2.623	–	ASTM C128
Coarse aggregate angularity (%)	100/100	Min. 95/90	ASTM D5821
Fine aggregate angularity (%)	52	Min. 45	ASTM D1252
Sand equivalent (%)	68	Min. 45	ASTM D2419
Los Angeles abrasion (%)	23	Max. 30	ASTM C131
Flat and elongated particles (%)	0.1	Max. 10	ASTM D4791
Sodium sulfate soundness (%)	0.2	Max. 15	ASTM C88

The used bitumen has a penetration grade of 60/70 and was produced by Iran's Pasargad oil refinery. Table 2 shows their characteristics based on the manufacturer data sheet. It is a common bitumen used globally for general weather and traffic conditions [46,47]. According to the ASTM D1559 standard, the optimal bitumen and air percentages for the prepared HMA mixture are 4.2 % and 5.1 %, respectively.

Table 2.

The characteristics and properties of binders.

Property	Value	Requirement	Test standard
Penetration at 25 °C (0.1 mm)	66	60–70	ASTM D5
Softening point (°C)	49.4	Min. 46	ASTM D36
Flashpoint (°C)	334	Min. 232	ASTM D92
Specific gravity at 25 °C	1.017	1010–1060	ASTM D3289
Ductility at 25 °C (cm)	100+	Min. 100	ASTM D113
Viscosity at 135 °C (cSt)	327	–	ASTM D2170
Change of mass (%)	0.003	Max. 0.8	ASTM D6

The current research used aliphatic epoxy resin with HA-11 polyamine hardener (ML506 commercial code, Mokarrar Company, Iran). Due to low viscosity, this resin is suitable for crack repair and mixing with silica aggregates. Table 3 presents the mechanical properties of the resin, as reported by the manufacturer. The Kavyan stone company in Iran supplied the silica aggregates. Due to the narrowness of cracks, the maximum particle size of silica aggregates was chosen as 1.2 mm (#16 sieve).

Table 3.

Mechanical properties of used epoxy resin.

Property	Value/unit	ASTM Standard
Compressive strength	97.4 MPa	D695M
Compressive modulus	937 MPa	D695M
Flexural strength	96.0 MPa	D790M
Flexural Modulus	2278.9 MPa	D790M
Tensile strength	76.1 MPa	D638M
Tensile modulus	2789 MPa	D638M

Woven E-glass fibers were used for the patches. Compared to other manufactured and natural fibers, glass fiber is a recycled and cheap material with good mechanical strength and durability, so it is suitable for use with pavements subjected to extreme loadings, harsh environments, and a wide range of temperatures. The used glass fiber has a tensile strength of 3450 MPa, a 2.5 g/cm³ density, an elastic modulus of 73 GPa, an elongation at break of 4.8 %, and a softening point of 846 °C.

3. Test method

3.1. Specimen

Experimental studies need to use suitable test specimens. The selection of test specimens depends on parameters such as the material type, manufacturing processes, or loading conditions. For example, disc and cylindrical shape specimens are among the favorite specimens for testing geo-materials [11,[48], [49], [50], [51]]. Disc and cylindrical shape test configurations can easily be extracted from the cylindrical cores obtained during field sampling or laboratory sample preparation (i.e., gyratory or marshal samples). In addition, due to the effect of loading configuration on results, the specimen loading configuration should be selected as similar to actual conditions (which in asphalt mixtures is the three points bend loading) [11,52]. Another important aspect in choosing a test specimen is the ability of the specimen to produce all loading modes (i.e., tensile or shear).

Some of the specimens that can be used in experimental studies of geo-materials, their abilities, advantages, and disadvantages are beam shape specimens subjected to three-point bend loading, which is unsuitable for studying geo-materials [[53], [54], [55], [56], [57], [58]]. Beam shape specimens are subjected to four-point bend loading, which is also unsuitable for studying geo-materials and has a complex loading configuration [[59], [60], [61]]. Brazilian disc (BD) specimens were subjected to compression loading, which required high-capacity loading devices, and produced significant gradients of compressive-tensile stresses during tests [[62], [63], [64], [65]]. Semi-circular bend disc specimens were subjected to three-point bend loading, which has difficulties due to specimen preparation [[66], [67], [68], [69], [70]]. Cube or cylinder shape specimens were subjected to compression loading, which cannot simulate pure shear, requires a high-capacity loading machine, and produces a significant gradient of compressive-tensile stresses during tests [[71], [72], [73], [74], [75]]. And other less common specimens, such as wedge-splitting specimens, have difficulties due to specimen preparation, test setup, or test execution [76,77].

All the above-mentioned specimens have some disadvantages, especially for asphalt material studies. Generally, disc shape specimens are more suited for asphalt material studies, the bending loading is better, and both tensile and shear loading modes should be simulated without test complexity. Recently, Aliha and coworkers [[78], [79], [80], [81]] developed such specimen for investigation of fracture in asphalt material, called Edge Notched Disk Bend (ENDB) specimen. ENDB is a disc-shaped specimen with a mid-section crack on one side. It is loaded using a three-point bend fixture. In this specimen, the symmetrical loading configuration simulates the tensile mode, while the shear mode can be simulated using a specific asymmetrical loading configuration. In addition, for the specimen without the crack, using the process previously developed by Aliha et al. [[78], [79], [80]], the test configuration that results in tensile and shear failure modes was extracted. Fig. 2 shows an illustration of the used specimens to investigate the effect of crack and crack repair on pavement behavior.

Fig. 2.

An illustration of test specimens and test configuration; left) Tensile mode, right) Shear mode.

The specimen's radius, thickness, and crack length are denoted by R, t, and a, respectively. S₁ and S₂ are the distances of span support in three-point bend fixtures to the center of specimens. To simulate tensile and shear failure modes, using studies previously conducted by Bahmani et al. [79], the loading configuration (S₁ and S₂ parameters) is chosen. In Table 4, the specimen dimensions and test configuration for simulating the mentioned modes are presented.

Table 4.

Specimen parameters, test configuration, and related geometry factors.

Mode	R (mm)	t (mm)	aa (mm)	a/ta	S1 (mm)	S2 (mm)
Tensile	50	30	12	0.4	47.5	47.5
Shear					47.5	3.0

^aFor cracked specimens.

3.2. Preparation of specimens, repair procedure

To prepare the specimens, aggregates were heated in an oven at 170 °C for 24 h, bitumen was heated to 135 °C, and combined after weighing. After thoroughly mixing the mixture, it was poured into a cylindrical mold. A Marshall device was utilized for sample compression. Afterward, samples were extracted from the molds and cut with a fixture and a rotary diamond blade. At this point, all specimens are in good condition (Fig. 3a). After securing the specimens in the milling machine, a diamond saw blade was used to create a crack along the diameter of the specimens (Fig. 3b). To repair the crack with bitumen, raw bitumen was liquefied (at 135 °C) and poured into the crack (Fig. 3c). The epoxy resin and hardener were mixed in the proportion of 9 parts resin to 1 part hardener for the repair of the crack with neat epoxy resin (Fig. 3c). After preparing the epoxy resin binder, the adhesive, and silica aggregates in a ratio of 2–8, the polymer concrete was poured into the crack (Fig. 3c). To repair the crack with the layered composite patch, the crack was first filled with epoxy adhesive or bitumen, then the patch was installed. The patch was laminated with a metal roller to ensure proper attachment (Fig. 3d).

Fig. 3.

The preparation process of repaired specimens.

3.3. Test method

As stated earlier, seven specimen conditions were prepared to investigate the influence of crack and crack repairing on pavement behavior. In Table 5, details of specimen labels and related conditions are presented.

Table 5.

The ID, condition, and description of prepared specimens.

ID	Condition	Description
INT	Intact	Specimen in intact condition.
CRC	Cracked	Specimen in cracked condition (ENDB specimen).
BMR	Repaired	Cracked specimen repaired with bitumen.
NER		Cracked specimen repaired with neat epoxy resin.
PCR		Cracked specimen repaired with polymer concrete made of epoxy resin (20 percent) and silica aggregates (80 percent).
BPR		Cracked specimen repaired with woven glass fiber/bitumen patch.
GPR		Cracked specimen repaired with woven glass fiber/epoxy resin patch.

After repairing, the samples were kept in +20 °C for seven days before conducting the tests (to ensure the proper epoxy resin cure). Fig. 4 shows the specimens after the preparation process and before the tests. For each testing condition, three replicate specimens were manufactured and tested.

Fig. 4.

Specimens used for the study: a) INT, b) CRC, c) BMR, d) NER, e) PCR, f) BPR, g) GPR.

Tests were conducted using a universal 6 kN servo-hydraulic testing machine with a load accuracy of 3.0 N and displacement accuracy of 1.0 μm. Loaded with a 1 mm/min crosshead speed.

To compare the behavior of specimens, the maximum load that a specimen can endure is recorded. This load is obtained from the load-displacement curve, known as the ultimate load, fracture load, or failure load. Work in a specimen, member, or structure represents the absorbed energy before failure. Enhancing the energy absorbance capacity prevents the collapse or destruction of structures and infrastructure under extreme loads such as impact [82,83]. The work parameter is determined from the area below the load-displacement (P-u) curve using the following equation:

$$W={\int }_0^{\Delta u}P\cdot du$$ (1)

Two important work parameters are defined as work of failure (W_F) which is obtained by calculating the area below the curve to the peak load (P_u) limit, and total work (W_T), which is obtained by measuring all area below the curve.

4. Results and discussion

4.1. Test records and observations

The failure patterns shown in Fig. 5 are a function of the specimen's condition and the loading mode. This figure depicts the failure pattern of intact specimens, which is a crack that developed in the weakest zone of the specimen, in the middle of the specimen (under the upper loading fixture). A straight crack propagates between the tip of the pre-crack and grows to the upper loading fixture (in the middle) in both cracked and BMR specimens, indicating that the two failure patterns are very similar. Bitumen's viscoelastic behavior at intermediate and high temperatures causes it to undergo large deformations, as is evident in the crack opening process as a result of testing. The failure patterns of NER and PCR specimens are different from BMR specimens; in these specimens, the crack deviates from the repaired zone (mid-section) and occurs at a distance (for both tensile and shear loading conditions). In other words, while in NER and PCR specimens, the crack initiates and develops far from the repaired zone, the crack in the BMR specimen initiates from the tip of the repaired crack and develops in the midsection.

Fig. 5.

Specimens after the failure, based on loading mode and specimen condition.

The failure pattern of BPR and GPR specimens is generally different from other specimens. In the BPR specimen, by reaching the applied load to the ultimate load, the delamination of the glass patch is seen, and in the following, the failure of specimens initiates from the pre-crack tip. In this case, the delamination of the fabric patch occurs due to the viscoelastic behavior of bitumen in intermediate temperatures, which cannot provide proper adhesion between the fabric and the base material (HMA). However, in GPR specimens, due to excellent adhesion of patch reinforcement with base material via epoxy resin, the crack did not initiate from the bottom of the part, and failure occurred in the support region.

Fig. 6 depicts the typical failure patterns of repaired specimens. While the cohesive (Fig. 6b), adhesive (Fig. 6c), and subtraction (Fig. 6a) failures of composite materials are three well-known failure modes. In the present study, BMR specimens exhibit a novel failure mode known as “necking of cohesive” (Fig. 6d). Due to bitumen's viscoelastic behavior, this mode of cohesive failure, which can be classified as a new pattern, causes adhesive materials to neck (necking is a mode of deformation where relatively large amounts of tensile strain localize in a small region of a ductile material). A crack starts from the pre-crack tip and develops in the middle of the specimen after necking.

Fig. 6.

Schematically depicted possible failure patterns; a) substrate failure, b) cohesive failure, c) adhesive failure (interface), d) necking of the cohesive.

Load-displacement curves obtained from tests are presented in Fig. 7. As seen in these curves, the specimen's general behavior is quasi-brittle; after a linear increase of load to the maximum stage, a decrease in the load-bearing capacity of specimens is seen. While the slope of the curve in the linear stage is the material's stiffness, as can be seen, repair of the specimen increases the material's stiffness, especially in tensile loading mode. The slope of the curve after the peak load is also an important parameter, which indicates behaviors such as post-failure energy absorption capability.

Fig. 7.

Load-displacement of specimens, a) tensile mode, b) shear mode.

The average failure loads and energies and related coefficient of variations (COV) are presented in Table 6. As seen, the coefficient of variation values for the ultimate loads is up to 13 percent, and for the work of failure and total work, up to 17 percent. Based on other studies, such COVs are typical for testing engineering materials [84,85].

Table 6.

Ultimate load, total work, work of failure values, and related coefficient of variations.

ID	Tensile			Shear
	Load (N)	WT (J)	WF (J)	Load (N)	WT (J)	WF (J)
INT	778 (10 %)	1.10 (11 %)	0.54 (8 %)	3725 (7 %)	4.01 (17 %)	1.82 (11 %)
CRC	287 (6 %)	0.38 (13 %)	0.14 (18 %)	2066 (9 %)	2.59 (18 %)	1.08 (13 %)
BMR	690 (5 %)	1.03 (2 %)	0.37 (15 %)	3127 (13 %)	3.42 (15 %)	1.42 (1 %)
NER	1460 (10 %)	1.89 (9 %)	1.13 (17 %)	4398 (4 %)	6.00 (10 %)	2.84 (3 %)
PCR	1396 (3 %)	1.83 (14 %)	1.04 (12 %)	4646 (3 %)	5.79 (5 %)	2.56 (5 %)
BPR	1075 (7 %)	1.46 (6 %)	0.75 (15 %)	3762 (8 %)	4.71 (13 %)	2.13 (2 %)
GPR	2788 (4 %)	8.55 (2 %)	2.45 (11 %)	5217 (2 %)	10.98 (5 %)	3.65 (5 %)

4.2. Efficiency of repairs

By defining two parameters, modified tensile strength ratio (MTSR) and modified shear strength ratio (MSSR), the specimen tensile or shear strength change due to crack appearance or crack repair can be calculated. The MTSR and MSSR parameters are defined as:

$${MTSR}_i^n=\frac{n_i-n_{intact}}{n_{intact}}\times 100$$ (5-1)

$${MSSR}_i^n=\frac{n_i-n_{intact}}{n_{intact}}\times 100$$ (5-2)

n = Load, W_F, or W_T; i = CRC, BMR, NER, PCR, BPR, or GPR.

The average ultimate tensile and shear loads, as well as the MTSR and MSST parameters of specimens, are shown in Fig. 8. As can be seen, the ultimate tensile load of the INT specimen is 778 N, whereas the ultimate tensile load of the CRC specimen is 287 N, which is 63 % less than the intact condition (${MTSR}_{cracked}^{Load}=-63\%$). The ultimate tensile load of the BMR specimen is also 690 N (${MTSR}_{BMR}^{Load}=-11\%$), which is higher than the cracked specimen but still lower than the intact condition. The ultimate tensile load of NER and PCR, on the other hand, is 1460 N (${MTSR}_{NER}^{Load}=88\%$) and 1396 N (${MTSR}_{PCR}^{Load}=79\%$), which is higher than the intact specimen. Finally, the ultimate tensile load of the BPR and GPR specimens is 1075 N (${MTSR}_{BPR}^{Load}=38\%$) and 2788 N (${MTSR}_{GPR}^{Load}=258\%$), respectively. Cracked and BMR specimens in shear condition have MSSR of −45 % and −16 %, respectively, while NER, PCR, BPR, and GPR have MSSR of 18 %, 25 %, 1 %, and 40 %, respectively.

Fig. 8.

a) Average ultimate loads and MTSR and MSSR parameters.

Fig. 9 displays the average work of failure and total work (respectively WF and WT) for both tensile and shear modes (see Fig. 7 for definitions). The trend of W_F and W_T values is the same as the ultimate load value. While work of failure for CRC specimen reduced by about 75 % for tension mode and 41 % for shear mode, respectively, the ${MTSR}_{BMR}^{WF}=-32\%$ and ${MSSR}_{BMR}^{WF}=-22\%$ values show that the work of failure that was lost due to the crack cannot be recovered after it was repaired with bitumen (BMR specimen). As with load data, repair with epoxy resin and polymer concrete (NER and PCR specimens) increased the work of failure even more than in intact condition (by at least 41 % and at most 111 %). Also, the work of failure of BPR and GPR specimens is calculated to be about 33 % and 680 % higher than the intact condition in tensile mode and 39 % and 355 % higher than the intact condition in shear mode. The same trends are also seen for the total work parameter.

Fig. 9.

a) Average work of failure and related MTSR and MSSR parameters.

4.3. Numerical modeling

Using CPE8 elements, a finite element model of the specimens is developed in the ABAQUS program. Approximately 60,000 elements were used after a series of analyses were performed as a convergence study to remove the model's sensitivity to mesh number. Simulation accuracy can be improved by making the mesh finer around the middle. The mesh pattern and boundary condition of the specimen used in the present modeling are shown in Fig. 10.

Fig. 10.

Finite element model and boundary conditions of the used ENDB specimen.

From the mechanical point of view, the unidirectional tensile stress in the midsection due to the application of ultimate loads (presented in Table 6) is obtained by numerical modeling and presented in Fig. 11. As seen, the ultimate tensile stress of tested asphalt materials is about 1.2 N/mm² which is obtained from intact specimens in the bottom side of the disc's midsection. Also, modeling of cracked and BMR subject to their ultimate loads shows that the ultimate tensile stress in asphalt reaches the maximum level (1.2 N/mm²) in the pre-crack tip, and failure initiates from the pre-crack tip in cracked and BMR specimens. Such tensile strength is consistent with other research results [[86], [87], [88]].

Fig. 11.

Diagram of the unidirectional stress in the mid-section of specimens obtained by numerical models.

Modeling shows that due to the repair of specimens with resin epoxy (NER) and polymer concrete (PCR), the tensile strength in the critical tensile stress of the specimen mid-section becomes about 2.05 N/mm², however as this stress occurs in the epoxy resin and polymer concrete region, they have a much higher tensile strength than existing stress (tensile strength of epoxy resin and polymer concrete used in the current paper are previously measured as 76.1 and 15 N/mm² respectively). In this case, failure is expected to occur somewhere distant from the repaired zone.

Using a glass fiber patch, the stress in the specimen asphaltic portion was reduced significantly and considering the high tensile strength of glass fiber (2950 N/mm² in the current case), the ultimate load capacity of the GPR specimen increased significantly. However, for the BPR specimen, the fabric patch and base material separation are seen due to the low elasticity of bitumen.

For better understanding, the normalized stresses in the mid-section of specimens are presented in Fig. 12b. To draw this diagram, the stress of each material is divided into its ultimate strength. As seen in this figure, for INT, CRC, and BMR specimens, the value of the existing tensile stress reaches the tensile strength value of asphalt material, so their curve reaches the tensile failure limit (red line on the right in Fig. 11b). In NER and PCR specimens, although existing tensile strength is higher than intact specimen, as the tensile strength of epoxy resin and polymer concrete is high, compare to their tensile strength, the existing stress is relatively low (3 % for epoxy resin and 15 % for polymer concrete). Based on that, it is expected that only in INT, CRC, and BMR specimens did the failure take place in mid-section (which is consistent with experimental observations), and in NER and PCR, the failure of specimen onset in other sections.

Fig. 12.

a) Failure pattern of GPR specimen under tensile loading mode, b) stresses in GPR specimen subjected to ultimate tensile load.

The stress distribution in the BPR specimen is interesting, as seen earlier during the test. In BPR specimens, the failure of the specimen occurred by separation of the fabric patch following the crack propagation from the pre-crack tip. Such behavior can also be predicted from the numerical moldings. As seen in Fig. 11b at the failure stage, the stress of the bitumen layer, which is used to attach the fabric patch, reached the ultimate stage, and the stress in the pre-crack tip is also in the ultimate range, so it is excepted to failure of BPR specimens were initiate and continued in the mid-section.

Failure of the GPR specimen under tensile loading is different from other specimens. While the failure of all other specimens occurs in a plane parallel to the loading plane, the failure path of the GPR specimen is initiated and developed in a S-shape route that has an intersection with loading and support lines (Fig. 12). Investigation of stresses in GRP specimen shows, the maximum tensile and compressive stress is specimen is below the tensile and compressive strength of materials, the tensile and compressive stress is the asphaltic region. In this condition and based on the failure pattern, the failure occurs due to shear stress. Such behavior can also see in concrete beams that failed in shear mode [89,90]. Using the Von Mises stress criteria, the maximum stress in the asphalt region of the GRP specimen is calculated as 4.32 N/mm² (Fig. 12b), which is consistent with the shear strength of asphalt calculated in the following section.

At first, to ensure the dominance of shear failure, after simulating the INT and CRC specimens in shear condition (asymmetrical loading condition), both tensile and shear stresses are checked. For this purpose, after the modeling, the ultimate loads (presented in Table 6) are applied to the specimen, and stresses are extracted. As seen in Fig. 13, in both INT and CRC specimens, the tensile stresses are lower than the tensile strength value of asphalt material (previously calculated as 1.2 N/mm²), so the failure of specimens is a result of shear stresses. Based on that, after reviewing the shear stresses in both specimens, the ultimate shear strength of asphalt material was obtained as 4.05 and 4.26 N/mm².

Fig. 13.

Tensile and shear plots of Intact and Cracked specimens under related ultimate shear load.

Fig. 14 shows the shear stress plots of the specimens in the midsection. The maximum shear stress in these specimens is in the range of 4.04 and 4.17, consistent with the shear strength of asphalt material calculated previously.

Fig. 14.

Shear stress plots of specimens a) BMR, b) NER, c) PCR, d) GPR.

5. Conclusion

The mechanical behavior of asphalt concrete is examined in this study through the use of a variety of construction materials, including bituminous, epoxy resin, polymer concrete, and glass/epoxy composite patches, to investigate the impact of fractures and breach restoration. The investigation is conducted both experimentally and numerically. The subsequent observations may be extracted from the investigation:

• •In the intact specimen, the failure pattern is a single crack that propagates in the specimen's midsection. The failure pattern of cracked and bitumen-repaired specimens is similar, with a straight crack that developed in the midsection between the pre-crack and the upper loading fixture. In the NER and PCR specimens, the crack deviates from the repaired zone (mid-section) and occurs at a distance.
• •In specimens repaired by fabric patch and bitumen adhesive, the delamination of the patch is seen because of poor adhesion. In addition, in patch-reinforced specimens that used epoxy resin as adhesive due to excellent adhesion, the crack did not initiate from the bottom of the part, and failure occurred in the support region.
• •A new failure mode named “necking of cohesive” is seen in bitumen-repaired specimens. Due to viscoelastic behavior, the necking phenomenon, a ductile failure mode in ductile materials, is seen in the adhesive material.
• •The ultimate tensile capacity of the cracked specimen is measured at 63 % lower than the intact condition. The ultimate tensile load of the bitumen-repaired specimen is higher than that of the cracked specimen, but it is still 11 % lower than that of the intact condition. The ultimate tensile capacity of epoxy resin repaired and polymer concrete repaired specimens are 88 % and 79 % higher than the intact specimens. The ultimate tensile load of the fabric patch reinforced specimen that used bitumen as the adhesive is 38 % higher than the intact specimen, while for the case of using the epoxy resin adhesive, this value is 258 %.
• •In shear mode, the ultimate load for cracked and bitumen-repaired specimens decreases by 45 % and 16 %, respectively, while epoxy resin and polymer concrete repairs increase the load by 18 % and 25 %. Fabric patch-reinforced specimens using bitumen adhesive show a 1 % improvement over the intact specimen, whereas using epoxy resin adhesive results in a 40 % increase.
• •While the cracked specimen's work of failure in tensile and shear modes was reduced by about 75 % for tension mode and 41 % for shear mode, repair with bitumen cannot recover the whole work of failure. Similar to the ultimate load, repair with epoxy resin and polymer concrete increased the work of failure, even higher than in intact condition, by at least 41 % and at most 111 %.
• •The failure work of the glass patch/bitumen-reinforced specimen shows an increase of 17 % and 33 % over the intact condition in tensile and shear modes, respectively. For tensile and shear modes, the increase reaches 680 % and 147 %. Similar trends are observed for the total work parameter across all specimens.
• •Assuming the repaired specimens are dual-material, the substance's elastic modulus, tensile strength, and repair material have a significant role. Regarding Young's modulus and tensile strength, different failure modes, including substance failure, cohesive or adhesive failure, or their combination, can occur.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

CRediT authorship contribution statement

M.R.M. Aliha: Supervision, Resources, Project administration. Hamid Reza Karimi: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ehsan Khedri: Writing – original draft, Resources, Methodology, Investigation, Formal analysis, Conceptualization. Sepehr V. Abdipour: Resources. Pegah Jafari Haghighatpour: Visualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.