Experimental and Statistical Evaluations of Recycled Waste Materials and Polyester Fibers in Enhancing Asphalt Concrete Performance

Abstract

This research aimed to evaluate the impact of using brick waste powder (BWP) and varying lengths of polyester fibers (PFs) on the performance properties of asphalt concrete (AC) mixtures. BWP was utilized as a replacement for traditional limestone powder (LS) filler, while PFs of three lengths (3 mm, 8 mm, and 15 mm) were introduced. The study employed the response surface methodology (RSM) for experimental design and analysis of variance (ANOVA) to identify the influence of BWP and PF on the selected performance indicators. These indicators included bulk density, air voids, voids in the mineral aggregate, voids filled with asphalt, Marshall stability, Marshall flow, Marshall quotient, indirect tensile strength, wet tensile strength, and the tensile strength ratio. The findings demonstrated that BWP improved moisture resistance and the mechanical performance of AC mixes. Moreover, incorporating PF alongside BWP further enhanced these properties, resulting in superior overall performance. Using multi-objective optimization through RSM-based empirical models, the study identified the optimal PF length of 5 mm in combination with BWP for achieving the best AC properties. Validation experiments confirmed the accuracy of the predicted results, with an error margin of less than 8%. The study emphasizes the intriguing prospect of BWP and PF as sustainable alternatives for improving the durability, mechanical characteristics, and cost-efficiency of asphalt pavements.

1. Introduction

Asphalt concrete (AC) is considered a composite material that contains aggregates, asphalt binder, and air voids (Vas). AC has been extensively used in the construction of road pavements. Nevertheless, AC mixes are susceptible to distresses such as cracking, stripping (separation of asphalt from aggregate), and rutting (permanent deformation) under the effects of contraction at low temperatures, moisture, and repeated vehicle loads at high temperatures [1]. Accordingly, there are a number of ways to enhance the performance of AC mixes. One way is to change the type of filler in the asphalt mix. Another way is to add fibers to the AC mix. Fibers have been used in various studies to enhance the mechanical properties of AC mixes and to alter the engineering properties of asphalt binders [2,3,4,5,6,7,8,9,10]. However, many researchers have investigated the impact of various filler types on the properties of asphalt and AC mix [11,12,13,14,15,16,17,18,19,20].

Nowadays, the use of waste materials as alternative fillers in road pavements is a critical subject due to their environmental impact and economic benefits. While some research focuses on recycled brick as fine aggregate [21,22], its potential as a filler material remains a viable area for sustainable pavement improvement. Several researchers have investigated the use of brick waste powder (BWP) instead of conventional filler in AC mixes and asphalt binder. Wu et al. [23] evaluated the influence of recycled red brick powder (RBP) as a filler on certain characteristics of asphalt mastic. Their findings suggested that RBP had a positive impact on the asphalt mastic’s high-temperature properties. Almadwi and Assaf [24] examined the possibility of using brick powder (BP) as filler in AC mixes and found that mixes containing BP had greater mechanical properties than those with limestone powder (LSP). The effects of waste materials like waste brick powder (WBP), waste glass powder (WGP), and rice husk ash (RHA) as fillers on the characteristics of AC mixes were investigated by Arabani et al. [25]. Their findings demonstrated that the mixes containing WGB and WBP outperformed the other mixes. Chen et al. [26] studied the mechanical characteristics of AC mixes containing WBP as a mineral filler. Their findings revealed that the inclusion of recycled brick powder enhanced AC’s mechanical performance, and they concluded that utilizing brick powder as a filler in AC mixes is feasible.

Previous researchers have used synthetic fibers, such as polyester fibers (PFs), to reinforce asphalt binders and AC mixes. Chen and Lin [27] assessed the effect of cellulose, rock wool, and polyester fibers on asphalt binders’ engineering properties, and they concluded that the addition of fibers stiffened the asphalt and had potential benefits in reducing drain down. The load-bearing capacity of the asphalt–fiber mastic was also improved by the strong adhesion between the fibers and the binder. Xu et al. [1] conducted an investigation into the mechanisms and reinforcing effects of polyacrylonitrile, lignin, polyester, and asbestos fibers on the performance of AC mixes. They discovered that the polyester and polyacrylonitrile fibers improved the split indirect tensile strength (SITS), fatigue life, and rutting resistance of the asphalt mixes more than the asbestos and lignin fibers, mainly because of their greater network function. The associated properties of PF-modified asphalt and PF-modified AC mix were studied by Wu et al. [28], and they confirmed that the incorporation of PF enhanced asphalt binder viscosity and the fatigue characteristics of the AC mixes, especially at lower stresses. Kim et al. [29] investigated the influence of polypropylene, carbon, polyester, and nylon fibers on the mechanical characteristics of AC. They showed that the mechanical performance of the majority of the fiber-reinforced AC specimens significantly improved compared to the unreinforced AC specimen. The impacts of polyester fiber (PF) and phosphorus slag powder (PSP) on the rheological properties of asphalt binders as well as the mechanical characteristics of AC mixes were evaluated by Sheng et al. [30]. They revealed that the use of PSP increased the viscosity of the binder, resulting in improved AC rutting resistance, while the addition of PF significantly enhanced the AC mix’s resistance to moisture damage and low-temperature cracking. Chen et al. [31] investigated the effect of using asbestos, polyacrylonitrile, lignin, and polyester fibers on the mechanical and volumetric properties of AC mixes, and their findings displayed that the addition of fibers to AC mixes resulted in a reduction in bulk density and an increase in optimum asphalt content, Marshall stability, voids in mineral aggregate, and Vas. They also found that the lignin and asbestos fibers had greater optimum asphalt content and voids filled with asphalt (VFAs) due to their higher asphalt absorption, whereas the polyacrylonitrile and polyester fibers had better stability values due to their higher network effect. In another study, Alnadish et al. [32] assessed the performance of the asphalt mixes incorporating coarse steel slag aggregate and three types of synthetic fibers: acrylic, polyester, and polyvinyl alcohol. Their findings revealed that asphalt mixes reinforced with fibers performed best in terms of rutting resistance, indirect tensile strength, tensile strength ratio, and cracking resistance.

The incorporation of brick waste powder (BWP) and polyester fibers (PFs) into asphalt concrete has garnered attention recently for their potential to enhance pavement performance and support sustainability efforts. BWP, sourced from recycled construction debris, is mainly utilized as a substitute for traditional mineral fillers. The inclusion has been demonstrated to markedly enhance the environmental profile of asphalt mixtures and improve specific mechanical properties [26,33]. BWP has been shown to enhance Marshall stability, tensile strength, and stiffness in asphalt mixes. Higher replacement levels of up to 100% improved rutting and moisture resistance without compromising volumetric properties, and had optimal performance at around 5.6% asphalt content [34]. PFs, derived from recycled polyethylene terephthalate (PET) waste, have been successfully employed to improve the high- and low-temperature performance of asphalt mixtures, extend fatigue life, and enhance water stability [35]. Furthermore, PFs facilitate the incorporation of increased asphalt binder content while maintaining rutting resistance, thus enhancing pavement durability and extending service life [36].

Both BWP and PF, while advantageous, exhibit limitations that require thorough assessment. While BWP enhances the mechanical strength and moisture resistance of asphalt mixtures, its inclusion may lead to reduced deformation resistance due to its brittle nature and lack of elasticity, requiring a balanced mix design to optimize overall performance [34]. The performance of PF varies based on fiber type, length, and dosage, resulting in inconsistent outcomes across different asphalt mix designs [37]. The long-term field performance of these materials under diverse climatic and loading conditions is not yet fully understood. The identified gaps highlight the necessity for integrated studies that assess both the individual effects of BWP and PF and their combined impact on the behavior of asphalt concrete.

Fiber length is among the most crucial factors affecting asphalt properties and AC mix performance. Qian et al. [38] assessed the reinforcing impacts of aramid and polyester fibers of different lengths on asphalt binders at low temperatures and found that the optimum length of PF was about 6 mm, which corresponds to the best tensile properties of asphalt. Putman and Amirkhanian [39] incorporated carpet, scrap tire, polyester, and cellulose fibers into AC mixes as stabilizing additives. They used a PF length of 6.35 mm and a tire fiber length of 3 to 13 mm, and the results showed that adding polyester, carpet, and tire fibers significantly enhanced the toughness of the AC mixes in comparison to the cellulose fibers. The influence of polyester, cellulose, and mineral fibers with lengths of 6 mm, 1.1 mm, and 6 mm, respectively, on the dynamic and fatigue properties of AC mixes was examined by Ye et al. [40]. They also stated that, after the incorporation of the fibers into the AC mixes, the phase angle and the dynamic modulus reduced compared to the control mix. Nonetheless, the incorporation of fibers in the AC mixes provided enhanced performance in terms of fatigue resistance as compared to the control mix. They also summarized that the PF had a better effect on improving the fatigue resistance of the AC mix. Hong et al. [41] added 12 mm-long PF and coal gangue powder (CGP) to enhance the performance of the AC mix. They stated that the interaction of PF and CGP at their optimum provided an excellent improvement in the AC mix’s low-temperature crack resistance.

Response surface methodology (RSM) is a mathematical tool that combines theoretical, statistical, and numerical procedures to create and optimize models: this methodology is commonly employed when multiple factors affect one or more responses [42,43,44,45]. The RSM is the most commonly designed experiment applied; it allows us to study multiple factors simultaneously and compute the interactions between the variables being considered with minimal experimentation [42,43,44]. The application of response surface methodology has been widely validated in materials research and optimization studies [46,47,48]. More concretely, RSM also aims to propose sequential strategies for the execution of experiments and the verification of the agreement between the experimental data and the constructed models [49,50,51,52,53]. RSM was used as a partial factorial design to minimize the number of experiments compared to a full factorial design. For example, to conduct a full factorial design with five variables at five distinct levels, it would require 3125 experiments. However, the RSM only requires 48 experiments [54]. Additionally, it is acknowledged that experimenters might lack the necessary time or resources to conduct full factorial experiments; therefore, they may opt for commonly used partial factorial designs [55].

The combination of brick waste powder (BWP) and polyester fiber (PF) was chosen for their synergistic advantages in improving performance and sustainability. BWP, abundant in silica and alumina, enhances filler–binder adhesion and moisture resistance, while facilitating the circular utilization of building waste. PF, sourced from recycled PET, offers tensile reinforcement and crack-bridging properties, enhancing fatigue resistance and ductility. Collectively, these additives provide a cost-efficient, eco-friendly, and technically feasible option to enhance the durability, stability, and resilience of asphalt mixtures under diverse loading and environmental circumstances.

Despite the promising mechanical and durability improvements reported in prior works, current studies on BWP often lack a detailed assessment of its performance across multiple interdependent properties using advanced statistical tools. Moreover, the combined influence of polyester fibers with BWP on asphalt concrete’s performance, especially considering varying fiber lengths, remains underexplored. Limitations in the literature also include minimal optimization studies to identify the most effective PF length when used alongside BWP. This research addresses these gaps by evaluating a range of performance indicators, integrating response surface methodology (RSM) and ANOVA to model interactions and validating findings experimentally. The novelty lies in identifying optimal design parameters for AC mixtures using recycled materials while enhancing both strength and durability. The goals of this investigation include the following:

(i)

Evaluating the interactive effects of brick waste powder (BWP) as a filler and polyester fibers (PFs) of varying lengths (3 mm, 8 mm, and 15 mm) on key mechanical and durability properties of asphalt concrete (AC) mixes;

(ii)

Developing and validating predictive models for bulk density, air voids, Marshall properties, indirect tensile strength, and tensile strength ratio using response surface methodology (RSM) and analysis of variance (ANOVA);

(iii)

Determining the optimal combination of filler type and fiber length to enhance AC performance, with a specific focus on identifying the most effective PF length (found to be approximately 5 mm) through multi-objective optimization.

Research Significance

This study presents a novel approach to improving asphalt concrete (AC) mixtures by combining recycled brick waste powder (BWP) as a filler and polyester fibers (PFs) of varying lengths. The originality lies in the simultaneous investigation of these two sustainable additives using a statistically optimized response surface methodology (RSM). While previous studies have individually explored waste fillers or fiber reinforcement, this research uniquely evaluates their interactive effects on both mechanical and moisture-resistance properties.

A key scientific contribution is the use of RSM and ANOVA to model 10 output responses, leading to an optimal PF length (approximately 5 mm) when paired with BWP. This optimization is validated experimentally with <8% error, confirming the reliability of the proposed method. The study demonstrates that incorporating BWP and PF can enhance indirect tensile strength by over 38%, improve moisture resistance (ITSR), and maintain mechanical integrity within standard specifications.

Practically, the materials are low-cost and recyclable, the formulations are production compatible, and the results support wider adoption of circular economy principles in pavement design. This work contributes both to material innovation and performance modeling, filling a gap in multi-additive asphalt optimization studies.

2. Materials and Experiments

This study began with the selection of materials, including brick waste powder (BWP) and limestone powder (LS) as fillers, coarse and fine aggregates, and polyester fibers (PFs) of three different lengths (3 mm, 8 mm, and 15 mm). Asphalt concrete (AC) samples were fabricated using a standard SCAC 0/14 gradation. A total of ten performance indicators were evaluated, including volumetric properties, Marshall characteristics, and indirect tensile strength (ITS) under both dry and moisture-conditioned states. The experimental results were analyzed using response surface methodology (RSM) followed by analysis of variance (ANOVA) to identify significant effects and interactions. Optimization and experimental validation were then conducted to determine the most effective PF length. The overall research methodology and workflow are illustrated in Figure 1.

2.1. Materials

2.1.1. Aggregates

The coarse and fine aggregates used in the present work were crushed limestone aggregates obtained from the ENOF quarry in Setif (eastern Algeria). Three aggregate fractions were used, 8/16 and 4/8 coarse aggregate as well 0/4 fine aggregate. Limestone powder (LS) was used as filler.

Table 1. Aggregate physical properties.

Property	Standard	Coarse Aggregate (8/16)	Coarse Aggregate (4/8)	Fine Aggregate (0/4)	Specification [56]
Specific gravity (g/cm3)	EN 1097-6	2.695	2.693	2.697	–
Water absorption (%)	EN 1097-6	0.32	0.44	0.33	–
Los Angeles value (%)	EN 1097-2	23.76	–	–	≤25
Micro-Deval value (%)	EN 1097-1	16.06	–	–	≤20
Friability coefficient (%)	NF P 18-576	–	–	32.77	≤35
Flakiness index (%)	EN 933-3	6.68	11.40	–	≤25
Surface cleanliness (%)	NF P 18-591	0.38	0.55	–	≤2
Sand equivalent (%)	EN 933-8	–	–	77.62	≥55
Methylene blue value	EN 933-9	–	–	0.35	≤2

and

Table 2. Chemical compositions of aggregate.

Component	Result (%)
Na2O	0.0290
MgO	1.37
Al2O3	0.161
SiO2	0.283
P2O5	0.0101
SO3	0.136
K2O	0.0213
CaCO3	97.3
Fe2O3	0.157
SrO	0.0596
BaO	0.463
Total	100

present the physical properties and chemical compositions of the aggregates studied, respectively. All physical properties of the aggregates are in accordance with Algerian specifications [56].

The selected aggregate gradation falls within the gradation limits of semi-coarse asphalt concrete 0/14 (SCAC 0/14), which is generally intended for the wearing courses of flexible pavements in Algeria [56]. The aggregate grading curve is displayed in Figure 2 and is obtained for the following proportions: 35% coarse aggregate (8/16), 22% coarse aggregate (4/8), 38% fine aggregate (0/4), and 5% filler (LS).

2.1.2. Asphalt

The asphalt binder utilized throughout this investigation was 35/50 penetration-grade pure bitumen supplied by the General Industry Bitumen Setifien Company, located in the Mezloug industrial zone, in Setif, Algeria. This grade of asphalt is used on a large scale for the construction of roads in Algeria.

Table 3. Asphalt physical properties.

Property	Standard	Result	Algerian Specification [56]
Penetration at 25 °C, 100 g, 5 s (1/10 mm)	EN 13880-2	38	35–50
Ring and ball softening point (°C)	EN 1427	51	47–60
Specific gravity at 25 °C (g/cm3)	NF T66-007	1.018	1.0–1.1

summarizes the binder properties that satisfy the requirements of the Algerian specification [56].

2.1.3. Fillers

The red brick waste was obtained from buildings damaged by the earthquake that occurred in Mila, Algeria, on 7 August 2020. The brick waste material was washed with water and then dried in an oven. After that, it was crushed using a jaw crusher and ground by a ball mill. Brick waste powder (BWP), which passed via a 0.080 mm sieve, was used as the filler in this investigation. Figure 3 illustrates recycled brick waste preparation. On the other hand, limestone powder (LS) passing the 0.080 mm sieve was employed as the conventional filler and was gathered locally. The percentage of both conventional and waste fillers was fixed at 5% by dry aggregate weight. The particle size distributions of LS and BWP were determined using a Laser Scattering Particle Size Distribution Analyzer, and the findings are displayed in Figure 4. The particle diameters D10, D50, and D90 were (6.92 µm, 37.02 µm, and 111.68 µm) and (3.17 µm, 23.80 µm, and 144.89 µm) for BWP and LS, respectively. The specific gravities of BWP and LS were 2.378 g/cm³ and 2.709 g/cm³, respectively. It can be observed that BWP has a lower specific gravity than LS.

Table 4. Chemical compositions of LS and BWP.

Component	BWP (%)	LS (%)
Na2O	0.714	0.0335
MgO	2.32	0.533
Al2O3	14.8	0.412
SiO2	54.8	1.05
P2O5	0.242	0.0073
SO3	0.439	0.0555
Cl	0.0716	–
K2O	1.90	0.106
CaO	14.9	–
CaCO3	-	97.6
TiO2	0.902	–
V2O5	0.0488	–
Cr2O3	0.0296	–
MnO	0.0929	–
Fe2O3	8.49	0.265
ZnO	0.0189	–
As2O3	0.0057	–
Rb2O	0.0090	–
SrO	0.111	0.0253
ZrO2	0.0338	–
Total	100	100

and Figure 5 provide the chemical and mineral compositions of the investigated fillers, respectively. The chemical compositions of the fillers were measured by X-ray fluorescence (XRF), while the mineral compositions of the filler materials were determined by X-ray diffraction (XRD). As presented in Table 4, SiO₂, CaO, Al₂O₃, and Fe₂O₃ are the main ingredients of BWP, and CaCO₃ is the major constituent of LS. As shown in Figure 5, the principal mineral-phase of BWP is quartz. On the other hand, LS consists mainly of calcite. The microscopic morphology of the fillers was performed using scanning electron microscopy (SEM). Figure 6 shows the SEM images of BWP and LS at various magnifications, with arrows showing major morphological aspects. The particles of BWP have an angular shape with a relatively rough and porous surface, while the particles of LS have an irregular shape with a rough surface. This suggests that BWP has a higher adsorption capacity compared to LS.

The chemical composition of brick waste powder (BWP), distinguished by elevated concentrations of silicon dioxide (SiO₂, 54.8%), aluminum oxide (Al₂O₃, 14.8%), and ferric oxide (Fe₂O₃, 8.49%), signifies substantial pozzolanic activity, which is crucial for enhancing the efficacy of asphalt binders and composites. These oxides enhance chemical interactions at the filler–binder interface, hence augmenting adhesion and surface reactivity. The pozzolanic reactions induced by SiO₂ and Al₂O₃ result in the generation of secondary hydration products, notably calcium silicate hydrates (C-S-H), which are crucial for enhancing microstructural integrity and long-term strength in cementitious and asphaltitious systems [57,58].

This reactive profile improves moisture resistance since the polar components in BWP create physicochemical connections with the bitumen, diminishing water susceptibility and stripping effects in hot-mix asphalt. This effect is especially advantageous in arid or desert regions, where binder–aggregate cohesiveness is frequently undermined [24].

BWP has shown a significant capacity to improve mechanical qualities when utilized as a partial replacement for traditional fillers or cementitious substances, beyond mere chemical contact. Research indicates that the integration of BWP can optimize pore architecture, augment tensile strain capacity, and improve bond strength in high-performance materials. In ultra-high-performance concrete (UHPC), brick powder enhanced packing density and internal bonding, resulting in better durability [59]. In engineered cementitious composites (ECCs), BWP was seen to diminish porosity and enhance tensile performance, confirming its viability as a supplemental cementitious material [60].

The mechanical improvements in cement-based systems are similarly applicable to asphalt concrete (AC), especially with enhanced indirect tensile strength (ITS), Marshall stability, and deformation resistance when BWP substitutes for traditional limestone filler.

The implementation of BWP in pavement and construction applications provides both technical advantages and substantial environmental and economic benefits. It facilitates waste valorization through the recycling of brick debris from demolition, thus alleviating landfill disposal and decreasing the reliance on virgin materials like limestone or fly ash—resources that are increasingly scarce or linked to elevated embodied energy [58,60].

Notwithstanding these benefits, obstacles including the inconsistency of brick waste composition, processing demands, and the logistics of extensive implementation must be resolved. Thorough environmental evaluations, encompassing particulate emissions, toxicity assessments, and life-cycle costing, must accompany technological validation to guarantee sustainable implementation.

Limestone powder (LS), mostly consisting of calcium carbonate (CaCO₃, ~97.6%), is extensively utilized in construction owing to its availability, economic efficiency, and function in filler gradation regulation [61,62]. Although LS can serve as a nucleation site for hydration products in cement systems, its chemical inertness and poor solubility considerably restrict its reactivity and filler–binder interaction, particularly in asphalt mixes [63,64].

This constraint is crucial in asphalt concrete, where moisture-related corrosion and adhesion failure are significant issues. The incapacity of CaCO₃ to establish robust chemical bonds with polar bitumen constituents limits its efficacy in preventing water infiltration and preserving binder cohesiveness under stress conditions [64]. Moreover, an overabundance of limestone filler can adversely affect mechanical strength and durability, as evidenced in specific cementitious applications where excessive LS content diminished binder concentration and decreased stiffness [61].

Unlike LS, BWP, characterized by its elevated silica and alumina content, establishes more robust polar interactions at the binder–filler interface, leading to enhanced adhesion, moisture resistance, and overall mechanical performance. Furthermore, carbonated alkaline solid wastes have been investigated as novel alternatives to LS, providing improved mechanical strength and sustainability through the utilization of industrial by-products [65]. Limestone is a cost-effective filler, yet its restricted chemical reactivity highlights the necessity for performance-oriented alternatives in sophisticated pavement designs. Brick waste powder, because of its chemical reactivity, mechanical reinforcing capabilities, and sustainability advantages, presents a compelling option for substituting or augmenting LS in forthcoming construction mixtures.

2.1.4. Fiber

The polyester fiber (PF) used in the present study was provided by Sebtex Fibers Company, located in the Chelghoum Laïd industrial zone in Mila, Algeria. It is made from recycled polyethylene terephthalate (PET) bottle waste. This fiber was used as an additive in the AC mix. Three different lengths of PF were used in this study, 3 mm, 8 mm, and 15 mm, as shown in Figure 7. The choice of these specific lengths of polyester fibers for reinforcing AC 0/14 is based on the different aggregate fractions (0/4, 4/8, and 8/16) and the maximum aggregate size of 14 mm, as indicated in 2.1.1. The primary reason for this selection is to achieve appropriate dispersion of the fibers within the AC mixes, which may enhance their properties. In addition, the PF content used is 0.3% by weight of dry aggregate.

Table 5. PF physical properties.

Property	Result
Color	White
Length (mm)	3 mm; 8 mm and 15 mm
Diameter (µm)	20
Aspect ratio	150; 400 and 750
Specific gravity (g/cm3)	1.36 ± 0.03
Tensile strength (MPa)	369 ± 27
Elongation (%)	39 ± 5.0
Melting point (°C)	251 ± 5.0

lists the basic physical properties of PF.

The selection of polyester fiber (PF) lengths—3 mm, 8 mm, and 15 mm—in asphalt mixtures is intentional, aimed at enhancing performance characteristics including toughness, fatigue resistance, and moisture sensitivity [38,39,40,41]. Polyester fibers are recognized for their cost-effectiveness and accessibility compared to other kinds of fibers, with those utilized in road building generally being factory-processed. These fibers effectively limit aggregate displacement, absorb asphalt, and prevent micro-crack formation, while also enhancing noise reduction and sound absorption in asphalt pavements.

Fibers of shorter lengths, such as 3 mm, are typically more conducive to uniform dispersion within the asphalt mixture. The ease of dispersion mitigates mix segregation, resulting in a more uniform material. Although certain fiber lengths may theoretically provide optimal tensile strength, practical factors frequently prefer shorter fibers due to their enhanced workability and successful integration. The 8 mm intermediate length is selected to provide a good equilibrium between usability and improved performance. This length seeks to ensure effective mechanical interlocking while mitigating the dispersion issues associated with longer fibers. Fibers often enhance the mechanical qualities of asphalt mixtures by bridging cracks and distributing loads, which is essential for augmenting tensile strength and ductility.

Longer fibers, such as those measuring 15 mm, are anticipated to enhance mechanical interlocking inside the asphalt matrix. Nonetheless, they entail an increased risk of inadequate dispersion and aggregation, which can adversely impact the mixture’s consistency and overall efficacy. The specified PF lengths are meticulously selected to correspond with the various particle sizes utilized in the asphalt mixture. This alignment guarantees the effective integration and dispersion of the fibers inside the asphalt matrix, thereby optimizing their reinforcing impact. Basalt aggregate, characterized by its angular form and coarse texture, facilitates this efficient incorporation. The meticulous selection of polyester fiber lengths is a strategic method to customize the qualities of the asphalt mixture, optimizing dispersion, mechanical interlocking, and aggregate characteristics for enhanced low-temperature fracture resistance and overall longevity.

2.2. Mix Design

The Marshall mix design method was utilized for the determination of the optimum asphalt content (OAC) of the reference mix (i.e., the mix containing LS as a filler) in accordance with EN 12697-34 [66]. A total of 12 samples of AC mixes were manufactured at 4 different asphalt proportions ranging from 5.5% to 6.2% by dry aggregate weight. Three samples were produced for each asphalt content, and the average values of Marshall stability, Marshall flow, Marshall quotient, bulk density, Vas, voids in mineral aggregate (VMAs), and VFAs were determined. From the results obtained, the OAC was found to be 5.8% by weight of the aggregate. Eight mixes with various combinations of LS, BWP, and PF were evaluated in this investigation, as shown in

Table 6. Different mixes used in the study.

Mix ID	Description of Mix Composition
ACLS	AC mix containing LS as conventional filler
ACLSF3	AC mix containing LS as conventional filler and PF with a length of 3 mm
ACLSF8	AC mix containing LS as conventional filler and PF with a length of 8 mm
ACLSF15	AC mix containing LS as conventional filler and PF with a length of 15 mm
ACBWP	AC mix containing BWP as a replacement for conventional filler
ACBWPF3	AC mix containing BWP as a replacement for conventional filler and PF with a length of 3 mm
ACBWPF8	AC mix containing BWP as a replacement for conventional filler and PF with a length of 8 mm
ACBWPF15	AC mix containing BWP as a replacement for conventional filler and PF with a length of 15 mm

. The same asphalt content of 5.8% was selected for all mixes studied. The Algerian specifications [56] for the required limits of certain properties of semi-coarse asphalt concrete 0/14 (SCAC 0/14) as used in this work, determined by the Marshall test, ITS test, and water sensitivity test, are given in

Table 7. Specification limits for certain characteristics of AC used in this study.

Property	Specification [56]
Marshall stability (kN)	Min. 10.5 kN
Va (%)	3–5
Marshall flow (mm)	Max. 4 mm
Indirect tensile strength ratio, ITSR (%)	Min. 75%

.

2.3. Test Procedures

2.3.1. Marshall Test

The Marshall test was performed according to EN 12697-34 [66]. Cylindrical, compacted specimens with a diameter of 101.6 mm and a height of 63.5 mm were placed in a 60 °C water bath for 40 min, then loaded at a constant deformation rate of 50 mm/min. Marshall stability (in kN) represents the maximum resistance to deformation, while flow (in mm) indicates deformation at maximum load. The Marshall quotient is calculated as the ratio of stability to flow (in kN/mm). For each mix type, the following properties were determined: Marshall quotient, Marshall flow, Marshall stability, VFAs, bulk density, VMAs, and Vas. For each AC mix, three replicate Marshall specimens were prepared and tested to ensure statistical reliability of stability and flow values.

2.3.2. Indirect Tensile Strength (ITS) Test

The indirect tensile strength (ITS) test was carried out at a standard temperature of 25 °C in accordance with EN 12697-23 [67]. The cylindrical specimen was subjected to a diametral load along the cylinder axis, at a constant deformation rate of 50 mm/min until failure. The ITS is the maximum tensile stress computed from the maximum load applied to failure and the specimen dimensions. The ITS test was also performed on three replicate specimens per mix type, and the average value was reported. The following equation was used to calculate the ITS:

$$\mathrm{I}\mathrm{T}\mathrm{S}={\displaystyle \frac{2\mathrm{P}}{\mathsf{\pi }\mathrm{D}\mathrm{H}}}$$

(1)

where ITS is the indirect tensile strength in kPa, P is the maximum load in kN, D is the diameter of the specimen in m, and H is the specimen height in m.

2.3.3. Water Sensitivity Test

The water sensitivity test was conducted as per EN 12697-12 [68]. Six cylindrical specimens of each mix type were divided into two equal sets. The dry set was stored on a flat surface at room temperature (20 ± 5) °C, while vacuum saturation was conducted at an absolute pressure of 6.7 kPa for 30 min to ensure sufficient moisture penetration in the specimen pores. Wet specimens were then immersed in a water bath maintained at 40 ± 1 °C for 72 h, followed by temperature equilibration at 25 ± 1 °C for 2 h before ITS testing. These temperature conditions were carefully controlled to simulate in-service pavement exposure and ensure repeatable moisture conditioning. The ITS of the samples was then measured according to the procedure outlined in EN 12697-23 [67]. The moisture sensitivity of AC mixes can be evaluated using the indirect tensile strength ratio (ITSR), expressed as a percentage, and is calculated as follows:

$$\mathrm{I}\mathrm{T}\mathrm{S}\mathrm{R}=\left({\displaystyle \frac{{\mathrm{I}\mathrm{T}\mathrm{S}}_{\mathrm{w}}}{{\mathrm{I}\mathrm{T}\mathrm{S}}_{\mathrm{d}}}}\right)\times 100$$

(2)

where ITS_w is the average ITS of the wet set in kPa and ITS_d is the average ITS of the dry set in kPa.

2.4. Variables and Mixtures by RSM

By applying the response surface methodology (RSM), we manipulated eight (8) combinations by adjusting the input parameters, resulting in ten output results. RSM generated eight (8) distinct configurations of factors, encompassing ranges from 1 to 2 for the filler type (limestone powder (LS) and brick waste powder (BWP)), as well as the length of polyester fibers (3, 8, and 15 mm).

Table 8. Input parameters and mixtures.

Mix Type	Filler Name	Filler Type	PF Length (mm)
ACLS	Limestone powder (LS)	1	0
ACLSF3		1	3
ACLSF8		1	8
ACLSF15		1	15
ACBWP	Brick waste powder (BWP)	2	0
ACBWPF3		2	3
ACBWPF8		2	8
ACBWPF15		2	15

displays the mixes generated by RSM and the parameters used in the production of the corresponding test samples for each mix.

3. Results and Discussion

3.1. Marshall Properties

Figure 8 summarizes the mechanical and volumetric characteristics as well as the moisture susceptibility (based on ITSR) of asphalt concrete (AC) mixtures containing limestone filler (LS), brick waste powder (BWP), and polyester fibers (PFs) of varying lengths (3 mm, 8 mm, and 15 mm). The results encompass essential metrics like bulk density, air voids (Vas), voids in mineral aggregate (VMAs), voids filled with asphalt (VFAs), Marshall stability, Marshall flow, Marshall quotient, indirect tensile strengths, and tensile strength ratio (ITSR). Each data point in the picture signifies the mean of three independently produced and tested specimens, while the accompanying error bars denote the standard deviation to illustrate the experimental heterogeneity within each group. This graphical depiction offers a comparison framework to evaluate the impact of filler type and fiber length on the internal structure, moisture sensitivity, and mechanical properties of the AC mixes.

Marshall test results for AC mixes are given in Figure 8a–g. Figure 8a displays the bulk density values for all mixes. The ACLS mix was selected as the reference mix. This type of mix has been extensively used in Algeria for the construction of the wearing course of flexible pavements. In comparison to the ACLS mix, the bulk density value of the ACBWP mix decreased by 0.89%. Choudhary et al. [69] also found this decrease. The observed decrease in bulk density of AC when substituting LS with BWP as a filler can be explained by the higher specific gravity of LS compared to that of BWP. As specific gravity more particularly affects the overall density of the asphalt mix, the use of BWP as a lighter filler contributed towards a lower bulk density of the mix. However, when PF was used in combination with LS and BWP, the bulk density values of the AC mixes decreased further as the length of the PF increased. This reduction might be further attributed to the specific gravity of the fiber, which is lower than that of the aggregate materials. In addition, the fiber length may also be longer, hence may contribute to increased air voids within the mix, thus resulting in lower bulk density.

Void is an essential property in AC mixes, as it is directly proportional to the total air voids and inversely related to the amount of asphalt binder [70]. The Vas for all the blends that were studied are presented in Figure 8b. The Vas value of the ACBWP mix is 5.56% higher than that of the ACLS mix. These findings are consistent with the results of the study by Arabani et al. [25]. The increase in Vas observed in an AC mix containing BWP is likely due to its lower specific gravity compared to LS. It is for this reason that at a given volume BWP has a lower density, which could lead to poor packing and compaction within the mix, hence increasing the Vas. In addition, as the length of the PFs increased, the Vas values also increased when PFs were incorporated into the AC mix with LS and the AC mix with BWP. This can be related to the fact that the specific gravity of PFs is less than that of the aggregates, and thus there are more voids in the mix. Additionally, longer fibers tend to ball together, disrupting the compaction process and further increasing the Vas. The ACLS, ACBWP, ACLSF3, and ACBWPF3 mixes had air voids within the specification limits of 3% to 5%, while the other mixes exceeded these limits.

VMAs are the voids between the particles of the aggregate in the mix, including voids that are filled with asphalt binder [22]. The VMAs of the mixes are illustrated in Figure 8c. As seen in the figure, the VMAs value for the AC mix containing BWP as a filler was slightly higher by 1.23% than that of the AC mix containing LS. This trend is consistent with the findings of Choudhary et al. [69]. The increase in VMAs might be caused by the higher porosity of BWP particles compared to LS, resulting in greater bitumen absorption in the mix. Moreover, when PFs were combined with LS and BWP, the VMA values of the AC mixes increased with increasing PF lengths. This is possibly due to the decrease in bulk density of the AC mix discussed above (a lower bulk density leads to higher VMAs) [31].

VFAs refers to the percentage of VMAs that are occupied by asphalt binder [71]. Figure 8d displays the values of VFA for all mixes. It can be observed that using BWP as a filler in the AC mix resulted in a 2.15% reduction in the VFA value compared to the AC mix with LS. These findings correspond with previous studies performed by Arabani et al. [25]. This decrease in VFA may be attributed to the higher asphalt binder absorption of BWP compared to LS, which results in less binder available to fill the voids in the mix. In addition, the values of VFA decreased with increasing PF lengths when PFs were added to the ACLS and ACBWP mixes. The reduction in VFA indicates a decrease in the effective asphalt film thickness between aggregates due to the effects of the fibers on asphalt absorption [31].

Marshall stability refers to the capacity of AC mixes to withstand deformation when subjected to applied loads [69]. Figure 8e depicts the Marshall stability values for all mixes studied. The replacement of LS with BWP resulted in a slight increase of 3.24% in the Marshall stability value of the AC mix in comparison to the mix containing LS. This enhancement aligns with the findings of Arabani et al. [25]. The increase in Marshall stability can be attributed to the lower specific gravity of BWP compared to LS, which enables it to occupy more volume within the mix. This increases the extent of the surface area that would interact with the asphalt and hence increases the adsorption of the asphalt binder. As a result, stronger adhesion between the asphalt and aggregates is promoted, improving the load-bearing capacity of the AC mix. When PFs were added to the ACLS and ACBWP mixes, the Marshall stability values initially increased with an increase in fiber length, reaching peak values at 8 mm (12.58 kN, a 16.59% increase, and 12.82 kN, an 18.81% increase, compared to the control mix). However, beyond this point, the Marshall stability decreased as the fiber length further increased to 15 mm, resulting in lower stability values of 10.12 kN and 10.66 kN. The improvement in Marshall stability of AC may be attributed to the fiber’s adhesion and networking effects. On the other hand, the reduction in stability with longer fiber length may result from poor fiber dispersion and fewer contact points between aggregates, leading to a less effective structure in the mix. These results suggest that the impact of PFs at optimal lengths on Marshall stability is more pronounced when used with BWP than with LS. The Marshall stability values for all mixes were greater than 10.5 kN, which is the minimum requirement, except for the ACLSF15 mix.

Flow is defined as the overall movement or strain that occurs in the specimen during the Marshall test, from no load to maximum load, and it represents the plasticity and flexibility of AC mixes [72]. Figure 8f shows the values of the Marshall flow for all the mixes. The minimum flow value indicates the brittle behavior of the asphalt mix, whereas the maximum flow value signifies its highest plasticity [73]. The type of filler directly affects the stiffness of the asphalt mix [74]. As illustrated in the figure, the flow value of the AC mix containing BWP increased by 14.01% compared to the mix with LS. This indicates a significant reduction in the stiffness of the AC when BWP was used as the filler material. The results obtained by Arabani et al. [25] also indicated that adding brick powder as a filler to the AC mix resulted in increasing the Marshall flow value. Moreover, with the inclusion of PFs in the ACLS and ACBWP mixes, and with the increase in PF lengths, Marshall flow values increased dramatically. The increase in Marshall flow may be due to the longer length of PFs, which causes poor dispersion in the mix, fiber entanglement, and reduced bonding between the asphalt and aggregates. This disruption in the mix structure leads to a less stable AC, allowing for greater flow underload. The ACLS, ACBWP, and ACLSF3 mixes met the requirement for Marshall flow. It can be concluded that the ACBWPF15 mix had the maximum plasticity behavior, while the ACLSF3, ACBWP, and ACLS mixes were more brittle.

A higher Marshall quotient value indicates that the AC mix is stiffer and less prone to permanent deformation (rutting) [75]. The values of the Marshall quotient for all the mixes are depicted in Figure 8g. The Marshall quotient of the AC containing BWP was 9.40% lower than that of the mix with LS. This reduction was also observed in the results of Choudhary et al. [69]. Notably, the AC mix with BWP exhibited a higher VMAs value (17.23%) than the mix containing LS, which may account for its lower MQ value. Previous studies have shown an inverse correlation between the rutting resistance of AC mixes and their VMAs [76,77]. In addition, the Marshall quotient values reduced with increasing PF lengths when PFs were incorporated into the ACLS and ACBWP mixes. This finding indicates that the addition of PFs to the AC reduced the Marshall quotient, with longer fibers causing a further reduction.

3.2. Cracking Properties (Indirect Tensile Strength)

The tensile properties of the AC mix influence the cracking potential of pavement [78]. Figure 8h illustrates the ITS for all the mixes studied. It is evident that substituting LS with BWP resulted in a 14.07% increase in the ITS value of the AC mix compared to the mix containing LS. This difference may be due to the greater surface contact area of BWP grains, which enhances the ITS of the AC mix more effectively than LS [24]. It is worth noting that BWP has a lower specific gravity than LS, meaning it occupies more volume and has a larger specific surface area. Another research study by Chen et al. [26] reported that the ITS value was enhanced by including recycled brick powder as a filler in AC. Incorporating PFs into ACLS and ACBWP mixes generally further enhances the ITS values. This improvement is primarily due to the ability of polyester fibers to form a 3D network within the AC mix. This network helps resist crack propagation and prevents aggregate sliding at the interface, distributes stress more evenly, and reduces stress concentration, ultimately improving the overall tensile strength of the asphalt mix. When PFs were used with LS, the ITS of the AC mixes marginally increased for fiber lengths of 3 mm and 8 mm but decreased for the longer fiber length of 15 mm. This decline in ITS could be explained by poor distribution of the fibers due to their longer length. In contrast, when PFs were used in conjunction with BWP, the ITS values of the AC mixes substantially improved, with the highest increase of 38.39% observed at an 8 mm fiber length, suggesting a synergistic effect between BWP and PFs. Overall, BWP combined with PFs at optimal lengths offers superior performance.

3.3. Mechanistic Insights into Material Interactions

The enhancements noted in the mechanical and moisture resistance characteristics of asphalt concrete (AC) mixtures containing brick waste powder (BWP) and polyester fiber (PF) can be ascribed to the synergistic interactions between these elements and the bituminous matrix. BWP serves as a chemically reactive and structurally beneficial filler, primarily owing to its elevated concentrations of silica (SiO₂), alumina (Al₂O₃), and ferric oxide (Fe₂O₃), which impart pozzolanic activity—an attribute that improves the long-term strength and durability of cementitious and bituminous systems [79,80]. These oxides promote physicochemical bonding with the polar groups in bitumen, leading to improved filler–binder adhesion [81]. Unlike traditional limestone fillers, BWP actively engages in mastic-phase interactions instead of only serving as an inert volumetric substitute.

The angular morphology and porous surface texture of BWP particles at the microstructural level substantially enhance the accessible surface area for bitumen interaction [82]. This results in enhanced mechanical interlock and binder absorption, hence increasing cohesion, stiffness, and indirect tensile strength in the asphalt mixture [79]. These advancements are essential for resisting deformation underload and reducing moisture susceptibility. The incorporation of BWP may influence the flowability and setting characteristics of the asphalt mastic, necessitating modification in the mix design to preserve compaction quality and workability [82]. The diversity in the mineralogical content of brick debris requires stringent quality control to maintain material uniformity across production batches [83].

Polyester fibers (PFs), obtained from recycled polyethylene terephthalate (PET), function as efficient reinforcements in the AC matrix. Their elevated tensile strength and elongation potential allow them to span micro-cracks and impede crack propagation, thus enhancing fatigue resistance and resistance to permanent deformation under cyclic traffic stress [84]. When effectively dispersed, PFs establish a secondary load-bearing network that limits aggregate displacement and enhances structural durability, especially under dynamic loads. The fiber-bitumen contact is crucial for load-transfer efficiency and the mechanical integrity of the composite system.

Fiber length and dispersion are essential for performance: medium-length PFs (5–8 mm) are optimum for balancing mechanical interlocking and dispersion homogeneity. These fibers possess adequate length to establish strong linkages between aggregates without causing clumping or tangling, which may undermine mixture homogeneity and compaction [84]. Moreover, the surface treatment of PFs, encompassing chemical or physical alteration, can markedly enhance their compatibility with bitumen, hence improving interfacial bonding and facilitating more effective stress transfer. Enhanced adhesion at the fiber–matrix interface results in greater resistance to cracking, rutting, and moisture damage, particularly under elevated temperature and strain conditions [85].

Notwithstanding these advantages, certain problems persist in the incorporation of PFs into asphalt mixtures. Concerns like fiber dispersion variability, the ecological consequences of synthetic fibers, and the enhancement of surface treatments necessitate additional research [86,87]. Nonetheless, the synergistic application of BWP and PFs establishes a multifunctional reinforcement system, with BWP augmenting the chemical and structural attributes of the mastic, while PFs provide tensile reinforcement and crack-bridging capabilities. This synergy results in an AC composite that is not only more durable and moisture-resistant but also more sustainable, advancing circular economy goals by valorizing industrial and plastic waste.

3.4. Water Sensitivity (Indirect Tensile Strength Ratio)

Indirect tensile strengths of wet and dry samples of AC mixes are presented in Figure 8i and Figure 8h, respectively. The ITSd and ITSw values of AC mixes with BWP increased by 14.07% and 19.17%, respectively, compared to those with LS. This increase might result from the enhanced binding properties of BWP compared to LS, leading to improved strength in the asphalt mixes. Furthermore, the use of PFs in conjunction with BWP significantly improved both the dry and wet indirect tensile strength (ITSd and ITSw) of AC mixes, with strength increasing as the fiber length increased. However, when PFs were combined with LS, ITSd showed minimal improvement or a decrease at longer fiber lengths, while ITSw showed moderate improvements. This finding indicates that BWP has a stronger positive effect on the performance of AC mixes when combined with PFs.

Figure 8j illustrates ITS ratios for AC mixes. A higher ITSR value suggests that the AC mix is more resistant to moisture damage. All mixes satisfied the minimum ITSR requirements. It can be noticed that the ITSR for the ACBWP mix showed an increase of 4.42% when compared to the ACLS mix, indicating that the AC mix with BWP exhibited better resistance against moisture damage than the AC mix with LS. This trend aligns with findings from several previous studies [21,23]. The improved moisture resistance observed in the AC mix containing BWP might be due to its higher pozzolanic activity compared to LS. The significant amount of silica (SiO₂) and the relatively low amounts of alumina (Al₂O₃) and iron oxide (Fe₂O₃) in BWP contribute to its enhanced pozzolanic properties. These properties promote stronger adhesion between the aggregates and the asphalt binder. As a result, improved adhesion helps to reduce the effects of moisture, leading to better resistance to striping and greater overall durability of the asphalt mix. The addition of PFs to the AC with LS significantly enhanced its water damage resistance, with ITSR increases of 7.26%, 9.79%, and 7.47% for fiber lengths of 3 mm, 8 mm, and 15 mm, respectively. The 8 mm PF length achieved the highest ITSR improvement of 9.79%, indicating the optimal performance with LS. In contrast, when PFs were added to AC with BWP, the ITSR values increased by 9.47%, 0.95%, and 4.53% for the same fiber lengths. The maximum ITSR improvement for BWP was 9.47% at 3 mm PF, indicating a lower but still significant improvement in water resistance. These results highlight the potential of PFs to enhance the durability of AC mixes when combined with both LS and BWP.

The enhanced performance of the 8 mm polyester fiber (PF) length is chiefly due to its ideal equilibrium between mechanical reinforcement and dispersion efficacy in the asphalt matrix. Fibers of this intermediate length are sufficiently long to provide a continuous load transfer network, thereby increasing internal friction and facilitating effective interlocking among aggregate particles. This network effect enhances stress distribution and crack-bridging abilities under load. Concurrently, the 8 mm length is sufficiently brief to reduce the likelihood of fiber entanglement or clumping, prevalent problems linked to longer fibers that can adversely impact compaction and uniformity. The optimal dispersion and structural reinforcement at this fiber length likely explain the maximum values seen in both Marshall stability and indirect tensile strength (ITS), signifying improved mechanical integrity and resistance to deformation [38,88].

3.5. Response Surface Methodology (RSM)

3.5.1. Model Development and ANOVA

The goal of using response surface methodology (RSM) in the preliminary phases of the design of experiments (DOE) is to develop predictive models of responses and perform optimization [42,43,89,90]. These response models, as illustrated in the generalized formats of Equations (3) and (4), can be the following linear or higher-order polynomials [91,92,93]:

y = β₀ + β₁x₁ + β₂x₂ + β_nx_n + ϵ

(3)

$$y={\mathsf{\beta }}_0+{\sum }_i^k{\beta }_i{x}_i+{\sum }_i^k{\beta }_{ii}{x_i}^2+{\sum }_{ij}^k{\beta }_{ij}.x_i.x_j+ϵ$$

(4)

where y is the desired response, β₀ denotes the coefficient of the regression for the constant term, β_i, β_ii, and β_ij represent the coefficients for the linear, quadratic, and interaction of the x_i and x_j terms, respectively, the random error is represented by ϵ, and the number of factors is represented by k.

RSM provides various modeling approaches, each with distinct characteristics and levels of accuracy. The precision of predictive results is not solely determined by the model type selected; it is also heavily influenced by the quality and relevance of the experimental data. High-quality, relevant experimental data are critical to ensuring accurate model predictions. In this study, the quadratic model was chosen to represent the responses, specifically bulk density (g/cm³), Vas (%), voids in mineral aggregate (VMAs) (%), VFAs (%), Marshall stability (kN), Marshall flow (mm), Marshall quotient (kN/mm), ITS (kPa), indirect tensile strength of wet specimen ITSw (kPa), and indirect tensile strength ratio ITSR (%). This model was selected for its superior accuracy compared to other alternatives. The strength of the quadratic model lies in its ability to account for nonlinear effects and complex interactions between input variables, which are vital for achieving reliable predictions when relationships among the variables are not purely linear.

These models are expressed in coded terms in Equations (5)–(14). Responses at various levels for each variable can be predicted by using equations expressed in coded factors. Generally, a value of −1 indicates low levels by default, whereas a value of +1 indicates high levels of a factor. The coded equation makes it easy to determine the variables’ relative importance through comparisons of the coefficients of the factors, denoted as A: (Type of filler) and B: (Polyester fiber length, mm). The equations are as follows:

Bulk density (g/cm³) = + 2.33 − 0.0085A − 0.0095B + 0.0052B²

(5)

Va (%) = + 5.32 + 0.06A + 0.39B + 0.003AB − 0.21B²

(6)

VMA (%) = + 18.33 + 0.04A + 0.34B − 0.19B²

(7)

VFA (%) = + 68.22 − 0.4A − 1.55B + 0.01AB + 0.89B²

(8)

Marshall stability (kN) = + 12.62 + 0.3A − 0.75B − 0.11AB − 1.48B²

(9)

Marshall flow (mm) = + 5.37 + 0.34A + 0.98B + 0.005AB − 0.3B²

(10)

Marshall quotient (kN/mm) = + 2.37 − 0.1A − 0.6B + 0.02AB − 0.05B²

(11)

ITS (kPa) = + 1234.25 + 125.7A − 24.75B + 21.67AB − 121.5B²

(12)

ITS of wet specimen (kPa) = + 1225.54 + 107.89A − 39.25B + 10.14AB − 92.79B²

(13)

ITS ratio (%) = + 99.86 − 1.57A − 1.125B − 0.99AB + 1.97B²

(14)

Analysis of Variance (ANOVA) is a statistical technique used extensively in research for the analysis and interpretation of experimental data. It is based on probability theory and mathematical statistics, aiming to validate models and assess the impact of input parameters on the variation in responses. An ANOVA tests the significance of observed differences between groups or treatments by partitioning the total variance into distinct components attributed to different sources, such as independent variables or interactions. This partitioning allows researchers to determine whether the observed differences are statistically significant or merely due to random chance [44,50,53]. An ANOVA is typically conducted with a 95% confidence interval, denoted by a significance level (α) of 0.05, which means that outcomes with a p-value that is below 0.05 are considered to be statistically significant [94,95,96]. The statistical parameters considered by the ANOVA are given in

Table 9. Statistical parameters of ANOVA.

Statistical Parameter	Equation	Definition
The squared sum (SSf)	SSf = ${\displaystyle \frac{N}{N_{n_f}}}{\sum }_{i=1}^{N_{n_f}}{\left({\overline{y}}_i-\overline{y}\right)}^2$	To estimate the deviation square from the overall average: $\overline{y}$: the average response, ${\overline{y}}_i$: mean of the measured responses for each level i of the F-factor, N denotes the total number of trials, $N_{n_f}$: number of levels for each factor f.
The squared mean (MSi)	$M{S}_i={\displaystyle \frac{S{S}_i}{{dl}_i}}$	The squared sum (SSi) divided by the number of degrees of freedom (dli) yields (MSi).
The F-value	$F_i={\displaystyle \frac{{MS}_i}{{MS}_e}}$	Is used to check that the mathematical model is compatible, since the computed F-values must be greater than the tabulated F. MSe is the mean squared sum of the errors.
Contribution (Cont.%)	$Cont.\%={\displaystyle \frac{{SS}_f}{{SS}_T}}\times 100$	It shows the factors’ contribution (SSf) to the total variance (SST), indicating the degree of percent effect on response.
The coefficient of determination (R2)	R2 = ${\displaystyle \frac{\sum {(y_i-\overline{y})}^2}{\sum {({\overline{y}}_i-\overline{y})}^2}}$	Goodness of fit is measured as the ratio of explained variation to total variation.

. (1) The parameter is significant if the p-value ≤ 0.05; (2) the parameter is insignificant if the p-value > 0.05.

A similar methodological framework was applied by Nouri et al. [97], who integrated optimization and an ANOVA to evaluate the significance of key parameters in glass fiber polymer–reinforced concrete beams. Their research illustrated how an ANOVA may proficiently discern predominant components and interaction effects in intricate material systems, affirming its use in optimization-driven experimental designs. This validates the application of an ANOVA in the current study to evaluate the statistical significance of input variables and the model’s appropriateness in forecasting asphalt mix performance.

Table 10. ANOVA for all AC characteristics studied.

Source	SSf	Df	MSi	F-Value	p-Value	Cont. %	Significant
Bulk density (g/cm3)	0.0008	2	0.0004	34.13	0.0086
A-Filler type	0.0004	1	0.0004	38.1	0.0086	50	Yes
B-PF length (mm)	0.0003	1	0.0003	30.16	0.0119	37.5	Yes
Residual	0	3	0
Cor Total	0.0008	5
Va (%)	0.5903	2	0.2952	15.62	0.0259
A-Filler type	0.0193	1	0.0193	1.02	0.387	2.982998454	No
B-PF length (mm)	0.571	1	0.571	30.22	0.0118	88.25347759	Yes
Residual	0.0567	3	0.0189
Cor Total	0.647	5
VMA (%)	0.4361	2	0.2181	15.14	0.0271
A-Filler type	0.0096	1	0.0096	0.6667	0.474	2.002920926	No
B-PF length (mm)	0.4265	1	0.4265	29.62	0.0122	88.98393491	Yes
Residual	0.0432	3	0.0144
Cor Total	0.4793	5
VFA (%)	10.07	2	5.04	15.16	0.027
A-Filler type	0.96	1	0.96	2.89	0.1877	8.672086721	No
B-PF length (mm)	9.11	1	9.11	27.43	0.0136	82.29448961	Yes
Residual	0.9965	3	0.3322
Cor Total	11.07	5
Marshall stability (kN)	3.3	2	1.65	1.71	0.3187
A-Filler type	0.546	1	0.546	0.5667	0.5063	8.806451613	No
B-PF length (mm)	2.76	1	2.76	2.86	0.1892	44.51612903	No
Residual	2.89	3	0.9635
Cor Total	6.2	5
Marshall flow (mm)	4.45	2	2.23	37.24	0.0076
A-Filler type	0.6801	1	0.6801	11.38	0.0433	14.68898488	Yes
B-PF length (mm)	3.77	1	3.77	63.09	0.0042	81.42548596	Yes
Residual	0.1793	3	0.0598
Cor Total	4.63	5
Marshall quotient (kN/mm)	1.51	2	0.7557	57.93	0.004
A-Filler type	0.0561	1	0.0561	4.3	0.1298	3.619354839	No
B-PF length (mm)	1.46	1	1.46	111.57	0.0018	94.19354839	Yes
Residual	0.0391	3	0.013
Cor Total	1.55	5
ITS (kPa)	96,944.5	2	48,472.25	5.27	0.1042
A-Filler type	93,001.5	1	93,001.5	10.12	0.0501	74.4012	Yes
B-PF length (mm)	3943	1	3943	0.429	0.5592	3.1544	No
Residual	27,573	3	9191
Cor Total	1.25 × 105	5
ITSw (kPa)	77,015.15	2	38,507.58	9.87	0.0479
A-Filler type	69,122.67	1	69,122.67	17.72	0.0245	77.91454754	Yes
B-PF length (mm)	7892.48	1	7892.48	2.02	0.25	8.896343388	No
Residual	11,700.85	3	3900.28
Cor Total	88716	5
ITSR (%)	18	2	9	0.8324	0.5157
A-Filler type	13.8	1	13.8	1.28	0.3407	27.36466389	No
B-PF length (mm)	4.2	1	4.2	0.3883	0.5774	8.328375967	No
Residual	32.44	3	10.81
Cor Total	50.43	5

presents the ANOVA findings for the AC properties.

The respective contributions of 50% and 14.688% of the filler type (LS and BWP) and 37.5% and 81.425% of the polyester fiber length are significant (p ≤ 0.05) for bulk density and Marshall flow, respectively. The respective contributions of 8.806% and 27.364% of the filler type (LS and BWP) and 44.516% and 8.328% of the polyester fiber length are insignificant (p > 0.05) for Marshall stability and ITS ratio, respectively. Additionally, the contributions of polyester fiber lengths of 88.253%, 88.983%, 82.294%, and 94.193% are more significant (p ≤ 0.05) compared to fillers of 2.982%, 2.002%, 8.672%, and 3.619%, which are insignificant (p > 0.05), for VAs, VMAs, VFAs, and the Marshall quotient, respectively. On the other hand, the contributions of fillers of 74.401% and 77.914% are more significant (p ≤ 0.05) compared to polyester fiber lengths of 3.154% and 8.896%, which are insignificant (p > 0.05), for the ITS in wet specimens.

The perturbation diagram is a commonly used graphical tool in engineering to visualize the effects of different factors on the output of interest. It aids in identifying and analyzing how variations or disturbances in input variables influence the overall output. In this diagram, the input variables are normalized and represented on a scale from −1 to +1. This normalization allows for easier comparison and analysis of the input–output relationships, regardless of the original units or scales of the variables. The diagram in Figure 9 illustrates how the system output reacts to simultaneous variations in the binary inputs A and B. It highlights which input combinations have the greatest effect on the output and reveals potential interactions between the variables. Notably, positive factor effects correspond to an increase in the model response, while negative effects lead to a decrease.

The bulk density increased considerably with a sharp increase in factors A and B to the left (level −1). In contrast, the ratio of ITS increased in both directions: quickly in the negative direction (−1) and slightly in the positive direction (+1). VAs as well as VMAs increased with a small increase in factor A and a large increase in factor B on the right side (level +1). Marshall flow increased with a strong increase in factors A and B on the right side (level +1). Furthermore, the responses (ITS of wet specimens, Marshall stability, as well as indirect tensile strength) increased as factor A increased to the right (level +1), while factor B increased near the reference point to the left side (level −1). The Marshall quotient and VFAs increased with a slight increase in factor A and a significant increase in factor B on the negative side (level −1).

3.5.2. The 2D Contours and 3D Response Surfaces

Response surface graphs (3D) and contour plots (2D) highlight the impact of variables such as filler type and PF length on the characteristics of AC, as illustrated in Figure 10. A contour plot, whether in 2D or 3D, presents the same data as Figure 8 but provides a clearer and more precise visual representation to express experimental results. For 2D contour plots as well as 3D response surface graphs, color coding allows identification of areas of the surface graph corresponding to different response intensities resulting from the interaction of input variables. Red zones indicate regions with the highest response intensities, while blue zones represent the lowest response intensities.

3.5.3. Multi-Objective Optimization of Response Using Desirability Functions (DFs)

Desirability functions (DFs) are widely used for optimizing multiple responses simultaneously. The fundamental idea behind this approach is to transform a multi-response problem into a single-response problem using mathematical transformations. For each response Y(x), where j = 1,2,…,m, a desirability function dj(Yj(x)) is defined with values ranging from 0 to 1. A value of 0 represents an undesirable response, while a value of 1 indicates optimal performance for the specified factors [98]. This study aims to use desirability functions (DFs) to determine the optimal values of filler type and PF length (mm) that maximize the average Nusselt number for each tested AC sample. To tackle this parameter design problem, the objective function F(x) is defined as follows [99]:

$$DF={\left({\prod }_{i=1}^n{d}_i^{wi}\right)}^{{\displaystyle \frac{1}{{\sum .}_{j-1}^{n}wi}}}$$

(15)

$$F\left(x\right)=-DF$$

(16)

where di represents the desirability linked to the target output, while wi is the weighting factor associated with di. The formulation of di may differ depending on the specific goal for each target output. For instance, if the objective is to achieve a specific value for Ti, the desirability di is expressed as follows:

$$di=0\mathrm{If}{Y}_i\le ln{f}_i$$

$$di=\left[{\displaystyle \frac{Yi-{Inf}_i}{T_i-{Inf}_i}}\right]\mathrm{If}ln{f}_i\le Y_i\le T_i$$

(17)

$$di=\left[{\displaystyle \frac{Yi-{Sup}_i}{T_i-{Sup}_i}}\right]\mathrm{If}{T}_i\le Y_i\le {Sup}_i$$

$$di=0\mathrm{If}{Y}_i\ge {Sup}_i$$

The desirability function for maximizing an objective is defined as:

$$di=0\mathrm{If}{Y}_i\ge {Inf}_i$$

$$di=\left[{\displaystyle \frac{Yi-{Inf}_i}{{Sup}_i-{Inf}_i}}\right]\mathrm{If}{Inf}_i\le Y_i\le {Sup}_i$$

(18)

$$di=1\mathrm{If}{Y}_i\ge {Sup}_i$$

Table 11. Optimization conditions.

Parameters	Lower Limit	Upper Limit
Filler Name	LS	BWP
PF length (mm)	0	15

shows the factor limits used in the optimization.

Optimization Results by DF

Table 12. Optimization criteria and solutions by DF.

	Parameters	Unit	Notation	Goal	Optimization Result	Desirability Value (D)
Input Factors	Filler Type	/	A	Maximize	2	0.718 (71.8%)
	PF length	mm	B	Target	5
Responses (Output Factors)	Bulk density	g/cm3	/	Maximize	2.32992
	Vas	%	/	Minimize	4.97694
	Voids in mineral aggregate	%	VMA	Minimize	18.0299
	Voids filled with asphalt	%	VFA	Maximize	69.4058
	Marshall stability	kN	/	Maximize	12.7528
	Marshall flow	mm	/	Minimize	4.82938
	Marshall quotient	kN/mm	/	Maximize	2.66429
	Indirect tensile strength	kPa	ITS	Maximize	1298.18
	Indirect tensile strength of wet specimen	kPa	ITSw	Maximize	1305.68
	Indirect tensile strength ratio	%	ITSR	Maximize	100.859

summarizes the optimization results obtained through the application of DFs. We selected the following objectives: two “BWP” filler types and a polyester fiber length of “5 mm” to match the experimental value range. The system was configured to maximize the bulk density (g/cm³), Marshall quotient (kN/mm), VFAs (%), ITS (kPa), Marshall stability (kN), and ITS of wet specimen (kPa), as well as ITS ratio (%), while minimizing the Vas (%), Marshall flow (mm), and VMAs (%). After carrying out an optimization process, it was found that the bulk density, ITS ratio, VFAs, ITS of wet specimen, Marshall quotient, ITS, and Marshall stability can be maximized at 2.32992 g/cm³, 100.859%, 69.4058%, 1305.68 kPa, 2.66429 kN/mm, 1298.18 kPa, and 12.7528 kN, respectively. Furthermore, it was determined that the voids in mineral aggregate, Vas, and Marshall flow can be minimized to 18.0299%, 4.97694%, and 4.82938 mm, respectively. These objectives can be achieved by adjusting the filler type and PF length to optimal levels of two (BWP) and 5 mm, respectively. The desirability value (D) of the optimization was set at 71.8%, which is high given the complexity that is required to meet the multi-objective criteria. The graphical representation of the optimization solutions from one to five is presented as ramps in Figure 11, while Figure 12 depicts a 3D response surface diagram for desirability (D).

3.5.4. Validation of Optimization Results

The validation process plays a key role in optimization, ensuring the reliability and robustness of solutions proposed by desirability functions (DFs). Following DF optimization to determine potential solutions for enhancing the responses, it is imperative to test and assess these solutions. This step identifies potential errors or inconsistencies that might impact performance and confirms the suitability and dependability of the optimized solutions for their intended applications. The last step of the RSM analysis includes experimental validation. In this phase, test samples are manufactured using optimal input factors determined by multi-objective optimization, selected from the DF-proposed solutions.

These samples are then subjected to tests to assess their performance. The findings obtained as well as the predicted optimal results are presented in

Table 13. Experimental validation of results.

Parameters	Unit	Notation	Goal	Model Prediction	Laboratory Experiment	Absolute Error, %	Desirability Value (D)
Filler Type	/	A	Maximize	2	2	-	0.718 (71.8%)
PF length	mm	B	Target	5	5	-
Bulk density	g/cm3	/	Maximize	2.32992	2.21	5.43
Vas	%	/	Minimize	4.97694	4.65	7.03
Voids in mineral aggregate	%	VMA	Minimize	18.0299	17.89	0.78
Voids filled with asphalt,	%	VFA	Maximize	69.4058	68.77	0.92
Marshall stability	kN	/	Maximize	12.7528	12.23	4.27
Marshall flow	mm	/	Minimize	4.82938	4.53	6.61
Marshall quotient	kN/mm	/	Maximize	2.66429	2.56	4.07
Indirect tensile strength	kPa	ITS	Maximize	1298.18	1297.85	0.025
Indirect tensile strength of wet specimen	kPa	ITSw	Maximize	1305.68	1304.62	0.081
Indirect tensile strength ratio	%	ITSR	Maximize	100.859	100.52	0.34

. To compare these two sets of results, Equation (19) [62,63] is used to calculate the absolute experimental error. The predictive model developed using RSM and desirability functions (DFs) exhibited excellent reliability and precision, with an error margin of only 8%. This small margin of error emphasizes the robustness of the DF approach, showcasing its capability to provide accurate and dependable results. These findings affirm the effectiveness of the applied methodology and its potential for broader use in similar predictive analyses and practical applications.

$$\mathrm{Absolute}\mathrm{Experimental}\mathrm{Error},\%=(1-{\displaystyle \frac{\mathrm{P}\mathrm{r}\mathrm{e}\mathrm{d}\mathrm{i}\mathrm{c}\mathrm{t}\mathrm{e}\mathrm{d}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{u}\mathrm{e}}{\mathrm{E}\mathrm{x}\mathrm{p}\mathrm{e}\mathrm{r}\mathrm{i}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}\mathrm{a}\mathrm{l}\mathrm{v}\mathrm{a}\mathrm{l}\mathrm{u}\mathrm{e}}})\times 100$$

(19)

The prediction error of approximately 8% between the experimental results and model predictions is within the acceptable range for multi-objective response surface optimization. Understanding the potential sources of this deviation is essential for assessing the reliability and robustness of the developed model.

A key factor is the variability in material properties, especially in aggregate gradation and filler characteristics, which can greatly influence mix compaction and mechanical behavior. Recycled materials such as brick waste powder (BWP) and polyester fibers (PFs) significantly complicate the asphalt matrix due to their inherent properties. The angular morphology and porosity of BWP may result in uneven filler distribution, which can lead to localized differences in binder absorption and internal architecture. Previous research indicates that recycled fillers including brick powder can exert variable impacts on hot-mix asphalt, diminishing moisture resistance and rutting performance, both of which are essential for the durability of pavement over time [72,73,74].

The length and distribution of PFs can similarly affect performance. Longer fibers can become entangled during mixing, which diminishes uniformity and adversely impacts reinforcement efficiency. Inconsistencies may affect essential mechanical properties, including tensile strength and Marshall stability [35].

Alongside material variability, operational and experimental uncertainties, including variations in mixing temperature, compaction effort, and uniformity of asphalt binder coating, can introduce noise into the results. Even with rigorous compliance to standardized protocols, complete control of these variables in laboratory settings remains challenging [72,75]. The selection of the compaction method significantly impacts void characteristics and aggregate distribution, subsequently influencing both mechanical and functional performance [75,76].

Recent studies have demonstrated the effectiveness of using RSM and optimization techniques to enhance the mechanical and volumetric properties of asphalt mixtures under various modifiers and fillers [100,101,102,103,104,105]. From a modeling perspective, response surface methodology (RSM) presumes polynomial relationships and smooth response surfaces, which may not sufficiently represent the complex nonlinear interactions found in real materials. This limitation is especially pertinent in the context of random particulate composites such as modified asphalt, where mechanical behavior is influenced by complex interactions among components. RSM treats all input factors as independent, neglecting potential hidden correlations that may affect outcomes [106,107,108]. The simplifications, although beneficial for decreasing experimental burden, inherently restrict the predictive accuracy of the model.

Alternative modeling approaches, including neural networks and machine learning meta-models, have been proposed to address these limitations. These techniques facilitate improved approximation of nonlinear behaviors, integrate physical knowledge, and more effectively capture complex variable interactions compared to traditional RSM [81,82,83]. The use of desirability functions and sensitivity analysis has further strengthened the interpretation and practical utility of RSM in pavement engineering and material research [109,110,111]. These methods present potential directions for future investigations into asphalt mixtures that include recycled and fibrous additives.

4. Conclusions

This study assessed the effects of brick waste powder (BWP) and polyester fibers (PFs) of various lengths on the performance of asphalt concrete (AC) mixtures. Based on experimental results and statistical modeling using response surface methodology (RSM) and ANOVA, the following conclusions are drawn:

• Substituting limestone filler (LS) with BWP enhanced moisture resistance, indirect tensile strength (ITS), and Marshall stability. BWP showed superior pozzolanic activity and stronger binder-aggregate adhesion.
• PF addition improved tensile strength, Marshall stability, and resistance to moisture damage, with performance trends depending on fiber length.
• Among the tested fiber lengths (3 mm, 8 mm, and 15 mm), 8 mm provided the best mechanical performance. However, RSM optimization indicated 5 mm as the statistically optimal length when multiple responses were simultaneously considered.
• The predictive RSM model showed an error below 8%, confirming its reliability for performance prediction and mix design optimization.

From a practical engineering standpoint, the use of BWP offers clear advantages in terms of sustainability and cost reduction as it repurposes readily available construction waste and reduces reliance on natural limestone filler. Similarly, incorporating PFs from recycled PET bottles aligns with circular economy principles and adds negligible cost due to the low dosage (0.3% by aggregate weight). The mix design does not require major modifications to conventional asphalt production or laying procedures, making the proposed approach feasible for field implementation.

4.1. Study Limitations and Future Research Directions

The current study presents strong evidence supporting the use of brick waste powder (BWP) and polyester fibers (PFs) in asphalt concrete (AC) mixtures; however, it is essential to recognize several limitations to properly contextualize the findings and inform future research directions.

Initially, all experimental procedures were performed in a controlled laboratory environment utilizing standardized short-term performance assessments. The study omits considering real-world variables, including traffic-induced fatigue, long-term environmental aging, and seasonal temperature fluctuations, which can significantly affect pavement behavior. The long-term durability and performance of the proposed mixtures under repeated vehicular loading and environmental stresses have yet to be verified. The physical stability of PFs, which may deteriorate due to oxidative aging or extended ultraviolet (UV) exposure, and the chemical resilience of BWP, potentially affected by freeze–thaw cycling, wet–dry conditions, or chemical leaching, warrant further investigation.

Field trials and accelerated pavement testing should be conducted to assess real-world parameters, including fatigue cracking, rutting resistance, stiffness degradation, and long-term moisture susceptibility. Simulated aging protocols, such as UV radiation, oxidation, and freeze–thaw cycles, will aid in determining the material resilience and life-cycle reliability of BWP–PF modified mixtures. Furthermore, observing crack propagation and healing behavior over time would yield more thorough insights into the structural evolution of these mixtures under sustained service conditions.

A significant limitation is the application of fixed proportions—5% BWP by aggregate weight and 0.3% PF. The selected values, derived from existing literature and standard filler dosage ranges, may not reflect optimal levels for all performance criteria. The relationship among filler content, fiber length, and dosage is expected to be nonlinear, necessitating a multi-dimensional experimental design for performance optimization. Future research should employ factorial or response surface methodologies to systematically investigate the impacts of varying BWP and PF contents. This would facilitate the identification of threshold limits, synergistic interactions, and optimal combinations designed to meet specific pavement performance objectives.

This study highlights the environmental and economic benefits of utilizing recycled materials; however, it does not adequately consider the potential environmental risks linked to large-scale implementation. The mechanical crushing of brick waste can produce respirable particulate matter, including PM₁₀ (particles with aerodynamic diameters ≤ 10 μm) and PM_2.5 (particles ≤ 2.5 μm), which may raise air quality issues. The leaching behavior of BWP under environmental exposure, especially in regions with high rainfall or groundwater sensitivity, remains unassessed. Future research must incorporate thorough environmental risk assessments, including particulate emission profiling, toxicity characterization, and life-cycle assessments (LCAs) to assess the ecological footprint of these sustainable AC mixtures from production to end-of-life.

Ultimately, regional adaptability continues to be an unresolved issue. The present findings are derived from a temperate climate and standard loading conditions; however, performance in extreme environmental contexts—such as arid regions with extended heat exposure, cold areas with frequent freeze–thaw cycles, or tropical climates characterized by heavy rainfall and humidity—requires validation through geographically focused studies. When evaluating generalizability, it is essential to consider regional material sourcing, construction practices, and traffic characteristics.

In conclusion, the findings of this study provide a robust basis for the application of BWP and PFs in asphalt concrete. However, it is imperative to conduct comprehensive field validation, performance optimization, environmental risk analysis, and climate-specific evaluation as essential subsequent steps to facilitate real-world implementation and policy adoption.

4.2. Economic and Practical Considerations

While this study primarily focused on mechanical and durability performance, the use of BWP and PFs also offers potential economic benefits. BWP is sourced from construction demolition waste, which is often freely available or low-cost compared to processed limestone filler. Its use can significantly reduce material costs and landfill burden. Similarly, the polyester fibers used in this study are manufactured from recycled PET bottles, contributing to reduced raw material costs and supporting circular economy goals. Although no formal cost–benefit analysis was conducted in this work, preliminary estimates suggest that substituting 100% of LS filler with BWP and incorporating 0.3% PFs could reduce material input costs by approximately 10–15%, depending on regional pricing and availability. Future research should include a detailed life-cycle cost assessment (LCCA) to evaluate cost-effectiveness over the pavement lifespan, including savings related to maintenance and durability improvements.

4.3. Industrial Feasibility

The asphalt concrete (AC) mixtures developed in this study are compatible with conventional asphalt production processes. The use of brick waste powder (BWP) as a filler does not require specialized processing beyond drying and milling, which are scalable using standard industrial equipment. Polyester fibers (PFs), applied at a low dosage of 0.3% per aggregate weight, can be uniformly mixed using existing fiber-feed systems commonly employed for cellulose or polymer additives. Furthermore, the mix design follows standard Marshall criteria and uses materials and equipment available in most asphalt plants. Therefore, the proposed formulations can be readily scaled up for full-scale production without significant adjustments to existing infrastructure.