Predictive Performance Evaluation of an Eco-Friendly Pavement Using Baosteel’s Slag Short Flow (BSSF) Steel Slag

Abstract

Predicting pavement performance is essential for highway planning and construction, considering traffic, climate, material quality, and maintenance. This study’s main objective is to evaluate Baosteel’s Slag Short Flow (BSSF) steel slag as a sustainable aggregate in pavement engineering by means of durability. The research integrates pavement performance prediction using BSSF and assesses its impact on fatigue resistance and percentage of cracked area (%CA). Using the Brazilian mechanistic-empirical design method (MeDiNa), eight scenarios were analyzed with soil–slag mixtures (0%, 25%, 50%, and 75% slag) in base and subbase layers under two traffic levels over 10 years. An asphalt mixture with 15% steel slag aggregate (SSA) was used in the surface layer and compared to a reference mixture. Higher SSA percentages were applied to the base layer, while lower percentages were used in subbase layers, facilitating field implementation. The resilient modulus (MR) and permanent deformation (PD) were design inputs. The results show that 15% SSA does not affect rutting damage, with %CA values below Brazilian limits for traffic of 1 × 10⁶. The simulations confirm BSSF as an effective and sustainable alternative for highway pavement construction, demonstrating its potential to improve durability and environmental impact while maintaining performance standards.

1. Introduction and Background

In the context of sustainability, prioritizing environmentally viable, durable, and technically sound solutions has become a fundamental goal in pavement design. In this scenario, adopting innovative materials, such as Baosteel’s Slag Short Flow (BSSF) steel slag, as substitutes for non-renewable natural aggregates emerges as a promising strategy for developing more resilient and environmentally responsible pavements.

The philosophy of the 4Rs of sustainability—rethink, reduce, reuse, and recycle—seeks to promote practices that minimize waste generation or ensure its proper disposal, thereby contributing to environmental preservation and sustainable development. According to the Center for Management and Strategic Studies [1], significant environmental change will only occur when waste treatment models adopt the same level of control that the steel industry applies to iron and steel production. This shift would result in the generation of more consistent byproducts with higher added value.

Before incorporating new materials into road construction, it is essential to conduct analyses using performance prediction methods to evaluate future structural parameters. Computational simulations with alternative materials have demonstrated positive results in this regard, with some authors [2,3] reporting superior performance in alternative pavements compared to conventional structures made with natural materials.

Given the lack of studies on the application of BSSF steel slag in pavement structures, this research seeks to fill that gap by integrating pavement performance prediction with the use of BSSF steel slag aggregate. Novelty materials require a larger amount of time to be tested, both in the laboratory and in the field, to propose recommendations on their use and to incorporate them into standards from roadway agencies; therefore, new research on these materials is important to develop and to increase databases. This innovative approach in this research not only aligns with sustainable practices but also highlights the potential to optimize pavement durability. The primary objective of this study is to evaluate the influence of aggregate content on pavement performance, specifically analyzing its actual effect on rut depth and percentage of cracked area (%CA) over a 10-year design period, compared to a conventional pavement structure.

1.1. Pavements Stabilized with Steel Aggregate

The need for financial resources to build new highways and rehabilitate old ones, growing concerns about the exploitation of natural resources, and the commitment to environmental preservation motivate researchers worldwide to explore technologies or materials that incorporate, partially or entirely, these principles. In this context, the use of byproducts from the steel industry in pavement structures has been proposed, such as steel slag [4,5,6,7], bringing significant economic benefits, especially in areas where steel production is relevant [8]. This is due to the fact that, according to the literature, the physical characteristics of slag are comparable to those of natural aggregates.

Polese et al. [9] state that the most significant application of steelmaking slags is in road bases; however, the biggest limitation is expansiveness, and they mention that treatments can be employed to reduce this effect. Pellegrino and Gaddo [10] note that the presence of free calcium and magnesium oxides is responsible for hydration and expansion phenomena typical of electric arc furnace (EAF) slags (black slags, resulting from the cold charging of scrap). This happens because contact with water promotes the rapid hydration of calcium oxide, causing quick volumetric expansion. Magnesium oxide, on the other hand, has slow hydration, resulting in more spaced expansion. However, the main disadvantage of using Linz-Donawitz (LD) slag, as with EAF slag, is expansion (volumetric instability) caused by the hydration of reactive MgO and free CaO.

LD steel slag contains a considerable amount of steel or metallic iron particles that become attached to the slag during the oxygen blowing process and may contribute to its expansion due to corrosion and oxidation of the residual metallic iron. The change in the allotropic forms of dicalcium silicate or calcium orthosilicate (C2S) may also be a potential source of volumetric instability for LD slags [11]. The heterogeneity of steel slags, which varies depending on the source and production process, is an important factor. The BSSF slag used in this study has the advantage of not requiring prior “curing”, unlike other steel slags.

Regarding the use of slag in asphalt pavement layers, Bosurgi et al. [12] investigated the effect of replacing natural basalt aggregates with steel slag in asphalt mixtures and found that the slag provided granular interlock similar to that of basalt aggregates, resulting in good compaction and higher crushing resistance. Despite the initial low affinity of slag with the asphalt binder due to its high silica content, the introduction of an anti-stripping agent substantially improved this issue, increasing the durability of the mixtures. Moreover, the asphalt mixtures enriched with slag showed a reduction in the optimal binder content, resulting in cost savings and environmental sustainability benefits. Additionally, these mixtures exhibited greater resistance to permanent deformation compared to control mixtures. Thus, the use of slag can represent an effective solution in practical applications, offering technical, economic, and environmental advantages.

Stabilization of clayey soils for road pavements with EAF slag in proportions of 15% and 30% was investigated by [6]. The soil, classified as A-4 (silty sand) and ML (low-plasticity silt), showed an increase in the California bearing ratio (CBR) and a reduction in expansion, which tended to zero with higher slag content. The studies also identified the influence of slag on the optimal moisture content (OMC) with the increment of slag. Rohde et al. [4] also highlighted the potential of steel slag as a substitute for aggregates or soil stabilizers to increase mechanical strength. The tests performed, such as grain size, sodium sulfate durability, Los Angeles abrasion, Proctor compaction, and CBR, revealed that the natural grain size of the slag needed adjustments through crushing or mixing. The corrected slag exhibited a CBR of about 200%, while the raw slag ranged between 80 and 100%. The mass loss in the Los Angeles abrasion test was 38%. These authors found an ideal curing period of 4 months in the air to prevent significant expansions.

Steel slag also has the potential of being a self-healing material to be used in asphalt mixtures [13]. Sun et al. [14] tested steel slag by means of microwave heating and concluded that this material provided a heating rate two times faster than a natural aggregate. Other researchers have also shown the capability of steel slag in self-healing for fatigue damage recovery [15,16,17] and deicing of hot-mix asphalt and cold-mix asphalts [18].

Although previous studies have shown good performance of slags from other sources in pavements, this study focused exclusively on BSSF slag, allowing for better control of variables while limiting the generalization of the results. Research evaluating the use of steel slag generated by the BSSF process for soil improvement in pavement layers is almost nonexistent. Therefore, this study aims to expand the knowledge on a potential additional application of this type of material in pavements. In recent years, a few studies have been conducted in order to incorporate this material into engineering materials and have resulted in some important remarks: The substitution of Portland cement with BSSF in their composition increases unconfined compressive strength (UCS) and elasticity [19], the use of BSSF as aggregates of asphalt mixtures enhances healing characteristics [20], the incorporation of BSSF fillers into asphalt mixtures maintains resilient modulus values in comparison to traditional fillers [21], the stabilization with BSSF increases UCS values of soils for pavements [22], and others.

1.2. Brazilian Empirical-Mechanistic Method (MeDiNa)

In road infrastructure, traffic loads on pavement structures induce deformations that are divided into two components: those that are permanent or irrecoverable (plastic) and those that are resilient or recoverable (elastic). Current pavement engineering practices primarily focus on material resilience but tend to overlook their plastic capacity. In recent years, researchers have been improving the understanding of pavement materials, especially regarding permanent deformation, through repeated load tests [23].

As highlighted by Bastos et al. [24], the phenomenon known as rutting is among the primary imperfections observed in asphalt pavements. This distress not only accelerates structural deterioration but also results in discomfort during driving, compromises user safety, and increases operational costs. Fatigue cracking is another inherent distress in asphalt pavement structures. Given its inevitable occurrence, the analysis of crack growth behavior assumes significant importance, as this study is essential for determining the appropriate time for pavement rehabilitation and evaluating the service capacity of highways [25].

In Brazil, the MeDiNa (National Pavement Design Method), an empirical-mechanistic method, was developed within partnerships between universities and specialized road research institutes in order to design asphalt pavements by limiting their main distresses at pre-established threshold values, e.g., 30% cracked area for fatigue and 10 mm for rut depth. The software provides the evolution of these distresses along the lifecycle of the pavement structure and can adjust the thicknesses of each layer in order to comply with the performance requirements. Among the main inputs considered in MeDiNa to design a pavement structure are the values of resilient modulus (RM) of asphalt, granular and soil materials, and PD calibration parameters for granular and soil materials. Fatigue curve calibration coefficients of the asphalt mixtures are also required.

2. Materials and Methods

2.1. Materials

The soil used in this research comes from the state of Ceará, Brazil. The steel slag originates from the BSSF process, produced in an oxygen (LD) furnace by a local steel industry in Ceará. These materials were applied in the surface, base, and subbase layers of pavement structures composed of natural soil (NS) and steel slag aggregate (SSA). The proposed mixture proportions were as follows: M1, with 25% NS and 75% SSA (NS25+SSA75); M2, with 50% NS and 50% SSA (NS50+SSA50); and M3, with 75% NS and 25% SSA (NS75+SSA25). For the surface layer, asphalt concrete containing 15% BSSF steel slag was used. The natural soil had a specific gravity of 2.574, and the steel slag aggregate had a specific gravity of 3.651. The soil was classified as SC by the Unified Soil Classification System (USCS), A-4 (AASHTO), and LA’ (Tropical Miniature Compacted), according to DNER-ME 258 [26] and DNER-ME 256 [27]. The BSSF steel slag was classified as SP (USCS) and A-1-a (AASHTO). The sand equivalent values were 6% for the soil and 98% for the SSA, following DNIT 450 [28]. The slag presented a Los Angeles abrasion resistance of 19%, based on DNIT 451 [29](DNIT, 2024), and a permeability coefficient of 6.02 × 10⁻³ cm/s, according to NBR 14.545 [30]. Regarding the reference mix and the asphalt mixture with 15% BSSF,

Table 1. Asphalt mix design properties.

Parameter	Reference Mix	15% SSA Mix
Asphalt binder content (%)	4.4	4.4
Maximum specific gravity, Gmm	2.463	2.564
Bulk specific gravity, Gmb	2.362	2.468
Air voids content, AV (%)	3.9	3.5
Voids in mineral aggregate, VMA (%)	14.1	14.2

presents the main asphalt mix design and volumetric parameters.

2.2. Laboratory Testing

To determine parameters that assess stiffness, fatigue cracking, and permanent deformation of mixtures and materials, several laboratory tests were conducted, such as particle size distribution, bulk density, Atterberg limits, and sand equivalent. Additionally, mechanical characterization tests were conducted: repeated load triaxial test to determine RM [31] and PD [32]. The latter involved the application of 150,000 load cycles to facilitate damage prediction necessary for the Brazilian mechanistic-empirical design of MeDiNa. Modified energy was applied to all the specimens tested. The six stress pairs (σ₃ × σ_d) applied for PD determination for each mixture analyzed were (40 × 40), (40 × 120), (80 × 80), (80 × 240), (120 × 240), and (120 × 360), at a loading frequency of 2 Hz. As for the PD values, the coefficients found by applying the Guimarães model [33] were inserted (Equation (1)).

$${\mathsf{\epsilon }}_{\mathrm{p}}^{\mathrm{e}\mathrm{s}\mathrm{p}}\left(\mathrm{\%}\right)={\mathsf{\psi }}_1\times {({\mathsf{\sigma }}_3/{\mathsf{\rho }}_0)}^{{\mathsf{\psi }}_2}\times {({\mathsf{\sigma }}_{\mathrm{d}}/{\mathsf{\rho }}_0)}^{{\mathsf{\psi }}_3}\times {\mathrm{N}}^{{\mathsf{\psi }}_4}$$

(1)

where

ε_p^esp (%) = specific permanent deformation;

σ₃ = confining stress;

σ_d = deviator stress;

ψ₁, ψ₂, ψ₃, and ψ₄ = calibration coefficients;

ρ₀ = reference stress (atmospheric stress);

N = number of load cycles.

For a comprehensive characterization of the surface layer, the following tests were conducted: (i) diametral compression fatigue test [34], (ii) resilient modulus test [35], and uniaxial repeated load rest [36], used to obtain the flow number (FN) parameter. The RM test was conducted for the asphalt mixtures with 15% of steel aggregate. However, since the result of this isolated test does not allow for the prediction of the performance of a pavement structure composed of this material, it is necessary to conduct other mechanical tests, such as the fatigue cracking resistance test. The diametral compression fatigue test was conducted following the guidelines established in DNIT 183/2018-ME [34]. For this study, the stress levels considered were 30, 40, and 50% of the indirect tensile strength (ITS) of the materials analyzed. The fatigue curve, known as the Wöhler curve, is represented by log–log plots, which allow for the modeling of the material behavior by obtaining a linear regression equation (Equation (2)).

$${\mathrm{N}}_{\mathrm{f}}={\mathrm{k}}_1\times {{\mathsf{\epsilon }}_{\mathrm{t}}}^{{\mathrm{k}}_2}$$

(2)

where

N_f = number of load cycles until failure;

ε_t = tensile strain observed at the beginning of the test;

k₁ and k₂ = calibration coefficients.

The uniaxial repeated load test, as outlined by the Brazilian standard DNIT 184/2018-ME [36], is conducted to determine the parameter known as “flow number”, which provides an assessment of the asphalt mixtures performance regarding permanent deformation. For the design software, the contribution of the asphalt surface layer is not considered in the total pavement rut depth; however, the method requirements indicate a range of FN values for each traffic volume as part of the acceptance or rejection criteria for the material under test. For FN values below 100, the indication is to apply only the maximum number of accumulated equivalent single axle loads (ESALs) of 10⁶, while FN values between 100 and 300 are considered for the number of accumulated ESALs from 10⁶ to 10⁷. To obtain the FN value, the Francken model was used, in which the inflection point of the model’s second derivative represents the FN. The model is defined by Equation (3).

$${\mathsf{\epsilon }}_{\mathrm{p}}=\mathrm{A}\times {\mathrm{N}}^{\mathrm{B}}+\mathrm{C}\times \left({\mathrm{e}}^{\mathrm{D}\times \mathrm{N}}-1\right)$$

(3)

where

ε_p = specific permanent deformation;

N = number of load cycles;

A, B, C, and D = constants of the model calibrated to experimental points.

2.3. Computer Simulations and Design

For the present paper, two levels of traffic were used to analyze the evolution of the pavement structure damage. The first level corresponds to the number of accumulated ESALs of 1 × 10⁶ (low-level traffic), and the second considers a traffic of 1 × 10⁷ (medium-level traffic). A Poisson’s ratio of 0.35 was used for the base and subbase layers, and a value of 0.30 was used for the asphalt surface layer. Predictions applied to a primary arterial system, with a reliability of 85%, indicate that the CA% is required to be below 30%, and the rut depth is required to be below 13 mm (Figure 1). The program allows for an analysis period between 1 and 20 years. A 10-year period was selected, which corresponds to half of the available range, for running the simulations.

For the configuration of the pavement structures evaluated, similar to typical structures in Brazil, soil–aggregate mixtures with higher steel slag content were chosen for the base layer in comparison to the mixtures used for the subbase layer. For the surface layer, asphalt concrete with 15% steel slag aggregate was used for structures #1, #3, #6, #7, and #8 (Figure 2). However, aiming to identify the influence of BSSF steel slag content on the performance of the layers executed with these mixtures, structure #5 was designed with materials traditionally employed in the field.

A traditional asphalt concrete was chosen for the reference surface layer. For the reference base layer, a soil–gravel material was employed with the following characteristics: non-lateritic clayey soil (NG’) + 30% of coarse gravel + 40% of fine gravel. Data from the software were used for the subgrade layers. For the steel slag aggregate–soil mixtures and for the natural soil, laboratory obtained RM and PD values were used. Figure 2 shows the layouts of the simulated pavement structures.

3. Results and Discussions

3.1. Characterization of Materials for Granular Layers and Asphalt Pavement

The composition of the materials used in the layers as well as the RM and PD values and models are provided in

Table 2. Material properties.

Material	Soil (%)	Steel Aggregate (%)	Composition	RM Average (MPa)	PD Model (%)
					ψ1	ψ2	ψ3	ψ4
Asphalt layer	-	-	Reference mix	7553	-	-	-	-
Asphalt layer	-	15	15% SSA	8330	-	-	-	-
Natural soil	100	0	NS	543	0.01	−0.87	1.04	0.13
M1	75	25	NS75+SSA25	687	0.004	0.472	0.261	0.191
M2	50	50	NS50+SSA50	931	0.01	−0.30	0.77	0.11
M3	25	75	NS25+SSA75	920	0.01	−1.23	1.14	0.15
Soil–gravel	-	-	-	433	0.24	−0.34	1.37	0.04
Subgrade	-	-	-	100	0.244	0.419	1.309	0.069

.

For the PD model, the analysis of the coefficients shows the following expectations: for ψ₁, this value is typically expected to be positive; for ψ₂ (corresponding to confining stress), this coefficient is expected to be negative because the natural behavior is that the higher the confining stress, the lower the permanent deformation; for ψ₃ (related to deviatoric stress), these values are expected to be positive, as the expected behavior is that the greater the stress applied to the material, the greater the deformation observed; for ψ₄ (related to the number of cycles), a positive value is also expected, as the trend is that an increase in cycles negatively influences permanent deformation, meaning it will increase. It is possible to observe that the coefficient with the greatest influence on the material is ψ₃, which also reflects expected behavior, as higher loads applied to the materials lead to greater deformations. All the analyzed samples met these conditions, except for mixture M1 (SN75+AS25), which showed an unexpected ψ₂ value with a positive result.

The fatigue coefficients were obtained through the diametral compression fatigue test of the asphalt mixture, with 15% SSA used in the eight simulations. Figure 3a illustrates the corresponding fatigue curve and the fitting coefficients, which are used as input in the pavement design software. As expected, with the increase in stiffness observed in the MR test and the incorporation of SSA, there is an increase in the fatigue resistance of the asphalt mixtures. A higher RM provides greater stiffness to the asphalt layer, resulting in smaller elastic deformations under traffic loads. Consequently, there is a reduction in the propagation of fatigue cracks. However, excessive stiffness can be detrimental to surface layers. Nevertheless, the 10.3% increase in RM of the asphalt mixture with steel slag aggregate (SSA) appears to be favorable and within the balance range between stiffness and flexibility. The increase in RM of the SSA mixture compared to the reference mixture’s RM may have also contributed to reduced accumulated deformations, as fatigue is associated with microcracks or microdeformations that, over time and under repeated loading cycles, lead to larger cracked areas.

The uniaxial repeated load test for permanent deformation was performed on two specimens of each asphalt mixture and provided FN values as indicated in Figure 3b. It is important to note that FN considerably decreased when the SSA slag was incorporated into the mixture. The stiffness identified by the RM test would indicate a better resistance to this mixture, contrary to what the results show. Such phenomenon might be related to the high percentage of the byproduct in the mixture. That percentage can indicate a higher modification on aggregate resistance. Also, the percentage of binder used in the asphalt mixtures might influence the results. Industrial aggregates tend to absorb and interact with the binder in a different proportion. For these mixtures, the asphalt binder content was the same (4.4%).

3.2. Evaluation of Structures Regarding Cracked Area and Rut Depth

For the scenarios analyzed, the software suggested a specific design, determining the thickness for each pavement layer considering the materials’ characteristics, as shown in Figure 4. To standardize the layers for each scenario, the same structure for the eight simulations on both traffic volumes was considered. For fatigue and stiffness test results, the software suggested an increase in the surface layer thickness with the increase in the traffic volume.

With the definition of the input data, the eight structures defined in this research were simulated.

Table 3. Rut depth for granular layers.

Structure	Number of Accumulated ESALs	Rut Depth (mm)
		Base	Subbase	Subgrade	Total
#1	1 × 106	1.07	0.49	1.33	2.89
	1 × 107	1.35	0.36	0.61	2.32
#2	1 × 106	1.05	0.49	1.34	2.88
	1 × 107	1.37	0.37	0.64	2.38
#3	1 × 106	1.09	0.16	1.23	2.48
	1 × 107	1.40	0.12	0.6	2.12
#4	1 × 106	1.07	0.16	1.24	2.47
	1 × 107	1.43	0.13	0.63	2.19
#5	1 × 106	2.63	0.55	1.64	4.82
	1 × 107	0.84	0.39	0.67	1.90
#6	1 × 106	0.19	0.51	1.41	2.11
	1 × 107	0.15	0.37	0.63	1.15
#7	1 × 106	1.06	0.06	1.32	2.44
	1 × 107	1.36	0.07	0.6	2.03
#8	1 × 106	0.19	0.06	1.40	1.65
	1 × 107	0.15	0.07	0.62	0.84

presents the results of rutting for the simulated traffic levels after 120 months.

When analyzing Table 3, it is observed that, in all scenarios for a traffic volume of 1 × 10⁶, higher total rutting values were found. The Brazilian mechanistic-empirical design method combines empirical aspects, based on field data, with mechanical aspects, grounded in mathematical models, to predict pavement performance. The counterintuitive reduction in rut depth observed under medium traffic compared to low traffic can be attributed to limitations in the models used by the program. These models sometimes fail to account for all factors influencing pavement behavior, such as the complex interactions between different materials, including SSA, as well as other specific conditions. Also, the asphalt layer considered for low-level traffic is much thinner than the structure used for medium-level traffic, and this could contribute to the rut depth results of the granular layers, even though this distress is not calculated for the surface layer.

When evaluating the impact of the steel slag on rutting damage, it was observed that this did not affect the pavement performance. Comparing structures #1 with #2 and structures #3 with #4, where the surface layer is the only differing layer, it can be observed that the rutting values are virtually the same.

Upon comparing structures #6 and #8, where the subbase layer differs, it is noted that the layer with only natural soil had a greater influence on rutting damage compared to the subbase layer utilizing mixture M1. Structure #5, composed of traditional materials, exhibited a rut depth of 4.82 mm for a traffic volume of 1 × 10⁶, higher than the other simulations at the same traffic level. For a traffic volume of 1 × 10⁷, structure #2 displayed the highest rutting result among the simulations at 2.38 mm. The fact that structure #5 exhibited the greatest wheel path depth can be attributed to the quality and mechanical properties of the materials used in the pavement layers. It is worth noting that the base layer in scenario #5 is stabilized with graded crushed stone, while scenario #2, for instance, features a base stabilized with steel slag aggregate, which provided greater resistance to this type of damage.

When comparing the simulations for structure #1 with structure #6 and isolating the base layer with mixtures M3 and M2, respectively, it is inferred that reducing the aggregate content contributed to less rutting damage. This observation is repeated in the analysis of structures #3 and #8, where the latter uses M2 as the base layer and mixture M1 as the subbase layer. Meanwhile, structure #3 employs M3 and M2 for the base and the subbase, respectively, presenting lower rutting in both traffic scenarios. Overall, the structures with subbase layers stabilized with aggregate exhibited lower rutting values when compared to the subbase layers with natural soil only.

In the analysis of average traffic level, the base layers showed higher values of settlement deformation compared to the subbase layers. This can be justified by the fact that the mixtures exhibit greater displacements in the permanent deformation test, showing values inversely proportional to the aggregate content. As with the M3 mixture, the permanent deformation and settlement values were greater than those provided by the design software. In summary, it can be inferred that increasing the aggregate content in the mixtures has a specific influence on the total value of settlement deformation.

The surface layer in the proposed structures, regardless of the material used, presented zero values for the depth of deformation due to settlement. The program only indicates the traffic volume range related to the FN value. According to the FN results, the asphalt mixture with 15% SSA should not be recommended for use in medium traffic volumes. It is not possible to observe the impact of using this destruction ratio in mixtures for different traffic volumes, since the software does not consider the surface layer contribution to the total rutting.

As for the deformation limit due to settlement, 13 mm, none of the structures reached this value. It is important to note, in Figure 5a, that structure #5 (taken as reference) presented higher rutting values compared to the other structures with steel slag aggregate in the low-level traffic (1 × 10⁶). It can be inferred that the inclusion of steel slag aggregate would help delay the onset of permanent deformations compared to the conventional structure. For a higher traffic level, an increase in the proportion of SSA in the soil mixture indicates lower rutting (Figure 5c). Furthermore, within a 10-year design period, conventional granular layers are more likely to develop premature deformation pathologies compared to granular layer structures with partial replacement of natural aggregates.

For the low-level traffic (1 × 10⁶), all the simulated structures presented %CA values below the limit (30%) for 10 years (Figure 5b). However, for the medium-level traffic (1 × 10⁷), structures #2, #4, and #5 exceeded the allowed %CA limit, reaching 39.10, 37.43, and 47.81%, respectively, all with a conventional surface layer. This limit was exceeded around 80 months, or 6.5 years, of the pavement’s life. This discrepancy highlights the advantage of SSA as a long-term mitigator of cracked area in pavement. In scenario #5, the surface layer and the other layers do not use steel slag, and this scenario exhibited the highest value of %CA for medium traffic. To delay this pathology, it is recommended to use high-quality materials with greater resistance to permanent deformation and representative resilient modulus. This is especially crucial for the surface layer, as it plays a fundamental role in cushioning and distributing the stresses caused by traffic as well as the effects of climatic and temperature variations. Figure 5d shows the results for %CA for medium-level traffic volume, where this pattern is also observed.

4. Conclusions

This study examined the influence of pavement layer composition on rutting and cracking behavior under different loading conditions. The analysis considered surface, base, and subbase layers, focusing on the incorporation of Baosteel’s Slag Short Flow (BSSF) steel slag aggregate (SSA). The main conclusions can be divided into the following aspects:

• Surface layer behavior: The application of 15% BSSF SSA in the surface layer did not result in measurable differences in rut depth compared to conventional mixtures under the evaluated conditions.
• Base and subbase layer behavior: Pavement structures with SSA in the base and subbase layers presented lower rut depths compared to those without SSA. Mixtures with higher aggregate content were associated with reduced permanent deformation.
• Cracking analysis: Configurations without SSA showed higher cracked area percentages (%CA) under medium traffic loading. Structures with conventional surface layers exceeded acceptable limits for %CA, indicating the need for material or structural adjustments.
• Predictive modeling and result variability: Some deviations in expected behavior were observed. These may be related to limitations in the predictive models applied or the sensitivity of the evaluation methods used.
• Material characterization and implementation: The characterization of granular and asphalt layers with and without SSA, together with the evaluation of observed pavement distresses, enabled the identification of performance patterns across different configurations. The use of BSSF SSA requires no pre-treatment, which simplifies field application compared to other steel slags.
• Implications for design and future studies: The relationship between material composition, traffic demand, and performance criteria highlights the importance of refining current design methodologies. Further studies are recommended to improve the accuracy of predictive models and to expand the analysis of alternative materials in pavement structures.

In summary, the results support the use of BSSF steel slag in base and subbase layers as part of flexible pavement systems. The behaviors observed under varied loading scenarios provide a basis for its consideration in future design practices and material specifications.