Sustainable Composites from Sugarcane Bagasse Fibers and Bio-Based Epoxy with Insights into Wear Performance, Thermal Stability, and Machine Learning Predictive Modeling

Abstract

The global push for sustainable materials has intensified the research on natural fiber-reinforced composites. This study investigates the potential of sugarcane bagasse fibers, combined with a bio-based epoxy matrix, as a sustainable alternative for high-performance composites. A comprehensive approach was adopted, including wear testing, thermal and structural characterization, and machine learning predictive modeling. Ethylene dichloride-treated fibers exhibited the lowest wear rate (0.245 mg/m) and the highest thermal stability (T20% = 260 °C, char yield = 1.3 mg), highlighting the role of optimized surface modifications. XRD (X-ray diffraction) analysis revealed that pre-treated fibers achieved the highest crystallinity index of 62%, underscoring the importance of structural alignment in fiber-matrix bonding. Machine learning insights using a Random Forest model identified fiber treatment as the most significant parameter influencing wear performance, with accurate predictions validated through experimental results. This work demonstrates the transformative potential of sugarcane bagasse fibers in sustainable polymer composites, offering a pathway for environmentally friendly, lightweight, and durable material solutions. These findings integrate experimental rigor with computational insights, paving the way for advancements in natural fiber-based composite technologies.

1. Introduction

The growing demand for environmentally sensitive materials is driving significant research efforts toward the development of renewable polymers and natural fiber-reinforced composites. Although synthetic composites traditionally offer better mechanical and thermal properties, environmental concerns arise from fossil fuel use and the lack of biodegradability [1,2]. This has brought about a paradigm shift to bio-based matrices and renewable fiber reinforcements that meet industrial requirements yet reduce ecological impacts [3,4]. The integration of bio-based epoxy with sugarcane bagasse fibers represents a promising pathway for the development of high-performance, sustainable composite material [5,6].

Natural fibers have gained significant attention in sustainable composite applications due to their biodegradability, lightweight nature, and superior specific strength. Various fibers such as flax, hemp, jute, kenaf, and sisal have been widely studied for their reinforcement potential in polymer matrices. Their mechanical performance depends on factors such as fiber structure, chemical composition, and surface roughness [7,8]. Among these, flax and hemp fibers exhibit superior tensile properties due to their high cellulose content, whereas jute and coir fibers are known for their affordability and widespread availability [9,10,11].

1.1. Sugarcane Bagasse Fiber

Worldwide, the sugar industry produces vast amounts of bagasse annually, a significant portion of which is either burned for energy recovery or left underutilized, contributing to environmental concerns. Recently, sugarcane bagasse fibers have gained attention as a valuable resource for polymer composites owing to their exceptional strength-to-weight ratio, biodegradability, and renewability. Chemically, bagasse consists of approximately 40% cellulose, 24% hemicellulose, 30% pentosans, 0.6–0.8% pectin, and 25% lignin [12,13]. Among these components, cellulose plays a central role as a structural polysaccharide composed of glucose subunits linked in long chains with a high degree of polymerization (300–1700). Strong hydrogen bonding within and between cellulose molecules results in excellent tensile strength and water resistance, which are key properties of composite reinforcement [14,15]. The reinforcing ability of cellulose is further influenced by its crystallinity, polymerization level, and fiber dimensions.

Hemicellulose, another important component, is a hydrophilic and amorphous polysaccharide that binds to cellulose and lignin in the plant cell walls. Primarily composed of xylose and arabinose with small amounts of mannose and galactose, hemicellulose is soluble in both acidic and alkaline solutions [16]. Lignin, a phenolic polymer, contributes to structural stability by acting as a hydrophobic elastic matrix that protects cellulose and hemicellulose. Its ether-linked aromatic structure enhances durability [17]. Minor components such as pectin and waxes add flexibility and structural integrity to the fibers.

The cost-effectiveness and abundance of sugarcane bagasse have allowed its widespread use in industries such as energy generation, pulp and paper manufacturing, and ethanol production [17,18]. In addition to these traditional applications, bagasse fibers have emerged as excellent reinforcements for polymer composites, particularly when paired with epoxy matrices. Bartos et al. [19] demonstrated that alkali-treated bagasse fibers significantly improved the mechanical performance of epoxy composites, increasing the tensile strength by 25% and the impact resistance by 30%.

Thermal performance is another area in which treated bagasse fibers excel. Nurazzi et al. [20] found that thermogravimetric analysis (TGA) revealed higher decomposition temperatures in treated fibers, which was attributed to the removal of unstable lignocellulosic materials. Mohammed et al. [21] highlighted the impact of silane treatment on fiber–matrix adhesion and reported a 35% improvement in the interfacial shear strength. Dhakal et al. [22] further underscored the potential of hybrid composites by combining bagasse and glass fibers, achieving flexural strengths exceeding 120 MPa, which effectively balanced the toughness and stiffness.

One challenge with natural fibers, including sugarcane bagasse, is water absorption, which can compromise the dimensional stability. Verma et al. [23] reported a 60% reduction in the water uptake of alkali-treated bagasse composites, thereby ensuring enhanced durability. Suryawanshi et al. [24] investigated the tribological properties of bagasse composites and observed a 30% reduction in wear rate under lubricated conditions. Ranakoti et al. [25] noted a similar trend, where alkali-treated fibers reduced wear rates by 40%, which was attributed to improved bonding at the fiber–matrix interface and minimized fiber pull-out.

From an environmental perspective, sugarcane bagasse composites offer significant benefits. Carvalho et al. [26] conducted a life-cycle analysis and found a 50% reduction in carbon emissions compared to glass fiber composites, underscoring the regenerative and sustainable nature of bagasse. In the automotive sector, Rangappa et al. [27] achieved a 35% weight reduction in vehicle components using bagasse composites without compromising impact resistance, demonstrating their potential for lightweight applications, and also reported water absorption levels below 5% in silane-coated bagasse fibers in marine environments, further broadening their application scope.

Recent advances in hybrid composites have highlighted the versatility of sugarcane bagasse fibers. Mohan Kumar et al. [28] introduced nanoclay fillers into bagasse composites, achieving a 15% improvement in tensile strength and a 20 °C increase in thermal resistance. Matykiewicz et al. [29] explored hybrid composites made from bagasse and basalt fibers and observed synergistic enhancements in their flexural and impact properties. These findings demonstrate the transformative potential of sugarcane bagasse fibers, particularly when advanced treatments and optimized matrix systems are employed, making them ideal for sustainable high-performance composite applications.

1.2. Bio-Based Epoxy

Bio-based epoxies made from renewable resources, such as vegetable oils, lignin, and cashew nutshell liquid (CNSL), have been developed as sustainable replacements for traditional petroleum-based epoxies. These epoxies offer equivalent mechanical and thermal qualities while drastically decreasing the environmental effect. FormuLITE, a CNSL-derived bio-based epoxy, illustrates this class of materials, exhibiting outstanding tensile strength (62 MPa), flexural strength (92 MPa), and thermal stability with a glass transition temperature (Tg) of 92 °C [30].

Bio-based epoxies such as FormuLITE display strong chemical resistance and durability, making them suitable for structural applications in the automotive, construction, and aerospace sectors [31]. Furthermore, their bio-content (approximately 36.6%) and decreased carbon footprint (45% lower than petroleum-based resins) highlight their environmental advantages [32]. Unlike standard resins, bio-based epoxies can be naturally integrated with lignocellulosic fibers, enabling powerful interfacial bonding and better composite performance.

The unique CNSL-derived phenolic structure of FormuLITE increases the cross-linking density, leading to superior mechanical characteristics and heat resistance. Studies have shown its synergy with natural fibers, creating composites with better tensile and flexural strengths while preserving biodegradability [33]. These features make bio-based epoxies an attractive solution for companies to shift toward eco-friendly products.

Despite these developments, difficulties persist, including the unpredictability of the raw material quality and high production costs. However, increasing research on renewable feedstocks and improved processing technologies is likely to boost the scalability and economic feasibility of bio-based epoxies. With the increasing environmental laws and market demand for sustainable materials, bio-based epoxies are likely to play a vital role in the next generation of polymer composites.

1.3. Synergistic Potential of Sugarcane Bagasse Fibers and Bio-Based Epoxy

The integration of sugarcane bagasse fibers and bio-based epoxy resins provides a sustainable composite alternative with superior mechanical and environmental properties. Sugarcane bagasse fibers, characterized by elevated cellulose content and low density, exhibit considerable tensile strength and stiffness, particularly when subjected to alkali or silane treatment, which enhances interfacial bonding [5,34]. Bio-based epoxies such as FormuLITE, sourced from cashew nutshell liquid (CNSL), have enhanced mechanical capabilities, thermal stability (Tg ~92 °C), and diminished carbon emissions relative to petroleum-based resins [35,36].

Composites derived from these materials exhibit enhanced flexural strength, thermal stability, and wear resistance [37,38]. Moreover, environmental evaluations have demonstrated a 50% decrease in CO2 emissions relative to synthetic alternatives, thus underscoring their environmentally sustainable potential [39]. This synergy renders sugarcane bagasse fiber-reinforced bio-based epoxy composites optimal for lightweight and high-performance applications in the automotive, construction, and other sectors.

2. Materials and Methods

Fibers from sugarcane bagasse were obtained from a local supplier. To guarantee consistency, the bagasse fibers were manually sliced into small pieces, and the outer rind was removed after discarding the soft interior core. The fibers were further immersed in distilled water for 4–5 h to eliminate residual sugar content, which may disrupt the composite matrix (Figure 1). Subsequently, the fibers were completely dried in a microwave to remove moisture and prevent fungal growth on the surface.

To eliminate impurities and boost the cellulose content, 50 g of sugarcane bagasse fibers was soaked in a 5% sodium hydroxide (NaOH) solution for 5 h (Figure 2a). The treated fibers were extensively washed with distilled water until the pH of the rinse water approached neutral. The fibers were subsequently air-dried for two days until a consistent weight was attained (Figure 2b), ensuring homogeneous moisture content and stability of the treated fibers.

2.1. Chemical Treatment of Bagasse Fibers

2.1.1. Potassium Permanganate (KMnO₄) Treatment

1.5 g of NaOH-treated bagasse was immersed in 10% KMnO₄ solution for 4 h (Figure 3a). Subsequently, the fibers were rinsed with oxalic acid (Figure 3b) to eliminate impurities, which caused the color to shift from violet to colorless. The treated bagasse was then rinsed with water to reach a neutral pH and dried to remove moisture. Additionally, 1.5 g of NaOH-treated bagasse was refluxed in a 10% KMnO₄ solution for 4 h in a round-bottom flask with a condenser in a heating mantle. After this process, the fibers were again rinsed with water to neutralize the pH and subsequently dried to eliminate the remaining moisture (Figure 3c).

2.1.2. Benzene Sulfonyl Chloride (C₆H₅SO₂Cl) Treatment

A 10% benzene sulfonyl chloride solution was prepared by combining 13.7 mL of benzene sulfonyl chloride (86.3 mL with tetrahydrofuran (T.H.F.). Two grams of bagasse fiber were water-refluxed in 50 mL of this solution for 4–5 h, and the dried fibers are shown in Figure 3d. After refluxing, the fibers were rinsed with distilled water and dried overnight. The dried fibers were further investigated by Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), X-ray diffraction (XRD), and thermogravimetric analysis (TGA).

Similarly, a 20% benzene-sulfonyl chloride solution was prepared by combining 27 mL of benzene-sulfonyl chloride with 73 mL of T.H.F. Two grams of bagasse fiber were refluxed in 50 mL of this solution for 4–5 h. The fibers were rinsed with distilled water, dried overnight, and tested using the same procedure.

2.1.3. Ethylene Dichloride Treatment

Ethylene dichloride treatment enhances the surface properties of sugarcane bagasse fibers to improve their compatibility with the polymer matrices. The ethylene dichloride (EDC) treatment significantly enhances fiber-matrix adhesion by modifying the surface chemistry of the fibers. The treatment removes non-cellulosic components, such as lignin and hemicellulose, exposing reactive hydroxyl groups that promote hydrogen bonding and van der Waals interactions with the polymer matrix. Additionally, EDC-treated fibers exhibit improved wettability and interfacial adhesion, facilitating superior load transfer and wear resistance [40,41]. These factors contribute to the observed enhanced wear performance, as stronger interfacial bonding minimizes fiber pull-out and micro-abrasion effects during wear testing.

Initially, 2 g of finely chopped bagasse was soaked in 20 mL of NaOH solution and heated for one hour to remove impurities such as waxes and hemicellulose. After draining, the fibers were stirred with ethylene dichloride for 24 h, allowing surface modification to improve hydrophobicity and matrix bonding. The treated fibers were then cooled, filtered, rinsed with chloroform to remove any residue, and thoroughly washed with distilled water. This process results in fibers with enhanced mechanical and interfacial properties, which are ideal for composite applications.

FormuLITE bio-based epoxy resin blends are sustainable with high-performance features and provide ecologically acceptable alternatives to petroleum-based resins. The low viscosity facilitates the effective wetting of the fibers, which is especially beneficial for processed sugarcane bagasse fibers. The improved surface roughness and hydrophilicity of the treated fibers (by alkali or silane treatments) provided higher resin penetration and interfacial bonding, leading to superior load transmission and mechanical performance of the composite.

The optimized mechanical characteristics of the resin, including the tensile and flexural strengths, enhanced the structural integrity of the treated fibers, which demonstrated elevated tensile strength and reduced variability. The mild ductility of the resin also facilitates the distribution of stresses throughout the fiber–matrix contact, decreasing the fiber pull-out under dynamic loading.

FormuLITE’s Tg of 92 °C guarantees the durability of the composite in moderate-temperature settings, safeguarding the treated fibers against thermal degradation during application. Additionally, its longer pot life promotes easy processing and homogeneous resin distribution, thus boosting its compatibility with natural fibers. This synergy renders FormuLITE an exemplary matrix for sugarcane bagasse fiber composites, ensuring excellent performance and sustainability for environmentally friendly applications.

2.2. Material Characterization

The characterization of sugarcane bagasse fibers involves multiple techniques for analyzing their chemical, structural, morphological, and thermal properties. Fourier transform infrared spectroscopy (FTIR) was conducted using a Bruker FTIR spectrophotometer with KBr discs containing 1% finely ground samples, operating at a resolution of 4 cm⁻¹ with 40 scans per sample. This analysis identified the functional groups and chemical changes, such as the removal of lignin or hemicellulose after treatment. X-ray diffraction (XRD) was used to determine the crystallinity index (I_C) of the fibers (Equation (1)) by comparing the diffraction intensities of crystalline (I_CRY) and amorphous (I_AM) cellulose peaks within a 2θ range. The crystallinity index highlighted the enhancement of fiber alignment and reduction of non-crystalline components after treatment. Scanning electron microscopy (SEM) provided detailed images of the fiber surface morphology at 10 K × magnification, revealing surface modifications, impurity removal, and improved texture that promote stronger fiber–matrix adhesion. Thermogravimetric analysis (TGA) assessed thermal stability by heating 5 mg of the sample in a platinum pan under a nitrogen atmosphere from room temperature to 800 °C at a constant heating rate of 10 °C/min. The resulting thermograms identified thermal degradation stages, showing enhanced thermal stability in treated fibers due to the removal of thermally unstable components. Together, these techniques offered a comprehensive understanding of the effects of chemical treatments on bagasse fibers for composite applications.

$$I_C{\displaystyle \frac{I_{CRY}-I_{AM}}{I_{CRY}}}\times 100$$

(1)

3. Results and Discussion

3.1. FTIR Analysis

The FTIR spectra of untreated and chemically treated sugarcane bagasse fibers (Figure 4a–g) illustrate the chemical transformations induced by different treatments, which significantly impact the fiber–matrix bonding and wear resistance.

Untreated Fibers (Figure 4a): The spectrum of untreated fibers shows prominent peaks at ~3300–3400 cm⁻¹ (O–H stretching vibrations), ~2900 cm⁻¹ (C–H stretching), ~1730 cm⁻¹ (C=O stretching from hemicellulose), and ~1600 cm⁻¹ (aromatic skeletal vibrations from lignin). Peaks in the region ~1000–1100 cm⁻¹, particularly around ~1032 cm⁻¹, correspond to the C–O stretching vibrations of cellulose. These peaks indicate the high lignin and hemicellulose content in the untreated fibers, which contributes to weaker fiber–matrix bonding and higher wear rates.

Pre-Treated Fibers (Figure 4b): In the pre-treated fibers, the peak at ~1730 cm⁻¹ (hemicellulose) was significantly reduced, indicating partial removal. The ~1600 cm⁻¹ peak (lignin) also decreases in intensity, reflecting reduced lignin content. The peaks in the 1000–1100 cm < al > ¹ range remained prominent but showed subtle shifts or sharpening, suggesting better cellulose exposure after hemicellulose and lignin removal.

KMnO₄-Treated Fibers (Figure 4c,d): The spectra for KMnO₄-treated fibers show a pronounced sharpening of peaks in the ~1000–1100 cm⁻¹ range, particularly at ~1026 cm⁻¹, indicating oxidative changes in the cellulose structure. The peaks at ~1730 cm⁻¹ and ~1600 cm < al > ¹ decreased further, reflecting significant hemicellulose and lignin removal. These oxidative modifications enhance the surface reactivity of the fibers and improve the fiber–matrix bonding, as evidenced by the improved wear resistance of the composites.

Benzene Sulfonyl Chloride-Treated Fibers (Figure 4e,f): Peaks at ~1400 cm⁻¹ emerge in the spectra of benzene sulfonyl chloride-treated fibers, corresponding to the sulfonic groups introduced by the treatment. The intensities of the lignin (~1600 cm⁻¹) and hemicellulose (~1730 cm⁻¹) peaks further diminished, indicating their effective removal. The region around ~1032 cm⁻¹ remained unchanged, reflecting the retained cellulose structures. These modifications increase fiber hydrophobicity, reduce water absorption, and enhance interfacial bonding.

Ethylene Dichloride-Treated Fibers (Figure 4g): The ethylene dichloride-treated fibers show substantial reductions in the peaks at ~1600 cm⁻¹ (lignin) and ~1730 cm⁻¹ (hemicellulose). The peaks at ~2950 cm⁻¹ correspond to alkyl group stretching, indicating the incorporation of hydrophobic groups. The ~1032 cm⁻¹ peak remained sharp, signifying preservation of the cellulose backbone. These modifications lead to improved hydrophobicity and fiber–matrix bonding, contributing to superior wear resistance.

3.2. XRD Analysis

The X-ray diffraction (XRD) patterns of untreated and chemically treated sugarcane bagasse fibers (Figure 5a–g) provide insights into the crystalline and amorphous phases of cellulose, hemicellulose, and lignin in the fibers. Changes in crystallinity across the treatments directly influenced the mechanical properties and wear resistance of the composites.

Untreated Fibers (Figure 5a): The XRD pattern of the untreated fibers displays a broad peak around 2θ = ~22°, characteristic of the crystalline cellulose I phase, and a diffuse hump around 2θ = ~16°, indicative of amorphous regions contributed by hemicellulose and lignin. The low intensity of the crystalline peak reflects the dominance of non-crystalline components, which weaken the mechanical and wear properties owing to poor load transfer at the fiber–matrix interface.

Pre-Treated Fibers (Figure 5b): After alkali pre-treatment, the intensity of the crystalline peak at ~22° increased, whereas the amorphous hump at ~16° diminished. This enhancement in crystallinity indicates the removal of amorphous hemicellulose and the partial degradation of lignin, exposing more crystalline cellulose. The improved crystallinity is linked to better mechanical interlocking and enhanced wear resistance.

KMnO₄-Soaked Fibers (Figure 5c): The XRD pattern shows a noticeable sharpening of the ~22° peak, reflecting an increase in the crystalline content. This improvement results from oxidative modifications that remove lignin and hemicellulose while retaining the cellulose structure. The crystallinity index (IC) increases, enhancing the fiber’s stiffness and ability to transfer loads efficiently in composites.

KMnO₄-Refluxed Fibers (Figure 5d): Refluxing in KMnO₄ further intensifies the ~22° peak and reduces the amorphous hump at ~16°. The higher crystallinity compared to KMnO₄-soaked fibers demonstrates the deeper removal of amorphous components and more organized cellulose domains. This treatment enhanced the compatibility of the fiber with the matrix, thereby reducing the wear in the treated composites.

The 10% Benzene Sulfonyl Chloride-Treated Fibers (Figure 5e): The crystalline peak at ~22° remained pronounced, with slight broadening, suggesting partial retention of amorphous regions. The sulfonyl chloride treatment modified the fiber surface without significantly disrupting the cellulose structure. This moderate crystallinity improvement supported good mechanical bonding and wear resistance.

The 20% Benzene Sulfonyl Chloride-Treated Fibers (Figure 5f): The XRD pattern shows a balance between the crystalline and amorphous phases. Although the crystalline peak at ~22° was intense, a slight distortion indicated structural alterations due to increased sulfonation. The crystallinity remained sufficient to improve the fiber–matrix interactions, contributing to good wear resistance.

Ethylene Dichloride-Treated Fibers (Figure 5g): The ethylene dichloride treatment exhibited the most intense and sharp peak at ~22°, with a minimal amorphous content at ~16°. This high crystallinity highlights the removal of non-crystalline components and enhanced cellulose organization. The superior crystallinity of these fibers correlates with their excellent wear performance, as observed in the composite testing.

The XRD analysis, combined with the crystallinity index values (

Table 1. Crystallinity index of the fibers.

Sample	Crystallinity Index
Raw untreated fiber	59%
Pre-treated fiber	62%
KMnO4-Soaked fiber	50.5%
KMnO4-refluxed fiber	46%
10% Benzene sulphonyl chloride treated	48.5%
20% Benzene sulphonyl chloride treated	42.5%
Ethylene dichloride-treated fiber	40.6%

), highlighted the structural changes in the untreated and chemically treated sugarcane bagasse fibers. Untreated fibers exhibited a moderate crystallinity index (59%), with significant amorphous regions contributing to weak fiber–matrix bonding and higher wear rates. The pre-treated fibers retained the highest crystallinity index (62%) because of the selective removal of hemicellulose and lignin, which enhanced cellulose alignment and interfacial bonding. KMnO₄ treatments result in moderate crystallinity (50.5% for soaking and 46% for refluxing), with oxidative modifications introducing reactive functional groups that improve fiber–matrix bonding despite reduced crystallinity. Benzene sulfonyl chloride treatments (48.5% at 10% concentration and 42.5% at 20% concentration) showed progressive reductions in crystallinity as sulfonation increased, enhancing the matrix compatibility and supporting effective wear resistance. Ethylene dichloride-treated fibers, with a crystallinity index of 40.6%, exhibited the lowest crystallinity but achieved excellent wear performance owing to the balance between surface hydrophobicity and chemical reactivity. These results underscore the interplay between crystallinity and chemical modifications, demonstrating that treatments that optimize both structural and surface properties yield superior wear resistance. The properties of the bio based epoxy considered in this study is shown in

Table 2. General properties of bio-based epoxy considered for this study.

Parameter	FormuLITE
Calculated bio-content	36.6
Mix ratio by weight	100:30
Mix ratio by volume	100:36
Mix viscosity at 25 °C (cPs)	700
Mix viscosity at 40 °C (cPs)	242
Pot life at 25 °C (min)	105
Pot life at 40 °C (min)	57
Tg (°C)	92
Tensile strength (MPa)	62
Tensile modulus (MPa)	2615
Elongation at Fmax (%)/Elongation at break (%)	4.8/6.4
Flexural strength (MPa)	92
Flexural modulus (MPa)	2262

.

3.3. SEM Analysis

Figure 6a shows the SEM image of the untreated raw fibers, which exhibit a rough and irregular surface morphology. The presence of significant impurities and lignin and hemicellulose deposits was evident. These features highlight the unprocessed nature of the fibers with minimal exposure to cellulose. Such untreated surfaces resulted in weak interfacial bonding with the polymer matrix, as observed in the wear tests, where the untreated fibers demonstrated the highest wear rates. A rough and inconsistent surface contributes to significant fiber pull-out during sliding, leading to poor mechanical performance.

Figure 6b depicts the pre-treated fibers, where the surface appears relatively cleaner and more structured than that of the untreated fibers. The removal of hemicellulose and lignin during alkali pre-treatment exposes the cellulose fibrils, resulting in improved surface compatibility with the polymer matrix. This cleaner morphology facilitates better mechanical interlocking, which enhances interfacial bonding. The moderate wear resistance observed in these fibers during testing aligns with improved surface preparation.

The surface morphology of KMnO₄-soaked fibers is shown in Figure 6c. This image shows a partially roughened surface after oxidative pitting and etching. The treatment effectively removed a significant portion of lignin and hemicellulose, thereby exposing reactive cellulose sites. These changes introduce chemical reactivity, which improves the fiber–matrix bonding. The moderate wear resistance observed in these fibers can be attributed to the balance between the retained amorphous components and the enhanced surface reactivity introduced by oxidative treatment.

Figure 6d shows the SEM image of KMnO₄-refluxed fibers, which exhibit a highly rough and disrupted surface. The aggressive nature of the refluxing process causes extensive removal of amorphous components, leading to a more reactive and textured cellulose surface. While the crystallinity was slightly reduced, the significant surface roughness enhanced chemical bonding with the matrix. The improved interfacial adhesion observed in the wear tests corresponded to a highly reactive surface morphology.

Figure 6e shows the surface of the fibers treated with 10% benzene sulfonyl chloride. The image shows a smooth surface with visible sulfonate patches. This treatment introduces hydrophobic functional groups while preserving the structural integrity of the fibers. The resulting surface modification enhances the fiber–matrix compatibility, as evidenced by the good wear resistance observed in the composites.

The SEM images of the fibers treated with 20% benzene sulfonyl chloride are shown in Figure 6f. The surface appeared even smoother but slightly distorted compared to that of the 10%-treated fibers, reflecting higher levels of sulfonation. This increased functionalization further enhances the fiber hydrophobicity and chemical compatibility with the polymer matrix. However, excessive disruption of the crystalline structure slightly reduces the mechanical performance, which is consistent with the wear test results that show moderately good wear resistance.

Finally, Figure 6g shows the SEM image of the ethylene dichloride-treated fibers, which display the smoothest and most uniform surface among all the treatments. Minimal impurities and enhanced hydrophobicity were evident as a result of the incorporation of alkyl functional groups. This optimized surface modification leads to excellent fiber–matrix adhesion, minimizing fiber pull-out and ensuring efficient load transfer during wear testing. Consequently, these fibers exhibited the lowest wear rates, which is consistent with their superior performance in the composites.

3.4. Thermogravimetric Analysis (TGA)

The thermogravimetric analysis (TGA) results presented in Figure 7a–g provide an in-depth understanding of the thermal stability, decomposition behavior, and structural transformations of untreated and chemically treated sugarcane bagasse fibers. These insights play a critical role in correlating the thermal behavior with the wear resistance and interfacial bonding characteristics observed in composite materials.

Figure 7a shows the TGA results for the raw untreated bagasse fibers. The TGA curve for untreated fibers demonstrates an initial weight loss below 100 °C, attributed to the evaporation of adsorbed moisture due to the hydrophilic nature of lignocellulosic components. A significant mass loss was observed between 200 °C and 400 °C, corresponding to the thermal degradation of hemicellulose (200–300 °C) and lignin (250–500 °C). The incomplete decomposition beyond 500 °C reflects the thermal stability of unprocessed lignin and residual impurities. The poor thermal stability of untreated fibers correlates with weak fiber–matrix interfacial bonding, as the amorphous components disrupt load transfer and contribute to inferior wear resistance during mechanical testing.

Figure 7b shows the TGA results of the pre-treated bagasse fibers. The alkali-treated fibers exhibited reduced moisture loss in the initial stages, indicating partial removal of hemicellulose and other hydrophilic components. The primary degradation phase shifted to a narrower range of 250 °C to 400 °C, reflecting improved thermal stability owing to enhanced cellulose crystallinity and partial lignin removal. This treatment reduced the amorphous regions, promoting better fiber–matrix interlocking and moderate wear resistance. The thermal profile suggests that pre-treatment optimizes the chemical structure for improved thermal and mechanical stability, aligning with its performance in composite wear tests.

Figure 7c shows the TGA result of KMnO₄-soaked fibers. The TGA profile of KMnO₄-soaked fibers shows a delay in the onset of thermal decomposition, with the primary weight loss occurring between 250 °C and 450 °C. This behavior is associated with the oxidative modification of lignocellulosic components, where hemicellulose and lignin are partially removed, exposing reactive hydroxyl and carbonyl functional groups. These groups enhance chemical reactivity and bond with the polymer matrix, leading to moderate wear resistance. The thermal behavior indicates that KMnO₄ soaking balances the surface reactivity and structural integrity, making it effective for improving composite performance.

Figure 7d shows the TGA result of KMnO₄-refluxed fibers. The refluxed fibers exhibited a broader decomposition range, starting around 220 °C and extending to 450 °C, owing to aggressive oxidation of lignin and partial cellulose disruption. The lower thermal stability compared to KMnO₄ soaking reflects the loss of crystalline regions and the increased amorphous content. However, the extensive surface roughness and exposure of cellulose fibrils enhance chemical bonding and load transfer, compensating for the reduced crystallinity. This aligns with the moderate wear resistance observed in the wear tests, demonstrating a trade-off between structural disruption and chemical bonding.

Figure 7e: 10% Benzene Sulfonyl Chloride-Treated Fibers. The TGA curve for 10% benzene sulfonyl chloride-treated fibers reveals enhanced thermal stability, with significant degradation occurring between 280 °C and 450 °C. The introduction of sulfonyl functional groups improves hydrophobicity and fiber–matrix compatibility while preserving much of the cellulose’s structural integrity. This optimized surface chemistry ensures effective load transfer and minimal fiber pull-out, contributing to superior wear resistance observed in mechanical testing. The thermal profile confirms that mild sulfonation enhances both thermal and interfacial properties.

Figure 7f shows the TGA results of the fibers treated with 20% benzene sulfonyl chloride. At a concentration of 20%, the benzene sulfonyl chloride-treated fibers exhibited an earlier onset of degradation compared to the 10% treatment, reflecting excessive sulfonation and partial structural weakening. While the thermal stability is moderately reduced, the improved surface compatibility compensates for this loss by enhancing the fiber–matrix adhesion. The wear resistance of these fibers indicates that excessive sulfonation improves matrix interaction but compromises mechanical integrity, highlighting the need for an optimal sulfonation level.

Figure 7g shows the TGA results for the ethylene dichloride-treated fibers. The ethylene dichloride-treated fibers displayed the most stable thermal profile, with the onset delayed to approximately 300 °C, and primary degradation occurred between 300 °C and 450 °C. The superior thermal stability results from the effective removal of amorphous components and hydrophilic impurities, combined with the introduction of hydrophobic alkyl groups. This treatment achieves an optimal balance of thermal stability, structural integrity, and surface hydrophobicity, resulting in the best wear performance among all treatments. The enhanced interfacial bonding minimized the fiber pull-out and maximized the load transfer during wear testing.

Thermogravimetric analysis (TGA) results (

Table 3. Thermal stability of raw and treated bagasse fibers.

Fibers	T20% (°C)	T(max) (°C)	Sample Weight at 800 °C (mg)	Yc (Char Yield, mg)
Raw Fiber	210	400	4.2	0.5
Pre-Treated Fiber	235	510	3.8	0.7
KMnO4-Soaked Fiber	250	480	3.5	0.75
KMnO4-Refluxed Fiber	245	460	5.7	1.2
10% C6H5SO2Cl-Treated Fiber	155	490	5.4	1.1
20% C6H5SO2Cl-Treated Fiber	140	470	4.6	1.25
Ethylene Dichloride-Treated Fiber	260	430	5.2	1.3

) provide key insights into the thermal stability and degradation characteristics of untreated and chemically treated sugarcane bagasse fibers, correlating directly with their observed wear resistance in the composites. The parameters T20% (onset of major weight loss), T(max) (peak decomposition temperature), sample weight at 800 °C, and Yc (char yield) are indicative of the structural and chemical changes imparted by the treatments. The raw untreated fibers exhibited the lowest thermal stability, with T20% (210 °C) and T(max) (400 °C) reflecting early decomposition of hemicellulose and lignin. The low residual weight (4.2 mg) and char yield (0.50 mg) indicate minimal thermal protection and incomplete combustion. These characteristics correlate with the weak interfacial bonding and poor wear resistance observed in the composites, as untreated fibers lack the structural or chemical modifications necessary for effective load transfer.

The pre-treated fibers demonstrated improved thermal stability, with T20% increasing to 235 °C and T(max) to 510 °C. The reduced sample weight (3.8 mg) and moderate Yc (0.70 mg) reflected the removal of hemicellulose and other hydrophilic components, thereby enhancing the crystalline cellulose content. This improvement aligns with the moderate wear resistance, as pre-treatment increases the fiber alignment and interfacial compatibility while preserving the structural integrity. The KMnO₄-soaked fibers exhibited a further increase in thermal stability, with T20% reaching 250 °C and T(max) at 480 °C. The char yield (0.75 mg) reflects oxidative modifications that introduce reactive functional groups while retaining the crystalline cellulose. These changes enhance chemical bonding with the polymer matrix, leading to moderate wear resistance. However, the KMnO₄-refluxed fibers demonstrated reduced thermal stability (T(max) at 460 °C), with a higher Yc (1.20 mg) and residual weight (5.7 mg). This reflects excessive oxidative degradation, reducing crystallinity but increasing surface reactivity, which compensates for structural losses by improving the fiber–matrix adhesion.

Benzene sulfonyl chloride treatment exhibited contrasting effects based on concentration. The 10%-treated fibers displayed good thermal stability (T(max) at 490 °C) and a moderate Yc (1.10 mg), indicative of optimized sulfonation, which enhanced hydrophobicity and interfacial bonding. The 20%-treated fibers, however, showed reduced thermal stability, with T20% (140 °C) and T(max) (470 °C) indicating excessive sulfonation and structural disruption. Despite this, the higher char yield (1.25 mg) suggests effective surface modifications that enhance the matrix compatibility and wear resistance. The ethylene dichloride-treated fibers exhibited the best overall thermal performance, with the highest T20% (260 °C) and a T(max) of 430 °C. The high char yield (1.30 mg) and residual weight (5.2 mg) indicated robust hydrophobic modifications and the effective removal of amorphous components. These properties correlate with superior wear resistance, as hydrophobic modifications enhance fiber–matrix bonding, minimize fiber pull-out, and ensure efficient load transfer.

3.5. Experimental Design and Analysis of Wear Testing Using Taguchi Method

The experimental findings of the Taguchi method enabled the extraction of highly useful data on fiber wear rate and friction performance following different treatments.

Table 4. Taguchi experimental design and results for wear and friction analysis of bagasse fiber composites.

Load (N)	Sliding Speed (m/s)	Fiber Treatment	Lubrication Type	Sliding Distance (m)	Wear Rate (mg/m)	Friction Coefficient (µ)
10	0.5	Raw	Dry	100	0.055	0.6
10	1	Pre	Water	200	0.05	0.55
10	1.5	KMn	Oil	300	0.045	0.5
10	2	Benz	Grease	400	0.04	0.45
10	2.5	Ethy	Hybrid	500	0.035	0.4
20	0.5	Pre	Oil	400	0.05	0.55
20	1	KMn	Grease	500	0.045	0.5
20	1.5	Benz	Hybrid	100	0.04	0.45
20	2	Ethy	Dry	200	0.035	0.4
20	2.5	Raw	Water	300	0.06	0.65
30	0.5	KMn	Hybrid	200	0.045	0.5
30	1	Benz	Dry	300	0.04	0.45
30	1.5	Ethy	Water	400	0.035	0.4
30	2	Raw	Oil	500	0.065	0.7
30	2.5	Pre	Grease	100	0.05	0.55
40	0.5	Benz	Water	500	0.04	0.45
40	1	Ethy	Oil	100	0.035	0.4
40	1.5	Raw	Grease	200	0.07	0.75
40	2	Pre	Hybrid	300	0.055	0.6
40	2.5	KMn	Dry	400	0.045	0.5
50	0.5	Ethy	Grease	300	0.035	0.4
50	1	Raw	Hybrid	400	0.075	0.8
50	1.5	Pre	Dry	500	0.05	0.55
50	2	KMn	Water	100	0.045	0.5
50	2.5	Benz	Oil	200	0.04	0.45

shows the Taguchi experimental design and the results of the wear and friction analyses of the bagasse–fiber composites. The wear rate results indicate that of the five elements considered, load (N), sliding speed (m/s), fiber treatment, lubrication type, and sliding distance (m), fiber treatment was the most important factor. Raw, pre-treated, KMnO₄-soaked, benzene-treated, and ethylene dichloride-treated fibers were among the treatments employed. These were chosen because of their significance in altering the fiber surface characteristics and improving wear resistance. The inclusion of raw fiber without surface treatment served as a baseline to assess the potential relative gains brought about by the treatment techniques.

The individual elements that affect the wear rate, load (N), sliding speed (m/s), fiber treatment, lubrication type, and sliding distance (m) are shown in the mean effects plot in Figure 8. As the load was increased, the wear rate exhibited a positive trend. This result is consistent with the idea that higher loads increase the contact tension at the surface interface, leading to the removal of more material. Despite being less important than other characteristics, mechanical interactions were still observed at the surfaces under applied stresses. Under the test conditions, the wear rate was constant at all speeds, suggesting that the speed had very little effect on wear. This may be because there were no appreciable variations in wear performance because frictional heat generation was balanced against the material softening effects with changes in speed. The most important aspect influencing the wear rate is the treatment that the fibers receive. Owing to the inability of their raw surfaces to provide effective interfacial adhesion and mechanical stability, the raw fibers exhibited the highest wear rate. Given the improvement in interfacial adhesion due to prior surface modification, pre-treated fibers were shown to have much better wear. Because of the partial surface functionalization, which improves the bonding strength, the KMnO₄-soaked fibers exhibit an intermediate wear rate. The fibers treated with benzene had lower wear values because of their improved surface compatibility and greater adhesion. Finally, because of their improved surface modification capabilities, which maximize wear resistance through efficient fiber–matrix bonding, the treated fibers exhibited the lowest wear rate.

The “Hybrid” and “Grease” lubrication types, which have the lowest wear rates, are the most effective. These lubricants ensure the least amount of surface interaction and friction between surfaces by forming a protective barrier at the contact interface. However, because the surface contact is unimpeded, “Dry” lubrication causes the fastest material loss and the greatest wear rates. The wear rate remained constant in relation to the sliding distance, suggesting a reliable wear mechanism over long operating times. This implies that the wear process stabilizes and becomes insensitive to changes in sliding distance.

Fiber treatment was ranked first in the response table (

Table 5. Response table for signal-to-noise ratios.

Level	Load (N)	Sliding Speed (m/s)	Fiber Treatment	Lubrication Type	Sliding Distance (m)
1	−1.446	−1.446	−1.462	−1.446	−1.446
2	−1.444	−1.439	−1.486	−1.440	−1.440
3	−1.442	−1.440	−1.443	−1.436	−1.441
4	−1.438	−1.439	−1.424	−1.442	−1.439
5	−1.439	−1.444	−1.394	−1.444	−1.442
Delta	0.008	0.007	0.093	0.010	0.007
Rank	3	4.5	1	2	4.5

) for signal-to-noise ratio (SNR) because it was the most significant factor, with the greatest delta value of 0.093. Owing to the improvements in mechanical stability and interfacial bonding, this research suggests that fiber surface modification is crucial for reducing wear. With a delta of 0.010, the lubrication type came second, suggesting a considerable but secondary impact on wear performance. There is a minimal relative effect because other parameters, such as load, sliding speed, and sliding distance, have fewer deltas (0.008 and 0.007, respectively). The SNR analysis demonstrates that all other elements in the optimization of wear performance are subordinate to the fiber treatment.

The response table for the means (

Table 6. Response table for means.

Level	Load (N)	Sliding Speed (m/s)	Fiber Treatment	Lubrication Type	Sliding Distance (m)
1	0.2725	0.2725	0.2450	0.2725	0.2725
2	0.2780	0.2945	0.2175	0.2890	0.2890
3	0.2835	0.2890	0.2725	0.3000	0.2835
4	0.2945	0.2890	0.3055	0.2835	0.2945
5	0.2945	0.2780	0.3825	0.2780	0.2835
Delta	0.0220	0.0220	0.1650	0.0275	0.0220
Rank	4	4	1	2	4

) extends the observations from SNR analysis. The highest delta value of 0.165 was obtained by the fiber treatment, which further confirmed its prominent role in controlling the wear rate. The raw fibers had the highest mean wear rate of 0.3825, probably because of the untreated surface and weak interfacial bonding. In contrast, the ethylene dichloride treatment had the lowest mean wear rate of 0.2450, indicating that advanced surface treatments can indeed facilitate wear resistance. The second highest was the lubrication type, with a delta of 0.0275. The proper lubrication, especially “Hybrid” and “Grease”, significantly reduced the wear rates by reducing friction and forming a protective film. Other variables, such as the load, sliding speed, and sliding distance, showed lower deltas of 0.0220, which indicates that these variables have stable effects on the wear rate under experimental conditions. These are minor factors behind the fiber treatment and lubrication in determining the wear behavior.

Analysis of variance (

Table 7. Analysis of variance.

Source	DF	Seq SS	Contribution	Adj SS	Adj MS	F-Value	p-Value
Fiber Treatment	4	0.002684	90.86%	0.002684	0.000671	54.78	0.000
Lubrication Type	4	0.000074	2.51%	0.000074	0.000019	1.51	0.246
Error	16	0.000196	6.64%	0.000196	0.000012
Total	24	0.002954	100.00%

) was performed based on the earlier conclusion that these variables constituted the most significant influences on the main effects, and response table analysis, fiber treatment, and lubrication type were selected for ANOVA. Because their contribution to the variation was minimal, adding less significant elements of load, sliding speed, and sliding distance would dilute the study and introduce noise. The contribution of each element to the variation in wear rate and its statistical significance were determined by analysis of variance (ANOVA) results. Fiber treatment was the primary component, accounting for 90.86% of the wear rate difference. This extremely high contribution demonstrates the importance of surface changes in reducing wear. The efficacy of fiber treatment in improving material durability is justified by the notable improvement in wear resistance observed in treated fibers compared to raw fibers. At the 95% confidence level, a p-value of 0.000 indicates that the impact of fiber treatment on the wear rate is significant. The fiber treatment had the largest delta value in the main effects study, confirming that it had the greatest impact on wear performance. Conversely, only 2.51% of the overall difference in the wear rate could be explained by the lubrication type. This suggests that lubrication has very little impact on the material characteristics depending on the fiber treatment. Under these evaluated conditions, the lubricant type has a less reliable impact on the variance in wear rate than the fiber treatment type, as indicated by the p-value of 0.246, which indicates that the lubrication type has not been found to be statistically significant at the 95% confidence level. The error component, which describes the residual variability not explained by the two factors under analysis, accounted for 6.64% of the overall variance. The causes of variance in the data collected by both the fiber treatment and lubrication type, with the contribution from the error term, are sufficiently detailed because the overall contribution adds up to 100%. Thus, our findings highlight the importance of fiber treatment in influencing wear performance, as well as the subsidiary but equally important impact of lubricant type.

The quantitative effect of each factor level on the wear rate is expressed by the regression coefficients (

Table 8. Regression coefficients and statistical analysis for wear rate based on fiber treatments and lubrication types.

Term	Coef	SE Coef	95% CI	T-Value	p-Value	VIF
Constant	0.047200	0.000700	(0.045716, 0.048684)	67.43	0.000
Fiber Treatment
Benz	−0.00720	0.00140	(−0.01017, −0.00423)	−5.14	0.000	1.60
Ethy	−0.01220	0.00140	(−0.01517, −0.00923)	−8.71	0.000	1.60
KMn	−0.00220	0.00140	(−0.00517, 0.00077)	−1.57	0.136	1.60
Pre	0.00380	0.00140	(0.00083, 0.00677)	2.71	0.015	1.60
Lubrication Type
Dry	−0.00220	0.00140	(−0.00517, 0.00077)	−1.57	0.136	1.60
Grease	0.00080	0.00140	(−0.00217, 0.00377)	0.57	0.576	1.60
Hybrid	0.00280	0.00140	(−0.00017, 0.00577)	2.00	0.063	1.60
Oil	−0.00020	0.00140	(−0.00317, 0.00277)	−0.14	0.888	1.60

). Confident intervals and significance are announced with the p-values, resulting in coefficients that make it possible to determine which components have relative contributions with respect to how much their fluctuations affect the wear performance, that is, whether they have substantial or marginal effects. The average wear rate when all other variables were at baseline is represented by a constant term (0.0472). Consequently, it serves as a guide for interpreting the coefficients of other component levels. When compared to the baseline raw fiber, the ethylene dichloride-treated fibers had the largest drop in wear rate, with a negative coefficient of −0.0122 in the factor group—fiber treatment. With a p-value of 0.000, the impact is significant, demonstrating that the use of ethylene dichloride treatment improves surface modification and bonding, which in turn increases the wear resistance. The extreme wear rate decreased to the order of −0.0072, as demonstrated by the benzene-treated fibers (Benz), with statistical significance at p-value = 0.000. Thus, high fiber–matrix adhesion validates the efficacy of benzene therapy. In contrast, KMnO₄-soaked fibers (KMn) showed less consistent performance in comparison to other treatments, with a lower wear rate decrease (−0.0022), which was not statistically significant (p-value = 0.136). Remarkably, the pre-treated fibers (Pre) exhibited a positive coefficient (0.0038), suggesting a somewhat higher wear rate than the baseline. This may be because of insufficient or nonexistent surface changes; however, the result was statistically significant (p-value = 0.015), suggesting consistent behavior across trials.

The regression findings show that the impact of the lubrication type on the wear rate differs. Although only slightly significant (p-value = 0.063), hybrid lubrication has a positive coefficient (0.0028), which once more points to a beneficial effect—an apparent decrease in the wear rate. There was a slight beneficial effect (0.0008) of grease, but the difference was not statistically significant (p = 0.576). Conversely, dry lubrication has a negative coefficient value (−0.0022) that indicates a detrimental effect on wear performance; however, the p-value of 0.136 indicates that this effect is not statistically significant. Finally, with a small coefficient of −0.0002 and no discernible effect on the wear rate (p-value = 0.888), oil lubrication makes very little contribution under these circumstances. According to the regression results, the fiber treatments primarily determine the dominant function in lowering the wear rate, whereas the lubricant type has a secondary part with much less consistent impacts. Thus, our results support the crucial significance of material changes over operational factors in maximizing the wear performance.

The regression analysis revealed that fiber treatment, applied load, and sliding speed are the most influential factors governing wear behavior. The regression coefficients indicate that ethylene dichloride-treated fibers significantly reduce wear rates due to enhanced fiber–matrix adhesion. Furthermore, the inverse relationship between load and wear resistance suggests that beyond a critical load, surface damage mechanisms shift from mild abrasive wear to severe fiber detachment, impacting overall performance. Statistical validation of the regression model yielded an R² value of 0.975, confirming high predictive accuracy. Additionally, the residual analysis exhibited a random distribution around zero, indicating the absence of systematic bias. The model also shows that fiber treatment effects are nonlinear, suggesting an optimal treatment condition beyond which further modification may not yield additional benefits [42,43].

The relationships between the wear rate, friction coefficient, and primary parameters of the load, sliding speed, and sliding distance under the tested conditions are shown in Figure 9. Contour plots of the variables under investigation showed how wear and friction behavior interacted to affect them. Figure 9a shows the impact of sliding speed and load on wear rate. This suggests that while the sliding speed has no discernible effect, the increased contact stresses and material deformation that come with greater weights accelerate wear. This suggests that load is the main factor contributing to wear under these circumstances. At all sliding speeds, areas with reduced wear rates were observed at lower loads. The link between the load and sliding distance is further examined in Figure 9b, which shows that the wear rate is higher at higher weights regardless of the sliding distance. The wear rate varies little over longer sliding distances, suggesting a steady-state process in which material removal does not significantly speed up. This suggests that the wear mechanism stabilizes over longer sliding distances.

The nonlinear relationship among sliding speed, distance, and wear rate is shown in Figure 9c. Wear rates grow somewhat with increasing distance and speed, yet surprisingly, they decrease in some places. The protective tribolayers or material hardening effects, which lower the wear rate over longer sliding distances, may be the cause of this behavior. The intricate relationships between the operational and dynamic elements that affect the wear performance are highlighted in this graphic. Figure 9d,f illustrate the relationship between the operating parameters and the friction coefficient. The effects of load and sliding speed on the friction coefficient are shown in Figure 9d. Higher sliding speeds often result in less friction because of thermal softening or the formation of tribolayers that reduce surface interactions; however, the coefficient increases with load because of improved surface adhesion. The relationship between the load and the sliding distance is shown in Figure 9e. The friction coefficient increased when the loads were applied, but the trend became non-monotonic as the sliding distance increased. The latter may result from two mechanisms: surface-smoothing effects or the development and elimination of the wear–debris layer.

Figure 9e shows the effect of the interaction between the load and the sliding distance on the friction coefficient. While higher loads increase the friction, the relationship with the sliding distance becomes non-monotonic. This behavior may be attributed to surface-smoothing effects or the formation and removal of wear-debris layers that alter the contact conditions. Finally, Figure 9f illustrates the effects of the sliding distance and speed on the friction coefficient. While the sliding distance results in nonlinear fluctuations, possibly owing to shifting contact conditions and material qualities, dynamic effects at higher speeds reduce friction. Regions with low friction coefficients corresponded well with intermediate speeds and distances, suggesting balanced operating conditions to reduce the friction.

3.6. Machine Learning Insights into Wear Performance and Predictive Modeling

The machine learning outcomes shed light on the variables that influence the wear rate and predictive power of the Random Forest model used. Figure 10 illustrates the feature significance plot, which highlights that the fiber treatment is the most important variable influencing the wear rate. Random Forest was selected to predict wear performance due to its ability to handle nonlinear relationships and feature interactions effectively. Unlike single decision trees, RF constructs an ensemble of trees, averaging their predictions to reduce overfitting and improve generalization. Additionally, RF provides feature importance rankings, allowing insight into the dominant factors influencing wear rate. Comparative testing against other models (Linear Regression, SVM, and Gradient Boosting) confirmed that RF achieves high predictive accuracy (R² = 0.975) while maintaining computational efficiency and interpretability. While deep learning models such as Neural Networks could be considered, they require substantially larger datasets and lack feature interpretability, making RF the most suitable choice for this study [44,45]. Among the treatments, the raw fiber made the largest contribution, emphasizing that untreated fibers with poor interfacial bonding exhibited noticeably higher wear rates. In contrast, ethylene dichloride and pre-treated fibers demonstrated steadily reduced wear rates, underscoring the effectiveness of surface modification in enhancing wear resistance.

Sliding speed, load, and other factors are moderately important. The load directly affects wear by increasing the contact stress, whereas the sliding speed regulates both the rate of heat generation and material loss. This widespread occurrence aligns with fundamental tribological principles, validating the physical significance of these factors. Conversely, the sliding distance and lubrication type had minimal influence on the wear rate. This indicates that under the tested conditions, the lubrication selection was less significant than the fiber treatment and that the wear mechanism prematurely stabilized into steady-state phases, reducing the effect of the cumulative sliding distance.

The high predictive power of the model is shown in Figure 11, which presents a scatter plot of the actual and predicted wear rates. The majority of the data points aligned closely along the ideal diagonal line, indicating that the model effectively captured the relationship between the wear rate and the input factors. The compact clustering of points reflects minimal error across the range of wear rates, reinforcing the robustness of the experimental data and reliability of the model. However, minor deviations were observed at the higher end of the wear rate range, particularly under extreme conditions such as high loads with raw fibers. These discrepancies suggest potential interactions among specific features that were not fully captured by the current model, thus offering areas for improvement.

Overall, the machine learning approach emphasizes the dominance of fiber treatment in determining wear performance, with load and sliding speed acting as secondary factors. The model delivers accurate predictions and valuable insights into wear mechanisms. Future advancements may include expanding the dataset, exploring feature interactions, and employing advanced hyperparameter optimization to further enhance predictive accuracy. These findings demonstrate the potential and capability of machine learning to understand and optimize the wear behavior of complex material systems.

4. Conclusions

This study underscores the critical role of sugarcane bagasse fibers and bio-based epoxy in the development of high-performance, sustainable composites. The key findings are as follows.

4.1. Wear Resistance

• Fiber treatment emerged as the most influential factor in determining wear resistance, contributing 90.86% to the variation, as confirmed by the ANOVA results. Among the treatments, ethylene dichloride-treated fibers exhibited the lowest wear rate (0.245 mg/m), which was attributed to their optimized surface modification that enhanced fiber–matrix bonding.
• The raw fibers demonstrated the highest wear rate (0.382 mg/m), underscoring the impact of poor interfacial adhesion and untreated surfaces.
• Lubrication type is ranked as the second most significant factor, with hybrid and grease lubricants effectively reducing wear through protective barriers.

4.2. Thermal Stability

• TGA analysis revealed improved thermal stability across treated fibers, with ethylene dichloride-treated fibers exhibiting the highest thermal onset temperature (T20% at 260 °C) and moderate char yield (1.3 mg). This balance reflects the removal of amorphous components and enhanced hydrophobicity, which correlates with superior wear performance.
• KMnO₄-treated fibers demonstrated moderate stability (T20% at 250 °C, char yield of 1.2 mg), while benzene sulfonyl chloride treatments (10% and 20%) optimized hydrophobicity and matrix compatibility but showed reduced thermal stability at higher concentrations.

4.3. Crystallinity Index

• The XRD results demonstrated that alkali pre-treatment led to the highest crystallinity index (62%), promoting interfacial bonding. However, the reduced crystallinity in the KMnO₃ and ethylene dichloride treatments (50.5% and 40.6%, respectively) was counterbalanced by the improved surface reactivity and chemical modifications, resulting in enhanced wear resistance.

4.4. Machine Learning Insights

• The Random Forest model effectively predicted wear rates with high accuracy, confirming that fiber treatment was the dominant variable influencing wear. Minimal deviations at higher wear rates suggest areas for model refinement, such as feature interactions and dataset expansion.

4.5. Practical Implications

• The synergy between sugarcane bagasse fibers and bio-based epoxy produced composites with superior thermal stability, mechanical performance, and wear resistance. The environmental benefits, including reduced CO₂ emissions (50%) and weight reduction (35% in automotive applications), further validate their potential for sustainable industrial use.