Heating Performance and Flexural Strength of Steel Fiber–Carbon Nanotube Cement Composites for Black Ice Prevention

Abstract

Cementitious composites incorporating multi-walled carbon nanotubes (MWCNTs), known for their superior mechanical, electrical, and thermal properties, present significant potential for multifunctional infrastructure applications. This study quantitatively evaluates the heat generation behavior, electrical resistivity, and flexural performance of hybrid cementitious composites reinforced with MWCNTs and steel fibers. A total of 72 specimens were tested following ASTM and KS standards, considering MWCNT concentrations of 0.0, 0.5, and 1.0 wt% and a steel fiber content of 2.0 vol%. The results revealed that increasing MWCNT concentration markedly enhanced heat generation, reaching a maximum temperature rise of 80.1 °C, while electrical resistance decreased by over 99% compared with basic mortar. The inclusion of steel fibers slightly reduced heat generation but improved flexural strength by up to twofold due to fiber bridging and crack control. Microstructural analyses (FE-SEM, XRD, and TGA) confirmed the formation of continuous CNT networks that facilitated electron transport and improved matrix densification. Although the findings are based on laboratory-scale specimens, the combined use of MWCNTs and steel fibers offers a promising pathway for developing self-heating, mechanically enhanced cementitious materials applicable to black-ice prevention and durable pavement systems.

1. Introduction

Black ice is a critical safety concern that can cause large-scale traffic accidents during winter. It forms when water or melted snow penetrates the pores of asphalt or concrete and subsequently freezes, producing a thin, nearly invisible layer that increases the risk of vehicle skidding and collisions. Statistical analyses have indicated that accidents related to black ice occur more frequently and result in greater casualties than those caused by heavy snowfall. Conventional de-icing methods, such as calcium chloride treatment, can temporarily remove surface ice but accelerate surface deterioration and chloride-induced corrosion, while contributing to environmental pollution and maintenance challenges. Consequently, recent research has focused on developing self-heating construction materials as a proactive strategy for ice prevention rather than post-treatment.

Multi-walled carbon nanotubes (MWCNTs), known for their excellent electrical and thermal conductivities, have shown strong potential to enhance the multifunctionality of cementitious composites [1,2,3,4,5]. Incorporating MWCNTs into cement matrices can facilitate electron transfer through conductive networks and improve heat generation when an external voltage is applied. Compared with other types of carbon nanotubes (CNTs), MWCNTs are more cost-effective and suitable for large-scale infrastructure applications.

For pavement systems, self-heating cementitious materials must provide not only efficient heat generation but also sufficient mechanical strength to withstand service loads [6,7,8,9,10]. Pavement design standards stipulate that the flexural strength of cement concrete should exceed 4.5 MPa [11,12,13,14]. Excessive heat accumulation, however, may induce temperature gradients between the interior and surface of the concrete, potentially leading to cracking depending on the thermal and mechanical properties of the material. Once cracking occurs, the continuity of conductive CNT networks may be disrupted, resulting in reduced electrical and heating performance. Maintaining flexural strength is therefore essential to sustain stable heating behavior under repeated thermal loading.

Hybrid reinforcement using MWCNTs and steel fibers offers a promising solution to address both performance requirements simultaneously. MWCNTs enhance electrical conductivity and Joule-heating efficiency by forming percolated conductive pathways, while steel fibers contribute to crack bridging and load transfer, improving flexural performance and durability. The synergistic use of both reinforcements is expected to achieve a balanced enhancement in thermal functionality and structural reliability.

Previous research has primarily focused on improving the heating performance of CNT-based cementitious composites. Zhang et al. [15] investigated the de-icing and snow-melting capabilities of cement composites, reporting that incorporating 3.0 wt% MWCNTs relative to cement mass facilitated the formation of thermally conductive layers. Li et al. [16] examined cement-based heating pavement systems with MWCNT concentrations of 1 wt%, 3 wt%, and 5 wt%, finding that 3 wt% MWCNTs achieved the highest thermal conductivity, whereas higher concentrations reduced performance due to dispersion challenges. Gomis et al. [17] analyzed the heating performance of cementitious composites containing carbon-based nanomaterials, reporting that specimens with 5 wt% CNTs reached a maximum temperature rise of 65 °C within 33 min. Kim et al. [18] studied thermal and electrical properties of CNT-incorporated cementitious composites at concentrations ranging from 0.1 wt% to 2.0 wt%, observing a maximum temperature rise of 67.8 °C at 2.0 wt%. Lee et al. [19,20,21,22] evaluated the performance of MWCNT cement composites under sub-zero conditions, showing that specimens with 1.0 wt% MWCNTs generated 51.4 times more heat than those with 0.1 wt%. Wei et al. [23] investigated thermoelectric characteristics of cement composites with CNTs at 5.0, 10.0, and 15.0 wt%, demonstrating substantial improvements in both electrical and thermal conductivity. Farcas et al. [24] assessed the heating performance of conductive cement pastes and concretes incorporating CNTs and graphite powder, identifying optimal performance in specimens containing 1% CNTs and 5% graphite powder.

Despite these advancements, most previous studies have emphasized heating performance while overlooking the mechanical strength requirements essential for practical pavement applications. Furthermore, limited research has addressed the simultaneous optimization of electrical, thermal, and mechanical properties in hybrid cementitious systems. The present study investigates the heating performance, electrical resistance, and flexural strength of hybrid cementitious composites incorporating MWCNTs and steel fibers. By analyzing the interaction between conductive nanomaterials and macro-scale fibers, the study seeks to identify an optimal combination that ensures both effective heat generation and structural stability, thereby addressing the existing research gap between thermal functionality and mechanical performance in self-heating cementitious materials. The reduction in electrical resistance by incorporating CNTs into cement has also been investigated. Li et al. [25] demonstrated that adding CNTs to cement paste significantly decreased electrical resistance and enhanced compressive sensitivity by leveraging the piezoelectric effect. Yu et al. [26] reported piezoelectric resistance sensitivity by disrupting contact between CNT particles, whereas higher CNT concentrations enhanced sensitivity. Gong et al. [27] examined the piezoelectric performance of CNT-reinforced cement composites under varying CNT concentrations and found that piezoelectric performance peaked at 0.3 vol% CNTs. Analyzing the effects of MWCNT concentration and water-to-cement ratio on piezoelectric resistance in cement composites, Han et al. [28] found that variations in piezoelectric resistance were proportional to MWCNT concentration, but beyond a certain threshold concentration, sensitivity decreased. Kim et al. [29] evaluated piezoelectric resistance sensitivity and stability in CNT cement mortars with low water-to-binder ratios, showing that reduced water content improved stability. Jang et al. [30] studied the influence of moisture on electrical conductivity, reporting that lower moisture increased porosity and decreased conductivity. Park et al. [31] assessed cement composites containing pitch-based carbon fibers and CNTs, finding that while electrical resistance was comparable to CNT-only composites, viscosity improved by 90%, enhancing workability. Yoo et al. [32] examined cement composites with various carbon nanomaterials (CNTs, graphite, and graphene oxide) and found that 1 wt% incorporation reduced electrical resistance in all cases, with CNTs exhibiting the most pronounced effect.

Beyond enhancements in heating performance and electrical conductivity, the incorporation of CNTs also improves mechanical strength. Li et al. [33] conducted strength tests and SEM analyses of cement composites incorporating MWCNTs, revealing that interfacial bonding and pore refinement improved mechanical performance. Konsta-Gdoutos et al. [34] studied microstructure and mechanical properties of cement composites incorporating MWCNTs, emphasizing that effective dispersion of MWCNTs is essential for mechanical enhancement and requires the use of optimally proportioned surfactants. Kumar et al. [35] investigated the mechanical strength of cement paste incorporating MWCNTs at concentrations of 0.5 wt%, 0.75 wt%, and 1.0 wt% relative to cement mass and found that compressive and tensile strengths improved most significantly at 0.5 wt% MWCNTs. Sobolkina et al. [36] examined the effect of CNT dispersion methods on the strength of cement composites. CNT dispersion in aqueous solution increased compressive strength by up to 40%. Xu et al. [37] analyzed the mechanical properties and microstructure of cement paste incorporating MWCNT, reporting that higher MWCNT concentrations improved strength and refined pore distribution. Numerous studies have explored the use of carbon nanomaterials to enhance the mechanical strength of cement composites [38,39]. CNTs and graphene have been incorporated into cement mortars to evaluate their effects on mechanical properties and pore structures [40,41,42,43,44,45,46,47,48]. However, most of this work has concentrated on compressive strength and porosity, with comparatively little attention given to the relationship between flexural strength and heating performance.

The present study aims to evaluate the heat generation performance, electrical resistance, and flexural behavior of steel fiber–carbon nano cementitious composites, assessing their feasibility as self-heating materials for black-ice prevention in pavement applications. Previous research has primarily focused on enhancing either heating efficiency or mechanical strength through nanomaterial incorporation, without addressing their simultaneous optimization. For practical implementation in field-scale pavement systems, self-heating composites must achieve both sufficient heat generation to maintain surface temperatures above freezing and design-standard flexural strength to resist traffic loading [48,49]. To meet these dual requirements, the present study adopts a hybrid reinforcement approach that integrates multi-walled carbon nanotubes (MWCNTs) and steel fibers. The MWCNTs enhance electrical conductivity and Joule-heating efficiency, while the steel fibers improve load-bearing capacity and crack resistance. The selected proportions—0.0, 0.5, and 1.0 wt% MWCNTs with 2.0 vol% steel fibers—are based on previous findings indicating that higher MWCNT contents lead to dispersion instability and that excessive fiber volumes (>2.5 vol%) induce fiber balling and workability loss. This design strategy enables a balanced improvement in both thermal and structural performance, providing a foundation for the development of durable self-heating pavement materials.

Heating and electrical resistance experiments were conducted to characterize the performance of steel fiber–carbon nano cementitious composites. Experimental parameters included MWCNT concentration, steel fiber content, curing duration, and supply voltage, selected for their influence on heating performance. Microstructural analyses, including X-ray diffraction (XRD), thermogravimetric analysis (TGA), and field emission scanning electron microscopy (FE-SEM), were performed to elucidate the mechanisms underlying conductivity and to assess the effects of MWCNTs on cement hydration and composite microstructure.

2. Materials and Methods

2.1. Experimental Overview

2.1.1. Heating and Electrical Resistance Tests

Heating and electrical resistance tests were conducted using parameters expected to influence the electrical and thermal performance of steel fiber–carbon nano cementitious composites (

Table 1. Test parameters.

Specimen Name	MWCNT Concentration (wt%)	Steel Fiber Content (vol%)	Curing Days (Day)
MW0.0-SF0.0-7D	0.0	0.0	7
MW0.5-SF0.0-7D	0.5
MW1.0-SF0.0-7D	1.0
MW0.0-SF2.0-7D	0.0	2.0
MW0.5-SF2.0-7D	0.5
MW1.0-SF2.0-7D	1.0
MW0.0-SF0.0-28D	0.0	0.0	28
MW0.5-SF0.0-28D	0.5
MW1.0-SF0.0-28D	1.0
MW0.0-SF2.0-28D	0.0	2.0
MW0.5-SF2.0-28D	0.5
MW1.0-SF2.0-28D	1.0

). High multi-walled carbon nanotube (MWCNT) contents can cause agglomeration within the cement matrix due to van der Waals forces between particles, resulting in reduced performance. Therefore, MWCNT concentrations of 0.0 wt%, 0.5 wt%, and 1.0 wt% relative to the cement mass were selected [6,8,27]. To ensure uniform dispersion, functionalized MWCNTs were first dispersed in an aqueous solution using ultrasonic agitation before being added to the cement mixture. Excessive steel fiber addition may lead to fiber balling and reduced mechanical strength; hence, the steel fiber volume fraction was fixed at 2.0 vol% of the specimen volume. Furthermore, higher fiber content tends to reduce the flowability of fresh mixtures, promoting nonuniform fiber distribution and causing agglomeration that interferes with the uniform dispersion of MWCNTs and increases variability in electrical and thermal performance. Curing durations of 7 and 28 days were chosen to evaluate both early-age and fully cured heating performance. Resistance measurements were conducted after 28 days to minimize the influence of residual moisture. Heating performance was evaluated under both low- and high-voltage conditions by applying direct current (DC) voltages of 10 V, 20 V, 30 V, and 60 V for 1 h. In total, 72 specimens were fabricated, with six specimens prepared for each test condition. The results of the heating and resistance tests were reported as the average values of the corresponding specimens. The 7-day measurements were conducted to examine the early-age electrical and thermal behavior influenced by residual moisture and ongoing hydration prior to full curing.

The specimen nomenclature was based on the sequence of MWCNT concentration, steel fiber content, and curing duration. The first notation represented the MWCNT concentration relative to cement mass; for example, “MW0.0” indicated specimens without MWCNTs. The second denoted steel fiber content by volume, where “SF0.0” indicated no steel fibers and “SF2.0” indicated 2.0 vol% steel fibers. The third represented curing duration, with “7D” and “28D” referring to curing periods of 7 and 28 days, respectively.

2.1.2. Flexural Strength Test

Flexural strength tests of steel fiber–carbon nano cementitious composites were conducted using parameters known to influence mechanical performance. The experimental conditions for flexural strength testing were identical to those used in the electrical resistance tests. MWCNT concentrations were set at 0.0 wt%, 0.5 wt%, and 1.0 wt% relative to the cement mass, consistent with the parameters used in the heating performance tests. Steel fiber content was expressed as a volume percentage in accordance with KCS 14 20 22. To prevent fiber agglomeration and the associated reduction in strength, the steel fiber volume fraction was fixed at 2.0 vol% [28]. All specimens were cured for 28 days to determine their 28-day flexural strength. A total of 36 specimens were fabricated, with six specimens prepared for each parameter set. Flexural strength results were reported as the average of the six specimens. The specimen nomenclature followed the same convention as that used for the heating performance tests, indicating the MWCNT concentration, steel fiber volume fraction, and curing duration in sequence. Specifically, the first term represents the MWCNT concentration, the second denotes the steel fiber volume fraction, and the third indicates the curing period.

2.1.3. Internal Microstructural Analysis

Internal microstructural analysis was performed to evaluate the effects of MWCNT incorporation on cementitious composites. XRD and TGA were employed to analyze the influence of MWCNTs on hydration products, while FE-SEM was used to observe MWCNTs within the cementitious matrix.

2.2. Specimen Fabrication and Experimental Procedure

2.2.1. Heating and Electrical Resistance Test

Figure 1a shows a schematic of the specimens used in this study. Currently, no standardized national or international testing methods exist for evaluating the heating performance and electrical resistance of cementitious composites. Therefore, the specimen dimensions for the steel fiber–carbon nano cementitious composites were set to 50 mm × 50 mm × 50 mm, consistent with the dimensions recommended in ASTM C109, the international standard for compressive strength testing of cement mortars [50].

The fine aggregate used for specimen fabrication was standard sand (KSL ISO 679 [Standard 1], Methods of Testing Cements, Seoul, Republic of Korea, 2022), and ordinary Portland cement (OPC) served as the primary binder [29,30,31,50]. Multi-walled carbon nanotubes (MWCNTs) with a purity exceeding 99% were incorporated using a functionalized aqueous suspension dispersed by ultrasonic agitation to ensure homogeneous distribution within the cement matrix [32,33,34,35]. The steel fibers were bundled type with a length of 30 mm and an aspect ratio of 50, selected to improve flexural performance while maintaining workability. Stainless-steel mesh electrodes were embedded at both specimen ends for electrical connection, and a T-type thermocouple was positioned at the geometric center to record internal temperature during heating tests (

Table 2. Mix proportion of specimen.

Specimen Name	MWCNT (g)	Steel Fiber (g)
MW0.0-SF0.0-7D	0	0
MW0.5-SF0.0-7D	0.4
MW1.0-SF0.0-7D	0.8
MW0.0-SF2.0-7D	0	19.6
MW0.5-SF2.0-7D	0.4
MW1.0-SF2.0-7D	0.8
MW0.0-SF0.0-28D	0	0
MW0.5-SF0.0-28D	0.4
MW1.0-SF0.0-28D	0.8
MW0.0-SF2.0-28D	0	19.6
MW0.5-SF2.0-28D	0.4
MW1.0-SF2.0-28D	0.8

).

Figure 2 illustrates the overall fabrication process of the steel fiber–carbon nano cementitious composites. The MWCNT aqueous suspension, steel fibers, cement, and standard sand were prepared according to the specified mix proportions (Figure 2a). Cement and sand were first dry-mixed for 2 min to ensure homogeneity (Figure 2b). The functionalized MWCNT suspension, dispersed by ultrasonic agitation, was then added and mixed with the cement–sand blend for 3 min to promote uniform distribution of nanotubes within the matrix (Figure 2c). The resulting mixture was cast in three layers, and each layer was compacted 30 times to remove entrapped air. Stainless-steel mesh electrodes were embedded after placement of the first layer, maintaining a 20 mm spacing to prevent fiber agglomeration and electrical short-circuiting (Figure 2d). After electrode installation, the second and third layers were sequentially cast and compacted. A T-type thermocouple was embedded at a depth of 25 mm from the surface to monitor internal temperature during heating tests (Figure 2e). After casting, specimens were covered with plastic film to prevent moisture loss and were cured for 24 h at 23 ± 2 °C. Following demolding, they were oven-dried at 45 °C until the mass variation was less than 0.1% over 24 h, ensuring consistent moisture conditions before electrical and thermal testing (Figure 2f).

2.2.2. Flexural Strength Test

Figure 1b shows a schematic of the specimen used to investigate flexural strength. For the flexural strength evaluation, specimens without steel fibers were prepared in accordance with KS F 2408, whereas those containing steel fibers followed KS F 2566. All specimens were fabricated as prisms with dimensions of 100 mm × 100 mm × 400 mm [49,51,52].

Steel fiber–carbon nano cementitious composites were prepared using standard sand and ordinary Portland cement. Consistent with the heating performance tests, MWCNTs were incorporated in an aqueous solution to ensure uniform dispersion, and the same bundle-type steel fibers were employed. The cement-to-sand-to-water mass ratio was maintained at 1:2.5, identical to that used in the heating and electrical resistance tests.

The fabrication procedure for the flexural strength specimens (Figure 3) is as follows. The MWCNT solution, steel fibers, cement, and standard sand were measured according to the designated mix proportions (Figure 3a). Cement, sand, and steel fibers were first dry-mixed for 2 min to ensure homogeneity (Figure 3b). The MWCNT solution and water were then incorporated and mixed for 3 min (Figure 3c,d). Subsequently, the mixture was cast in three layers into molds, with compaction performed at each layer. After casting and compaction, the specimens were demolded after 24 h and dried in a 45 °C oven, consistent with the procedure for heating performance specimens (Figure 3e).

Flexural strength tests were conducted in accordance with ASTM C78 [10], KS F 2408 [51], and KS F 2566 [52]. Figure 4c shows the test setup. The test was performed using a stress-controlled four-point loading method, during which the mid-span deflection and fracture load of the flexural strength specimens were measured. Deflection was recorded using a universal testing machine, and the loading rate was controlled at 0.06 MPa/s until specimen failure. The flexural strength was calculated based on the specimen cross-sectional area, span length, and applied load at failure.

The heating performance of the steel fiber–carbon nano cementitious composites was evaluated by applying a constant direct current (DC) voltage to the specimens. Figure 4a presents the experimental setup. A DC power supply (EX-200, EXTECH Instruments, Nashua, NH, USA) provided voltages of 10, 20, 30, and 60 V, which were applied for 60 min through stainless-steel mesh electrodes embedded at both ends of each specimen. The electrodes were connected to the power supply using insulated copper leads and separated by a 20 mm gap to prevent electrical short-circuiting. The positive and negative terminals were connected through an insulated rubber plate to maintain stable contact and avoid leakage. The internal temperature was continuously monitored using a T-type thermocouple (accuracy ±0.1 °C) embedded at the geometric center of each specimen during fabrication and connected to a data logger (TDS-303, Tokyo Measuring Instruments Laboratory Co., Tokyo, Japan). Simultaneously, the surface temperature was captured using an infrared thermal camera (FLIR E8-XT, 7.5–13 μm, NETD ≤ 0.06 °C) positioned 0.5 m normal to the specimen surface. The camera emissivity was set to 0.95, and ambient conditions were maintained at 22 ± 2 °C. Thermal images were recorded at the time corresponding to the peak internal temperature measured by the data logger to evaluate the uniformity of heat distribution. Electrical resistance measurements were subsequently conducted using a digital multimeter (Keithley 2701, Tektronix Inc., Beaverton, OR, USA) as shown in Figure 4b. The multimeter’s terminals were connected directly to the embedded stainless-steel electrodes, and the resistance readings were recorded after stabilization. Each test was repeated three times, and the mean resistance value was reported to minimize random error.

2.2.3. Internal Microstructure Analysis

The internal microstructure of the steel fiber–carbon nano cementitious composites was analyzed using XRD, TGA, and FE-SEM. XRD was employed to identify and characterize the crystalline phases present in the composites. When X-rays interact with a crystal lattice, diffraction occurs at characteristic angles and intensities that reflect the material’s internal structure. In XRD analysis, the diffraction angle (2θ)—twice the incident angle (θ)—is measured to obtain structural information. The scanning range for 2θ was set between 10° and 70°. Comparative analyses of ordinary cement mortar and steel fiber–carbon nano cementitious composites were conducted to evaluate the influence of MWCNT incorporation on hydration products.

TGA was used to assess the mass changes in inorganic and organic constituents as a function of temperature. The resulting mass–temperature curve was used to identify the decomposition temperatures of various components. TGA of the steel fiber–carbon nano cementitious composites was performed under a nitrogen atmosphere, with a maximum temperature of 1100 °C and a heating rate of 10 °C/min.

3. Results

3.1. Heating Performance Test Results

Table 3. Heating generation performance results.

Specimen Name	Maximum Temperature Variation
	10 V	20 V	30 V	60 V
MW0.0-SF0.0-7D	$0.1 \pm$$0.0$	$0.1 \pm$$0.0$	$0.2 \pm$$0.0$	$0.3 \pm$$0.0$
MW0.5-SF0.0-7D	$1.5 \pm$$0.2$	$4.6 \pm$$0.3$	$9.2 \pm$$0.4$	$43.1 \pm$$1.8$
MW1.0-SF0.0-7D	$4.3 \pm$$0.3$	$15.7 \pm$$0.5$	$32 \pm$$1.0$	$80.1 \pm$$2.3$
MW0.0-SF2.0-7D	$0.1 \pm$$0.0$	$0.1 \pm$$0.0$	$0.1 \pm$$0.0$	$0.2 \pm$$0.0$
MW0.5-SF2.0-7D	$1.5 \pm$$0.2$	$4.4 \pm$$0.3$	$8.4 \pm$$0.4$	$31.5 \pm$$1.6$
MW1.0-SF2.0-7D	$4.0 \pm$$0.3$	$13.2 \pm$$0.6$	$25.6 \pm$$0.8$	$50.8 \pm$$2.1$
MW0.0-SF0.0-28D	$0.1 \pm$$0.0$	$0.1 \pm$$0.0$	$0.1 \pm$$0.0$	$0.2 \pm$$0.0$
MW0.5-SF0.0-28D	$0.9 \pm$$0.1$	$3.0 \pm$$0.2$	$7.1 \pm$$0.3$	$28.0 \pm$$1.2$
MW1.0-SF0.0-28D	$2.6 \pm$$0.2$	$10.1 \pm$$0.4$	$20.8 \pm$$0.7$	$52.8 \pm$$2.0$
MW0.0-SF2.0-28D	$0.1 \pm$$0.0$	$0.1 \pm$$0.0$	$0.1 \pm$$0.0$	$0.2 \pm$$0.0$
MW0.5-SF2.0-28D	$0.8 \pm$$0.1$	$2.8 \pm$$0.2$	$6.3 \pm$$0.3$	$22.1 \pm$$0.9$
MW1.0-SF2.0-28D	$2.5 \pm$$0.2$	$9.1 \pm$$0.4$	$16.5 \pm$$0.6$	$33.3 \pm$$1.5$

presents the results of the heating performance tests for steel fiber–carbon nano cementitious composites. The highest heat rise was observed for specimen MW1.0-SF0.0-7D, whose temperature reached 80.1 °C. Under identical conditions, except for curing duration, the temperature of specimen MW1.0-SF0.0-28D reached 53.0 °C, representing the highest heat generation among the 28-day cured specimens. Specimens without MWCNTs or steel fibers (MW0.0-SF0.0-7D and MW0.0-SF0.0-28D) showed a maximum temperature increase of only 0.3 °C, which was up to 79.9 °C lower than that of the other mixtures. Conventional cementitious materials exhibit negligible self-heating capacity and are therefore ineffective in preventing black ice. In contrast, specimens containing 1.0 wt% MWCNTs or more achieved temperature increases exceeding 50 °C, which is sufficient to melt snow and ice.

Figure 5 illustrates the time–temperature variation graphs at different supply voltages for specimens cured for 7 days. Each curve demonstrates a rapid initial temperature increase, reaching a peak, followed by stabilization. For the initial slope analysis, temperature changes after 10 min from the start of testing were examined. Figure 5a shows the results for a supply voltage of 10 V. Specimens MW0.0-SF0.0-7D and MW0.0-SF2.0-7D exhibited no temperature change after 10 min. Specimen MW0.5-SF0.0-7D, containing 0.5 wt% MWCNT, exhibited a temperature rise of 0.6 °C. The highest temperature increase after 10 min occurred in specimen MW1.0-SF0.0-7D, reaching 1.7 °C, which was 2.8 times greater than that of MW0.5-SF0.0-7D. For specimens with incorporated steel fibers (MW0.5-SF2.0-7D and MW1.0-SF2.0-7D), the 10 min temperature increases were 0.6 °C and 1.7 °C, respectively, identical to those observed in the corresponding specimens without steel fibers.

Figure 5b presents the time–temperature variation curves under a supply voltage of 20 V. Specimens without MWCNTs exhibited temperature changes of less than 0.1 °C after 10 min. Specimen MW0.5-SF0.0-7D showed a temperature rise of 2.0 °C after 10 min, while MW1.0-SF0.0-7D recorded the highest increase of 6.8 °C, approximately 3.4 times higher than that of MW0.5-SF0.0-7D. The 10 min temperature change for MW0.5-SF2.0-7D was 1.9 °C, comparable to that for MW0.5-SF0.0-7D. In contrast, MW1.0-SF2.0-7D showed a 10 min temperature rise of 5.6 °C, representing a 17.6% reduction compared with MW1.0-SF0.0-7D.

Figure 5c depicts the results for a supply voltage of 30 V. Specimens MW0.0-SF0.0-7D and MW0.0-SF2.0-7D displayed temperature changes of less than 0.1 °C after 10 min. Specimen MW0.5-SF0.0-7D showed a temperature rise of 4.3 °C, while MW1.0-SF0.0-7D exhibited the highest increase of 17.0 °C, approximately four times greater than that for MW0.5-SF0.0-7D. For specimens incorporating steel fibers, MW0.5-SF2.0-7D and MW1.0-SF2.0-7D recorded 10 min temperature increases of 3.5 °C and 11.6 °C, respectively. The incorporation of steel fibers reduced the 10 min temperature rise by 18.6% at 0.5 wt% MWCNT and by 31.8% at 1.0 wt% MWCNT, compared with the corresponding specimens without steel fibers.

Figure 5d presents the results for a supply voltage of 60 V. Specimens MW0.0-SF0.0-7D and MW0.0-SF2.0-7D each recorded a 10 min temperature change of only 0.1 °C. Specimen MW0.5-SF0.0-7D exhibited a 19.0 °C temperature rise, while MW1.0-SF0.0-7D showed the highest increase of 65.0 °C, approximately 3.4 times greater than that observed for MW0.5-SF0.0-7D. When steel fibers were incorporated, MW0.5-SF2.0-7D recorded a temperature increase of 13.9 °C, which was 26.8% lower than that for MW0.5-SF0.0-7D, whereas the temperature of MW1.0-SF2.0-7D reached 21.8 °C, representing a 66.5% reduction compared to MW1.0-SF0.0-7D.

Analysis of the time-dependent temperature variations for 7-day cured specimens indicates that specimens without MWCNTs exhibited negligible temperature changes, with initial temperature–time gradients converging to zero. In contrast, specimens containing MWCNTs presented steeper initial gradients that increased with higher MWCNT concentrations, as shown in Figure 6. The incorporated MWCNTs formed conductive networks among hydration products, and higher concentrations generated more continuous pathways that enabled greater current flow through the cement matrix. The temperature change was directly proportional to the electric current, resulting in higher temperature rises with increasing MWCNT content [35].

Even at identical MWCNT concentrations, the inclusion of steel fibers reduced the initial temperature–time gradient, and the degree of reduction increased with supply voltage and MWCNT content. The reduction occurred because the internal temperature rise increased the resistivity of steel fibers, restricting current flow and limiting further heat generation. Controlling temperature gradients is therefore essential, and the incorporation of steel fibers can function as an effective regulating mechanism that prevents excessive internal thermal stress.

Figure 6a illustrates the time–temperature variation of 28-day cured specimens under a supply voltage of 10 V. Specimens MW0.0-SF0.0-28D and MW0.0-SF2.0-28D exhibited temperature changes of less than 0.1 °C after 10 min, confirming the absence of an effective conductive network. Specimen MW0.5-SF0.0-28D recorded a 0.2 °C increase, while MW1.0-SF0.0-28D reached 0.6 °C, approximately three times higher, indicating that an increase in CNT concentration enhanced charge transfer efficiency through a percolated network.

Figure 6b presents the results at a supply voltage of 20 V. Specimens MW0.0-SF0.0-28D and MW0.0-SF2.0-28D again exhibited temperature changes below 0.1 °C after 10 min. Specimen MW1.0-SF0.0-28D recorded a temperature rise of 4.2 °C, approximately 3.5 times greater than that of MW0.5-SF0.0-28D. For specimens with 0.5 wt% MWCNTs, the inclusion of steel fibers produced minimal influence, while MW1.0-SF2.0-28D exhibited a 4.8% reduction compared with MW1.0-SF0.0-28D. The results suggest that the conductive contribution of steel fibers becomes less significant at moderate voltages where the CNT network dominates the current flow.

Figure 6c shows the results under a supply voltage of 30 V. Specimen MW0.5-SF0.0-28D recorded a temperature rise of 3.0 °C, while MW1.0-SF0.0-28D reached 8.8 °C, approximately 2.9 times greater. The inclusion of steel fibers reduced the 10 min temperature rise to 2.5 °C for MW0.5-SF2.0-28D (a 20.0% reduction) and to 6.9 °C for MW1.0-SF2.0-28D (a 21.6% reduction) compared with their respective fiber-free specimens. The trend reflects a transition toward a competitive–synergistic conduction mechanism, where steel fibers partially interrupt CNT pathways but simultaneously stabilize current flow by moderating excessive heat accumulation.

Figure 6d shows the results under a supply voltage of 60 V. The highest 10 min temperature change of 25.4 °C was recorded for MW1.0-SF0.0-28D. Specimen MW0.5-SF2.0-28D exhibited a temperature rise of 5.7 °C, 40.6% lower than MW0.5-SF0.0-28D, while MW1.0-SF2.0-28D recorded 11.4 °C, representing a 55.1% reduction compared with MW1.0-SF0.0-28D. The reduction in heating rate at elevated voltage conditions suggests that the increased resistivity of metallic fibers and the nonuniform heat distribution act as a self-regulating mechanism, preventing thermal concentration and potential cracking within the composite matrix.

Analysis of the time-dependent temperature variations in specimens cured for 28 days shows trends consistent with those observed for specimens cured for 7 days. Specimens without MWCNTs exhibited negligible temperature changes, characterized by very low initial gradients. In contrast, specimens containing MWCNTs displayed higher initial gradients with increasing MWCNT concentrations, attributable to the formation of more extensive CNT networks. However, compared with the 7-day cured specimens, the initial gradients of the 28-day cured specimens decreased by up to 160%, likely due to reduced pore water content as hydration progressed during the extended curing period [35]. As with the 7-day cured specimens, the inclusion of steel fibers reduced the initial gradient relative to specimens without fibers, and this reduction became more pronounced with increasing supply voltage and MWCNT concentration.

Figure 7 illustrates the heat generation performance of specimens cured for 7 days. For specimens without steel fibers, maximum heat generation consistently occurred in those containing 1.0 wt% MWCNTs across all supply voltages (Figure 7a). At 10 V, specimen MW1.0-SF0.0-7D exhibited a temperature increase of 4.3 °C, representing 43-fold and 2.9-fold increases relative to MW0.0-SF0.0-7D and MW0.5-SF0.0-7D, respectively. At 20 V, heat generation reached 15.7 °C, corresponding to 157-fold and 3.4-fold increases compared with specimens containing 0.0 wt% and 0.5 wt% MWCNTs, respectively. At 30 V, MW1.0-SF0.0-7D reached 32.0 °C, which was 160-fold and 3.5-fold higher than the temperatures of the respective controls. At 60 V, this specimen recorded 80.1 °C, 267-fold and 1.9-fold greater than MW0.0-SF0.0-7D and MW0.5-SF0.0-7D, respectively.

For specimens containing steel fibers, MW1.0-SF2.0-7D consistently exhibited the highest heat generation across all voltages (Figure 7b). At 10 V, MW1.0-SF2.0-7D reached 4.0 °C, which was 40-fold and 2.7-fold higher than MW0.0-SF2.0-7D and MW0.5-SF2.0-7D, respectively. At 60 V, the same specimen achieved 50.8 °C, corresponding to 254-fold and 1.6-fold increases compared to MW0.0-SF2.0-7D and MW0.5-SF2.0-7D, respectively.

Overall, heat generation increased with supply voltage due to higher current flow through the specimens, with the maximum temperature rise consistently observed at 60 V. Heat generation also increased with MWCNT concentration, with specimens containing 1.0 wt% MWCNTs demonstrating the highest performance. The formation of interconnected CNT networks facilitated electron transport within the composite matrix, thereby enhancing electrical conductivity and thermal response. Higher MWCNT concentrations promoted the development of more extensive networks, further improving heat generation efficiency. In contrast, specimens without MWCNTs (MW0.0-SF0.0-7D and MW0.0-SF2.0-7D) exhibited minimal heat generation of only 0.3 °C, due to the absence of conductive pathways. These results confirm that MWCNT incorporation is essential for imparting electrical and thermal functionality to cementitious composites.

However, the incorporation of steel fibers reduced the heat generation performance. At 10 V, specimens containing 0.5 wt% MWCNTs exhibited comparable heat generation regardless of the presence of steel fibers. In contrast, specimens with 1.0 wt% MWCNTs showed a decrease in the maximum heat generated when steel fibers were added. At 20 V, the specimen MW0.5-SF2.0-7D displayed a 4.3% reduction in heat generation relative to MW0.5-SF0.0-7D, while MW1.0-SF2.0-7D exhibited a 15.9% reduction compared to MW1.0-SF0.0-7D. At 60 V, the inclusion of steel fibers led to decreases of 26.9% and 36.6% in heat generation for specimens with 0.5 wt% and 1.0 wt% MWCNTs, respectively. The reduction in heat generation increased with both higher MWCNT concentrations and higher supply voltages. As heat generation is positively correlated with MWCNT content and applied voltage, specimens exhibiting higher intrinsic heat generation experienced larger reductions when steel fibers were incorporated. This effect is attributed to the increase in electrical resistance of metallic steel fibers at elevated temperatures, which limits current flow through the composites.

Figure 8 presents the heat generation performance of specimens cured for 28 days. Consistent with the results observed for specimens cured for 7 days, specimens without steel fibers exhibited the highest performance at an MWCNT concentration of 1.0 wt% across all applied voltages (Figure 8a). At 10 V, the temperature of MW1.0-SF0.0-28D reached 2.6 °C, representing increases of 26-fold and 2.9-fold compared to the temperatures of MW0.0-SF0.0-28D and MW0.5-SF0.0-28D, respectively. At 60 V, the temperature of MW1.0-SF0.0-28D reached 52.8 °C, corresponding to 237-fold and 1.9-fold increases relative to the temperatures of the respective control specimens.

For specimens containing steel fibers, specimen MW1.0-SF2.0-28D consistently exhibited the highest heat generation at all applied voltages (Figure 8b). At 10 V, the temperature of specimen MW1.0-SF2.0-28D reached 2.5 °C, which was 25 and 3.1 times greater than the temperatures of specimens MW0.0-SF2.0-28D and MW0.5-SF2.0-28D, respectively. At 60 V, the temperature reached 33.3 °C, corresponding to increases of 167 and 1.5 times relative to the temperatures of MW0.0-SF2.0-28D and MW0.5-SF2.0-28D, respectively. The consistent increase in heat generation with voltage and MWCNT concentration demonstrates that the percolated CNT network governs the primary conduction mechanism, while steel fibers provide secondary conduction pathways that stabilize current flow at higher voltages. The presence of metallic fibers also prevents local overheating by dissipating heat through their interconnected structure, indicating a self-regulating thermal behavior within the composite.

The heat generation trends of 28-day cured specimens were consistent with those of 7-day specimens regarding supply voltage, MWCNT concentration, and steel fiber incorporation; however, overall heating performance decreased with curing age. In specimens without MWCNTs, heat generation remained nearly unchanged regardless of curing duration, confirming that ionic conduction alone contributed minimally to heat generation. In specimens with 0.5 wt% MWCNTs, the maximum temperature decreased by approximately 15.1 °C compared with 7-day specimens, whereas specimens with 1.0 wt% MWCNTs exhibited a reduction of 27.3 °C. The decrease in heating performance with extended curing age can be attributed to the reduction in internal moisture, which initially assists current conduction during early hydration but diminishes as the matrix densifies and pore connectivity decreases.

Figure 9 presents the thermal images captured at the point of maximum heat generation for 7-day cured specimens under a 60 V supply. Surface temperatures were obtained by adding the generated heat to the initial ambient temperature of approximately 22.3 °C. The thermal images of specimens without steel fibers, shown in Figure 9a–c, confirm that specimens without MWCNTs generated negligible heat, with surface temperatures similar to the surroundings. The highest surface temperature, 102.5 °C, was recorded for specimen MW1.0-SF0.0-7D (Figure 9c), corresponding to a 57.9% increase compared with specimen MW0.5-SF0.0-7D.

Figure 9d–f displays the thermal images of 7-day cured specimens containing steel fibers. Specimens without MWCNTs exhibited minimal heat generation, with surface temperatures remaining close to the ambient temperature. The highest surface temperature, 73.0 °C, was measured for specimen MW1.0-SF2.0-7D (Figure 9f), representing a 36.4% increase compared with specimen MW0.5-SF2.0-7D. The reduction in peak surface temperature in steel-fiber-reinforced specimens compared with CNT-only specimens can be explained by the increased resistivity of steel fibers at elevated temperatures and the heat-dissipating effect of the metallic phase. The conductive interaction between CNT networks and steel fibers therefore establishes a competitive–synergistic mechanism, in which CNTs dominate early-stage heating through efficient electron transport, while steel fibers enhance thermal stability and prevent localized overheating at higher voltages.

Figure 10 presents thermal images of 28-day cured specimens captured at peak heat generation. Similar to the 7-day specimens, the maximum heating occurred under 60 V, with identical initial temperatures. For specimens without steel fibers, Figure 10a–c show that those lacking MWCNTs exhibited negligible heating, resulting in indistinct thermal images. The highest surface temperature was observed in specimens containing 1.0 wt% MWCNTs, reaching 75.3 °C–53.7% higher than specimens with 0.5 wt% MWCNTs (Figure 10c).

For 28-day cured specimens containing steel fibers, thermal images are presented beginning with Figure 10d. Consistent with the results for the 7-day cured specimens, specimens without MWCNTs showed minimal heating, with surface temperatures close to ambient. The maximum surface temperature occurred in specimens with 1.0 wt% MWCNTs, reaching 55.8 °C–25.7% higher than in specimens containing 0.5 wt% MWCNTs (Figure 10f).

Overall, the surface temperature of steel fiber–reinforced carbon nano-cementitious composites increased with MWCNT concentration, accompanied by clearer thermal images. Further analysis revealed that heat distribution was concentrated between the electrodes connected to the power supply and diffused outward. This pattern reflects the tendency of electric current to follow the shortest path, resulting in heat generation originating primarily in the electrode regions.

3.2. Electrical Resistance Test Results

Table 4. Electrical resistance results.

Specimen Name	Steel Fiber Content (vol%)	MWCNT Content (wt%)	Electrical Resistance (Ω)
MW0.0-SF0.0-28D	0.0	0.0	$\mathrm{50,549.0} \pm 3250.4$
MW0.5-SF0.0-28D		0.5	$990.2 \pm$$42.6$
MW1.0-SF0.0-28D		1.0	$384.9 \pm$$21.3$
MW0.0-SF2.0-28D	2.0	0.0	$\mathrm{68,330.5} \pm$$4180.2$
MW0.5-SF2.0-28D		0.5	$1489.1 \pm$$58.9$
MW1.0-SF2.0-28D		1.0	$731.7 \pm$$34.5$

summarizes the electrical resistance measurements for steel fiber–carbon nano cementitious composites cured for 28 days. The lowest resistance (384.9 Ω) was observed for specimen MW1.0-SF0.0-28D. Among the fiber-reinforced specimens, MW1.0-SF2.0-28D exhibited the lowest resistance of 731.7 Ω. Under a constant applied voltage, lower resistance corresponds to higher current flow through the composite, resulting in enhanced heating performance. Specimens without MWCNTs, MW0.0-SF0.0-28D and MW0.0-SF2.0-28D, exhibited substantially higher resistances of 50,549.0 Ω and 8330.5 Ω, respectively, indicating negligible electrical functionality.

Figure 11 presents the electrical resistance of the 28-day cured specimens. The resistances of MW0.5-SF0.0-28D and MW1.0-SF0.0-28D decreased by 98% and 99.2%, respectively, compared with that of MW0.0-SF0.0-28D. When steel fibers were incorporated, MW0.5-SF2.0-28D and MW1.0-SF2.0-28D exhibited reductions in resistance of 97.4% and 98.7% relative to the resistance of MW0.0-SF2.0-28D. The decrease in electrical resistance with increasing MWCNT content is attributed to the formation of more extensive CNT networks, which create additional conductive pathways.

Conversely, the inclusion of steel fibers increased the electrical resistance of steel fiber–carbon nano cementitious composites. The resistance of MW0.5-SF2.0-28D increased by 50.4% relative to that of MW0.5-SF0.0-28D, and at 1.0 wt% MWCNTs, the fiber-reinforced specimens exhibited a 90.1% higher resistance compared with the fiber-free specimens. The electrical and thermal performance of MWCNT-incorporated composites is highly sensitive to the dispersion quality. Nonuniform dispersion leads to local CNT agglomeration, increasing the resistance and diminishing the heating efficiency. Steel fibers reduce the workability of fresh composites, impeding uniform MWCNT dispersion; as a result, their incorporation contributes to higher electrical resistance.

3.3. Flexural Strength Test Results

Figure 12 presents the flexural strength results of steel fiber–reinforced carbon nano-cementitious composites. Among the specimens without steel fibers, MW1.0-SF0.0-28D exhibited the highest flexural strength (Figure 12a), measuring 3.1 MPa. This represents an increase of 0.4 MPa and 0.1 MPa compared with the flexural strengths of MW0.0-SF0.0-28D and MW0.5-SF0.0-28D, respectively. For fiber-reinforced specimens, MW1.0-SF2.0-28D achieved the highest flexural strength (Figure 12b), reaching 9.0 MPa, which is 1.3 MPa and 0.8 MPa greater than the corresponding strengths of the 0.0 wt% and 0.5 wt% MWCNT specimens, respectively. Overall, the flexural strength of the steel fiber–carbon nano-cementitious composites increased with higher MWCNT content. This enhancement is attributed to the formation of CNT networks among the hydration products, which reduces the porosity and improves the mechanical performance. As the MWCNT concentration increases, these networks become more extensive, resulting in higher flexural strength.

The inclusion of steel fibers increased the flexural strength compared to that of specimens without fibers. Specifically, the flexural strength of MW0.0-SF2.0-28D exceeded that of MW0.0-SF0.0-28D by 5.0 MPa. The addition of 0.5 wt% MWCNTs further increased the strength of fiber-reinforced specimens by 5.2 MPa relative to specimens without steel fibers. When 1.0 wt% MWCNTs were incorporated, the strength difference rose to 5.9 MPa. Steel fibers enhance flexural strength by improving the crack resistance and toughness; thus, flexural strength increased regardless of MWCNT concentration. In self-heating construction materials, thermal gradients between the interior and exterior can induce microcracking, which degrades the heating performance. The incorporation of steel fibers mitigates microcrack formation, thereby enhancing the durability of such materials.

In practical pavement engineering, the trade-off between heating efficiency and flexural performance can be optimized through layer-specific mix design strategies. The surface course can incorporate a lower steel fiber volume fraction and a higher MWCNT dosage to enhance Joule-heating efficiency and surface temperature rise, whereas the base or sublayer can employ a higher fiber volume fraction to ensure structural integrity and load-bearing capacity. Such a layered optimization framework provides a rational design basis for achieving both efficient black-ice mitigation and long-term mechanical reliability under service conditions.

3.4. Internal Microstructure Analysis Results

Figure 13 presents FE-SEM images of specimens MW0.0-SF0.0-28D, MW1.0-SF0.0-28D, and MW1.0-SF2.0-28D, with representative positions of MWCNTs indicated by red crosses. The specimen MW0.0-SF0.0-28D did not exhibit visible CNT filaments among the hydration products (Figure 13a), confirming the absence of conductive features within the plain cement matrix. In contrast, specimens containing MWCNTs (Figure 13b,c) showed localized CNT filaments bridging adjacent hydration products, suggesting the formation of micro-scale conductive links within the matrix. These qualitative observations indicate that MWCNTs can contribute to electron transport and interfacial bonding at the microscopic level. While SEM provides limited information confined to local regions, the overall dispersion uniformity was ensured through the ultrasonic aqueous dispersion procedure described in the Section 2. The observed CNT–hydration interfaces may also enhance local densification and stress transfer, supporting the improved mechanical and electrical response of the hybrid composite.

Figure 14 presents the XRD patterns of specimens MW0.0-SF0.0-28D, MW1.0-SF0.0-28D, and MW1.0-SF2.0-28D. All specimens exhibited diffraction peaks at 2θ = 20–21°, 39–40°, and 60–61°. The peaks located at 2θ = 20–21° correspond to calcium hydroxide (Ca(OH)₂) and silica (SiO₂), while the peak at 2θ = 39–40° is attributed to calcium carbonate (CaCO₃). The peak at 2θ = 60–61° represents calcium silicate hydrate (C–S–H), which is primarily responsible for mechanical strength development in hydrated cement matrices. Additional peaks at 2θ = 26–27° were observed in specimens MW1.0-SF0.0-28D and MW1.0-SF2.0-28D but were absent in MW0.0-SF0.0-28D. The appearance of the 2θ = 26–27° peak corresponds to the characteristic diffraction of MWCNTs [36,37], confirming their stable incorporation within the hydration matrix. The coexistence of C–S–H and the CNT-related peak suggests that the introduction of MWCNTs promoted nucleation and densification of hydration products without altering their fundamental crystalline phases. The CNTs likely acted as heterogeneous nucleation sites, accelerating C–S–H formation and enhancing interfacial bonding between hydration phases. The resulting microstructure supported improved mechanical integrity and facilitated electron transport through a denser matrix network. Overall, the XRD results confirm that MWCNT incorporation strengthened the internal bonding structure and contributed indirectly to the enhanced flexural and electrical performance of the hybrid composites.

Figure 15 presents the TGA curves of the specimens analyzed by FE-SEM and XRD. Specimens MW1.0-SF0.0-28D and MW1.0-SF2.0-28D exhibited pronounced mass losses within the temperature ranges of 400–450 °C, 550–700 °C, and 750–850 °C, whereas specimen MW0.0-SF0.0-28D showed significant mass losses only at 400–450 °C and 750–850 °C. The mass loss observed between 400 °C and 450 °C corresponds to the thermal decomposition of calcium hydroxide (Ca(OH)₂), and the mass loss between 750 °C and 850 °C is attributed to the decomposition of calcium carbonate (CaCO₃). Within the range of 550–700 °C, the TGA curves of specimens containing MWCNTs displayed a sharp mass reduction, in contrast to the gradual decline observed for the specimen without MWCNTs. The sharp loss within this range indicates the decomposition of carbonaceous materials, confirming the presence of thermally decomposed MWCNTs [38]. The additional mass loss in MWCNT-containing specimens reflects both the oxidation of CNTs and the concurrent release of chemically bound water from the C–S–H structure, suggesting a strong interfacial interaction between CNTs and hydration products. The thermal stability of the CNT-containing composites implies that the MWCNTs were effectively embedded within the cement matrix rather than remaining as unbound clusters. This interfacial bonding contributes to improved matrix densification, enhanced load transfer, and reduced electrical resistance. Consequently, the TGA results support the FE-SEM and XRD findings, confirming that MWCNT incorporation enhances both the thermal and structural stability of hybrid cementitious composites through improved microstructural integration.

The FE-SEM images confirmed the formation of continuous CNT networks bridging adjacent hydration products, which facilitates electron transport and thereby reduces bulk electrical resistance. These interconnected CNT pathways also contribute to stress transfer across microcracks, resulting in enhanced flexural strength. The XRD results showing intensified C–S–H peaks indicate improved matrix crystallinity and bonding, while the TGA data reveal additional mass loss attributed to CNT decomposition, confirming their stable incorporation within the hydrated matrix. Collectively, these microstructural features explain the concurrent improvement in both electrical conductivity and mechanical performance observed in the hybrid composites.

4. Conclusions

In this study, cementitious composites incorporating MWCNT and steel fibers were fabricated to achieve superior flexural strength. Heating performance tests, electrical resistance tests, flexural strength tests, and microstructural analyses were conducted. Based on the experimental findings, the following conclusions were drawn:

• The heating performance of steel fiber–carbon nano cementitious composites increased with higher MWCNT concentration. Specimen MW1.0-SF0.0-7D exhibited the best heating performance among all parameters, with a 10 min temperature rise of 65.0 °C and a maximum heating value of 80.1 °C. MWCNTs formed CNT networks between the hydration products. As the MWCNT concentration increased, the number of CNT networks also increased, resulting in improved 10 min temperature rise and maximum heating values.
• The heating performance decreased with longer curing duration. The heating value of 7-day cured specimens was approximately 50% higher than that of 28-day cured specimens. The residual moisture in the early curing stage increased the current flow through specimens, thereby enhancing the heating performance. As curing progressed, hydration consumed the moisture, diminishing the effect of residual water and reducing the heating performance.
• The maximum heating value and 10 min temperature rise decreased with the incorporation of steel fibers. Among fiber-reinforced specimens, MW1.0-SF2.0-7D showed the best heating performance, with values of 21.8 °C and 50.8 °C for the 10 min temperature rise and maximum heating value, respectively. These values were approximately 70% and 37% lower than those of MW1.0-SF0.0-7D with identical MWCNT concentration. The reduction in the heating performance was attributable to the increase in the steel fiber resistance at higher temperatures. However, compared to the values for specimens without MWCNT, the 10 min temperature rise and maximum heating value increased by approximately 21.7 °C and 50.5 °C, respectively, demonstrating superior heating performance relative to conventional construction materials.
• Electrical resistance decreased with higher MWCNT concentration. Increased incorporation of highly conductive MWCNTs promoted the formation of extensive CNT networks, thereby reducing the resistance. Electrical resistance increased with steel fiber incorporation. Steel fibers influenced the workability of fresh mortar; reduced workability hindered the dispersion of MWCNTs, leading to increased resistance.
• Flexural strength improved with MWCNT incorporation. MWCNTs, possessing high tensile strength, formed CNT networks within specimens, which enhanced their flexural strength. Steel fibers exerted a pronounced reinforcing effect; the flexural strength of fiber-reinforced specimens exceeded that of specimens without steel fiber by more than threefold. In addition to enhancing the flexural strength, steel fibers mitigated microcracking, thereby improving the durability of steel fiber–carbon nano cementitious composites.
• FE-SEM imaging confirmed that MWCNTs were dispersed within steel fiber–carbon nano cementitious composites and formed CNT networks. These CNT networks improved the heating performance of the specimens and reduced their electrical resistance. XRD and TGA analyses verified that hydration products identical to those observed in ordinary mortar were generated in steel fiber–carbon nano cementitious composites. MWCNT incorporation did not alter the cement hydration reactions.
• Steel fiber–carbon nano cementitious composites demonstrated adequate heating performance for black ice prevention. Steel fiber incorporation contributed to the adjustment of thermal gradients and enhancement of durability. In addition, the marked improvement in flexural strength supports the application of these composites as self-heating construction materials with both functional and structural advantages.
• Laboratory-scale experiments verified the enhanced heating and mechanical performance of steel fiber–MWCNT cementitious composites. The results provide a basis for large-scale validation and long-term durability evaluation under real environmental conditions. Further research will focus on optimizing energy efficiency and assessing life-cycle sustainability. Experimental data and procedures are available upon reasonable request to ensure reproducibility.