Researchers at ESPOL Polytechnic University in Guayaquil, working with colleagues in Spain, evaluated eight concrete mixtures labeled M0 through M7, including a plain control mix, single-fiber systems, and hybrid combinations of Dramix steel fibers and SikaForce-48 polypropylene macrofibers. Their goal was to identify a combination that maximized energy absorption and post-crack behavior while remaining workable and affordable for real civil engineering projects such as slabs, industrial floors, and pavements. The work, published in Frontiers in Built Environment, addresses a gap in locally generated performance data that has limited the adoption of fiber-reinforced concrete in the region.

The logic of hybridization is rooted in scale. Steel fibers, with tensile strengths near 2,300 megapascals and an elastic modulus around 207 gigapascals, act as stiff bridges that carry load across wide, established cracks and sustain residual strength after failure. Polypropylene macrofibers, by contrast, are far lighter, with a density of roughly 0.91 to 0.92 grams per cubic centimeter against steel’s 7.85, and contribute primarily to crack control and deformation stability. Combining them means the two fiber types operate at different stages of the cracking process, producing a composite whose response is more balanced than either fiber alone can deliver.

To measure performance, the team relied on EFNARC panel tests, casting 600 by 600 by 100 millimeter panels and loading them while recording complete load-deflection curves. Two indicators dominated the analysis: E25, the energy absorbed up to 25 millimeters of deflection, and R25, the residual load the panel still carries at that deflection. In total the experimental program produced 24 panels and 72 cylinders, with compressive strength tested at 7, 14, and 28 days and workability tracked through slump measurements.

The results were dramatic. The plain control mixture, M0, absorbed just 36.09 newton-meters of energy and retained a residual load of only 0.20 kilonewtons, equivalent to a mere 1.77 percent of its peak strength of 11.29 kilonewtons. In other words, conventional concrete lost more than 98 percent of its load-bearing capacity once it cracked. A mixture reinforced solely with steel fiber, M1, reached E25 of 506.03 newton-meters and retained 19.00 kilonewtons, while the all-polypropylene mix, M5, achieved 266.09 newton-meters and 7.80 kilonewtons.

The hybrids outperformed everything. Mixture M7, dosed with 11.25 kilograms of steel fiber and 3.75 kilograms of polypropylene macrofiber per cubic meter, recorded E25 of 653.63 newton-meters, R25 of 21.70 kilonewtons, a peak load of 33.89 kilonewtons, a 28-day compressive strength of 257.98 kilograms per square centimeter, and a slump of 18.33 centimeters. Compared with the plain control, that represents roughly 18 times more absorbed energy and more than 100 times greater residual load. Its sister hybrid M6 posted a marginally higher energy figure of 656.86 newton-meters, but with lower slump, lower residual load, and a material cost 13.9 percent higher.

Statistical analysis reinforced the findings. A one-way ANOVA on absorbed energy yielded an F-value of 41.41 with a p-value of 1.08 multiplied by ten to the minus eighth power, confirming significant differences among the mixtures, and Tukey’s test showed all fiber-reinforced mixes significantly outperformed the control while M6 and M7 were statistically indistinguishable from each other. Compressive strength differences were significant at 7 and 28 days, with M7 leading, although the authors note that some mixtures showed higher variability, likely tied to local differences in fiber distribution and orientation, factors the study did not independently quantify.

The economic dimension proved decisive in selecting a winner. At 179.20 US dollars per cubic meter, M7 cost nearly 29 dollars less than M6 while sacrificing less than 1 percent of energy absorption, and it delivered 12.1 percent higher residual load, 25.3 percent greater compressive strength, and better workability. The lesson is counterintuitive and important: simply loading concrete with the maximum fiber content does not produce the most efficient material. Higher dosages reduce flowability, hinder homogeneous dispersion, and raise the risk of fiber clumping, eroding the very performance gains they are meant to provide.

Beyond the laboratory, the implications reach into sustainability. Cement and concrete production is associated with roughly 7 to 8 percent of global carbon dioxide emissions, so any technology that extends service life, limits cracking, and reduces repair cycles carries environmental weight. The authors frame their contribution as a replicable, performance-oriented experimental protocol combining material characterization, EFNARC fracture testing, compressive strength verification, and cost-per-cubic-meter analysis, giving engineers a defensible multi-criteria basis for specifying hybrid fiber concrete rather than relying on peak strength alone.

The team is candid about limitations: three panels per mixture constrain statistical precision, fiber orientation was not characterized through imaging or computed tomography, and the fiber-matrix interface was not directly tested through pull-out experiments. Future work should address fatigue, impact, abrasion, chloride and carbonation durability, full-scale validation, and life-cycle cost analysis. Even so, the study demonstrates that in the unglamorous world of industrial flooring, a precisely tuned cocktail of steel and plastic fibers can turn brittle, fragile concrete into a material that bends, absorbs, and keeps carrying the load long after it has cracked.

Subject of Research: Mechanical performance and energy absorption of steel-polypropylene hybrid fiber-reinforced concrete

Article Title: Mechanical performance and energy absorption capacity of concrete reinforced with hybrid fiber systems

Article References: Briones-Bitar, J., Pinto-Ponce, B., Pacaji-Ulloa, M., Vera-Navarrete, E., Santos-Baquerizo, E., Berrezueta, E., Carrión-Mero, P., & Morante-Carballo, F. (2026). Mechanical performance and energy absorption capacity of concrete reinforced with hybrid fiber systems. Frontiers in Built Environment, 12, Article 1965114. https://doi.org/10.3389/fbuil.2026.1965114

Image Credits: AI Generated

DOI: 10.3389/fbuil.2026.1965114

Keywords: fiber-reinforced concrete, hybrid fibers, steel fibers, polypropylene macrofibers, EFNARC panel test, energy absorption, toughness, compressive strength, workability, cost analysis, pavements, construction materials

The material at the heart of the research is ultra-high-performance alkali-activated concrete, or UHP-AAC. Conventional ultra-high-performance concrete achieves compressive strengths of 120 megapascals or more by packing cement, fine particles, and fibers into an extraordinarily dense matrix, but it typically requires 900 to 1100 kilograms of cement per cubic meter. UHP-AAC replaces that cement entirely with industrial by-products, chiefly ground granulated blast furnace slag and silica fume, which are activated by an alkaline compound rather than by clinker hydration. Previous work has shown that this substitution can cut carbon dioxide emissions by up to 70 percent and production costs by up to 64 percent while still reaching strengths above 130 megapascals.

Designing such mixtures, however, is notoriously difficult. The compressive strength of UHP-AAC depends on a web of nonlinear, interacting variables: the balance between slag and silica fume, the dose of solid alkaline activator, the amount of recycled steel fiber, the water content, the curing regime, and the age of the specimen. Traditional mix design relies on exhaustive laboratory trial batches, each consuming time, labor, and money. The research team, led by K. Behfarnia and A. Ghasemi Marnani at Isfahan University of Technology in Iran, set out to replace much of that experimental burden with data-driven prediction.

What distinguishes this study from earlier machine learning efforts in concrete research is the quality and scale of its dataset. Rather than pooling results from many laboratories, where differences in materials, curing, and testing procedures can muddy the signal, the authors generated 2048 compressive strength records from 256 unique mixture designs in their own laboratory. Each mixture was tested under two curing conditions, wrapped and ambient, at four ages: 7, 14, 28, and 56 days. Seven input features described every record, including curing method, slag content, silica fume content, sodium metasilicate pentahydrate as the solid activator, recycled steel fiber content, water content, and curing age. Measured strengths ranged from about 27.6 to 137.5 megapascals, a wide and well-distributed span that is ideal for training robust models.

The team benchmarked six algorithms: linear regression, decision tree, random forest, k-nearest neighbors, gradient boosting, and extreme gradient boosting. Each model was evaluated with unusual statistical rigor. Because multiple measurements came from the same mixture, the researchers grouped all data from an identical mix design together during train-test splitting, preventing information leakage between related records. They repeated the entire modeling pipeline ten times with different group-based splits, used five-fold group cross-validation during training, and tuned hyperparameters through grid and randomized searches. This design ensured that models were always judged on genuinely unseen mixtures, the fairest possible test of generalization.

The results were striking. Gradient boosting emerged as the clear winner, achieving a coefficient of determination of 99.93 percent on the unseen test set with a mean absolute error of just 0.406 megapascals, meaning its typical prediction missed the measured strength by less than half a megapascal. Extreme gradient boosting followed at 99.83 percent, then random forest at 98.6 percent, decision tree at 97.1 percent, k-nearest neighbors at 95.9 percent, and linear regression at 90.2 percent. Every prediction from the top three models fell within 20 percent of the observed value. Residual analysis across low, medium, and high strength ranges confirmed that the boosting models remained accurate whether the concrete was expected to reach 40 or 130 megapascals.

The gap between simple and ensemble methods carries a technical lesson. Linear regression assumes additive, linear effects, but the study’s own correlation analysis showed that all relationships between the input variables and strength were nonlinear, with correlation coefficients below 0.7. Boosting algorithms, which build decision trees sequentially with each new tree correcting the errors of its predecessors, excel precisely at capturing such interactions. Random forest, which averages many independently grown trees, reduces variance but lacks the sequential error correction that gives boosting its edge. Single decision trees and k-nearest neighbors, meanwhile, fit training data almost perfectly but generalize less well to new mixtures.

Beyond raw accuracy, the researchers applied SHAP analysis, a technique borrowed from game theory that quantifies each feature’s contribution to every individual prediction. The results were physically coherent and, in places, surprising. Curing method proved the single most influential variable: wrapped curing, which seals specimens in plastic sheeting to preserve moisture, consistently outperformed ambient curing by sustaining the dissolution and gel-forming reactions that build strength. Steel fiber content ranked second, reflecting the fibers’ ability to bridge cracks and restrain lateral deformation under compression. Water content and curing age followed, with higher water diluting the alkaline activator and slowing gel formation, while longer curing allowed continued microstructural densification, especially before 28 days.

The finer details of the binder chemistry also emerged clearly from the model. Increasing slag content raised strength because slag is rich in calcium oxide, promoting the calcium-rich binding gels that give the matrix its density. Silica fume, by contrast, showed a negative effect in this design space: because it contains almost no calcium, substituting it for slag lowers the system’s calcium-to-silicon ratio and limits the formation of those critical binding phases. The optimal activator dose sat near 170 kilograms per cubic meter, and the best fiber contents clustered around 230 to 240 kilograms per cubic meter, roughly a 3 percent volume fraction.

The practical implications extend well beyond the laboratory. The authors position their gradient boosting model as a decision-support tool that engineers can use at the mix design stage to screen candidate formulations and estimate strength at different ages without extensive preliminary trials, dramatically accelerating the development of new cement-free mixtures. They caution, appropriately, that predictions are only reliable within the range of materials and proportions tested, and that laboratory verification remains essential for critical structural applications. Still, the combination of a large, internally consistent dataset, rigorous group-based validation, and interpretable machine learning offers a template for how artificial intelligence can help decarbonize one of humanity’s heaviest industries, one optimized mixture at a time.

Subject of Research: Machine learning prediction of compressive strength in ultra-high-performance alkali-activated concrete

Article Title: A comparative machine learning study for predicting the compressive strength of ultra-high-performance alkali-activated concrete

Article References: Behfarnia, K., & Marnani, A. G. (2026). A comparative machine learning study for predicting the compressive strength of ultra-high-performance alkali-activated concrete. Case Studies in Construction Materials, 25, Article e06577. https://doi.org/10.1016/j.cscm.2026.e06577

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06577

Keywords: machine learning, ultra-high-performance concrete, alkali-activated concrete, compressive strength, gradient boosting, SHAP, ground granulated blast furnace slag, silica fume, steel fibers, sustainable construction, cement replacement, mix design

The research, led by Jiuyang Li and colleagues, systematically explored a hybrid conductive system that combines four different types of steel fibers with short-cut carbon fibers. The team’s central question was deceptively simple: can a single concrete recipe simultaneously deliver high compressive strength, low electrical resistivity, and a reliable piezoresistive response, meaning a resistivity that changes predictably under load? Previous work usually optimized one property at the expense of the others, and the authors argue that only a multi-phase fiber network can break that trade-off.

To answer it, the researchers designed a rigorous three-factor, four-level orthogonal experiment, producing sixteen hybrid mixtures plus a plain concrete control. The variables were steel fiber type, ranging from shear-type and milled fibers to copper-coated and hooked-end variants, steel fiber volume fraction from 0.35 to 1.25 percent, and carbon fiber content from 0.12 to 0.48 percent. Each mixture was evaluated for cubic and axial compressive strength, volume resistivity, and sensing behavior under both monotonic and cyclic compression, with range analysis, analysis of variance, and matrix analysis applied to disentangle each factor’s contribution.

The mechanical results were unambiguous. Every hybrid mixture outperformed the plain control, which registered a cubic compressive strength of 45.06 megapascals, and the strongest performer, specimen T-11 containing copper-coated steel fibers at 0.95 percent volume with 0.12 percent carbon fibers, reached 60.17 megapascals. Statistical testing showed that steel fiber content was the dominant factor for strength, followed by fiber type, while carbon fiber content played a comparatively minor mechanical role. The optimum combination for coupled mechanical performance was copper-coated steel fibers at 0.95 percent with carbon fibers at 0.48 percent.

Microscopy helped explain why. Scanning electron micrographs revealed carbon fibers fracturing across the cement matrix rather than pulling out, evidence of strong interfacial bonding that lets their exceptional tensile modulus resist crack growth. Steel fibers, meanwhile, overlapped and interlocked into a local skeleton that both absorbs energy and establishes long-range conductive pathways. Carbon fibers filled micropores and gel pores between the larger steel elements, tightening the network. Too much carbon fiber, however, caused agglomeration that loosened the matrix and weakened the bond, a reminder that in conductive concrete, more is not always better.

The electrical findings were the most dramatic. Plain concrete exhibited a resistivity of roughly 243,859 ohm-meters, dominated by slow ionic conduction through pore water. The hybrid fiber concretes collapsed that figure by as much as 99.93 percent, with the best specimens falling below 200 ohm-meters. Here the hierarchy of influence reversed: carbon fiber content mattered most, followed by steel fiber content, then fiber type. The authors identified a percolation threshold near 0.95 percent steel fiber, beyond which adding more metal yielded almost no additional conductivity because the interconnected fiber network had already become the primary conduction route.

That percolated network is precisely what makes the material sense pressure. Under compression, pores close and fiber spacing shrinks, lowering both contact resistance between touching fibers and tunneling resistance across nanoscale gaps between carbon fibers. The result is a measurable drop in resistivity that tracks the applied load. Under monotonic loading to 400 kilonewtons, all sixteen mixtures showed a clear piezoresistive response, with fitted curves achieving coefficients of determination above 0.9. The copper-coated steel fiber hybrids were the most sensitive of all, with specimen T-10 recording a fractional change in resistivity of minus 78.02 percent at peak load.

Cyclic loading tests on the four copper-coated hybrids revealed both the promise and the remaining challenge. The resistivity of the materials oscillated in sync with each of six loading cycles, falling as load rose and recovering as it fell, exactly the behavior a built-in stress gauge would need. Yet the baseline drifted. After six cycles, the fractional change in resistivity shifted irreversibly, from as little as minus 8.90 percent in the most stable group to minus 21.06 percent in the most drift-prone, reflecting internal compaction and microcracking that permanently rearranged the conductive network. The T-11 mixture, with 0.95 percent copper-coated fibers and just 0.12 percent carbon fibers, showed the smallest cumulative drift, averaging minus 1.48 percent per cycle.

The authors are candid about the caveats. Copper-coated steel fibers raise questions of electrochemical corrosion and long-term durability that this study did not address, and they recommend future work introducing the coating as an explicit variable, alongside accelerated corrosion, chloride exposure, and long-term electrical stability testing. They also note that because carbon fiber content governs conductivity, engineers prioritizing sensing over strength could reduce the expensive steel fiber dosage, cutting both cost and environmental footprint while preserving the self-monitoring function.

Even with those qualifications, the study marks a meaningful step toward infrastructure that reports its own condition. A concrete column laced with a percolated steel and carbon fiber network could, in principle, tell engineers continuously how much load it carries and when internal cracks begin to form, without a single attached sensor. As monitoring demands grow across bridges, tunnels, and towers, the idea of buildings wired from within, by the very material that holds them up, is moving from laboratory curiosity toward engineering reality.

Subject of Research: Multi-phase conductive fiber reinforced concrete for mechanical performance and piezoresistive structural health monitoring

Article Title: Study on the mechanical properties, electrical conductivity, and piezoresistive behavior of multi-phase conductive fiber reinforced concrete

Article References: Li, J., Luo, C., Wang, B., Luo, J., Wang, Z., & Shao, M. (2026). Study on the mechanical properties, electrical conductivity, and piezoresistive behavior of multi-phase conductive fiber reinforced concrete. Case Studies in Construction Materials, 25, Article e06580. https://doi.org/10.1016/j.cscm.2026.e06580

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06580

Keywords: concrete, steel fibers, carbon fibers, piezoresistivity, electrical conductivity, structural health monitoring, self-sensing materials, orthogonal experiment, compressive strength, percolation threshold, smart materials, infrastructure

The study, published in Case Studies in Construction Materials, tackles a bottleneck that has constrained protective design for years. Existing C120-grade high-strength concrete typically exhibits densities between 2,400 and 2,500 kilograms per cubic meter, which inflates the self-weight of protective structures, raises construction costs and limits use in weight-sensitive applications. Conventional C80 concrete is lighter, but its strength and anti-penetration performance fall short of the demands of high-end protective engineering. The researchers set out to achieve what they describe as a synergistic optimization of protective efficiency and structural economy: ultra-high compressive strength to withstand high-velocity projectile impact, combined with a substantially reduced density.

The key to the new material lies in its mix design, which was guided by particle close-packing theory. The team systematically analyzed how varying amounts of cement, silica fume and fly ash microspheres affect the packing density of particles in the cementitious system, then tuned the aggregate ratio, sand-binder ratio and fiber content to optimize apparent density, mechanical properties and workability. The resulting C120 lightweight high-performance concrete contains 682.5 kilograms of cement, 210 kilograms of fly ash and 157 kilograms of silica fume per cubic meter, achieving a 28-day compressive strength of 121.3 megapascals and a flexural strength of 17.2 megapascals at a density of just 2,050 kilograms per cubic meter.

Two ingredient choices stand out. First, all natural aggregates were replaced with ceramsite sand, a lightweight ceramic aggregate with particle sizes between 0.5 and 5 millimeters, a bulk density of 1,450 kilograms per cubic meter and a cylinder compressive strength of 8.5 megapascals. Its rough, porous surface forms a strong bond with the cementitious matrix, while its moderate particle strength preserves load-bearing capacity at a much lower density. Second, steel fibers were added at a volume fraction of about 2.0 percent. Each fiber is 13 millimeters long, 200 micrometers in diameter and has a tensile strength of 2,000 megapascals, providing the tensile reinforcement and impact toughness needed to restrain crack propagation during penetration. For comparison, the conventional C80 mix used haydite aggregate with natural sand and basalt fibers, reaching 82.6 megapascals of compressive strength at 2,450 kilograms per cubic meter.

To test the material under realistic ballistic conditions, the team cast cylindrical concrete targets measuring 900 millimeters in diameter and 600 millimeters thick, each enclosed in a 5-millimeter steel plate that provided lateral restraint while minimizing boundary effects. The dimensions were chosen so that the radial size exceeded 30 projectile diameters and the thickness exceeded 1.5 times the expected maximum penetration depth of 360 millimeters, ensuring that neither lateral boundaries nor the rear free surface would distort the results. After 28 days of standard curing at 20 degrees Celsius and at least 95 percent relative humidity, the targets were struck with pointed ogive-nose penetrators made of 35CrMnSiA steel, each 120 millimeters long, 24 millimeters in diameter and weighing 352 grams, launched from a 30-millimeter gun at approximately 700 meters per second at normal incidence.

The results were striking. Against the conventional C80 target, struck at 699 meters per second, the projectile penetrated 0.251 meters and opened a large frontal crater with a maximum diameter of 0.421 meters and a minimum of 0.348 meters, a nearly circular shape ratio of 1.21 characteristic of dispersed brittle failure. High-speed photography at 100,000 frames per second captured the concrete at the impact point being pulverized, radial cracks spreading outward and large chunks shearing away as the projectile head entered the material. The C80 matrix has low compactness and a weak transition zone between aggregate and paste, so stress waves rapidly induced interfacial cracking that extended irregularly, dissipating energy through large-scale crushing and spalling while severely damaging the specimen’s integrity.

The lightweight C120 target, struck even faster at 726 meters per second, performed dramatically better. Penetration depth was only 0.158 meters, and the frontal crater shrank to a maximum diameter of 0.263 meters and a minimum of 0.157 meters, a shape ratio of 1.68 indicating localized, concentrated compressive failure rather than widespread destruction. The steel fibers raised tensile strength and impact toughness, increasing the energy absorbed during penetration, while the tightly bound lightweight aggregate interface suppressed crack initiation and radial expansion. Impact energy was dissipated through plastic compression within a small contact zone, leaving the overall structure intact. Notably, recovered projectiles showed more extensive surface abrasion on those that had penetrated the C120 target than on those that hit the C80, evidence that the lighter material actually imposed greater axial and sidewall frictional resistance.

A cross-literature comparison reinforced the finding. The team benchmarked its results against a published test of C120 ultra-high performance concrete with a density of 2,450 kilograms per cubic meter and a cube compressive strength of 117.9 megapascals, impacted at 728 meters per second by a smaller 15-millimeter, 101.8-gram projectile. Because both projectiles remained essentially rigid, the impact energies and penetration resistances could be compared quantitatively. The new lightweight concrete faced roughly 3.44 times the impact kinetic energy and a 34.3 percent higher sectional energy density, yet its penetration depth was 10.7 percent shallower. Its average penetration resistance reached 1.30 gigapascals, about 51.2 percent higher than the 0.86 gigapascals of the denser UHPC, and the strength-normalized resistance index rose from 7.3 to 10.7, all while the target density was 16.3 percent lower.

The data also revealed an unexpected mathematical pattern. The researchers defined a penetration depth-density coupling coefficient K, the ratio of penetration reduction to density reduction, and found K equals 2.54: every 1 percent decrease in density corresponded to a roughly 2.54 percent decrease in penetration depth. This superlinear inverse coupling suggests the material offsets the loss of inertial resistance from being lighter through mix optimization and improved interface performance, shifting anti-penetration behavior from density-dominated inertia toward domination by intrinsic material properties. Marginal-efficiency analysis told a similar story: moving from C80 to the new C120 raised compressive strength by 50 percent, but basic anti-penetration efficiency, the reciprocal of penetration depth, rose 58.86 percent, and density-normalized intrinsic efficiency rose 85.96 percent, gains that far exceed what strength increase alone would predict and that break the conventional trend of diminishing returns in traditional high-strength concrete.

Finally, the team built a finite element model using the K&C concrete constitutive model with a quarter-symmetry mesh of hexahedral elements, and validated it against the experiments. The simulations reproduced the full penetration process, including radial cracks from constrained lateral expansion, circumferential cracks from unloading tensile waves after the projectile’s compressive effect vanished, the funnel-shaped spallation crater, and the steady-state cylindrical tunnel phase, with projectile deceleration curves and instantaneous penetration depths matching the measured data closely. The authors note that their tests covered a single projectile size at roughly 700 meters per second under normal incidence, so velocity dependence, scale effects and oblique or repeated impacts remain for future study. Even so, the validated model and the material itself point toward practical use as external protective layers for nuclear power plants, substations and oil storage facilities, where a shield that is both lighter and harder to penetrate could reshape how critical infrastructure is defended.

Subject of Research: Anti-penetration performance of a new lightweight high-performance C120 concrete

Article Title: Research on the anti-penetration performance of a new type of lightweight high-performance concrete

Article References: Liu, J., Zhang, C., Wang, G., Cao, S., Li, L., Dang, A., & Liu, X. (2026). Research on the anti-penetration performance of a new type of lightweight high-performance concrete. Case Studies in Construction Materials, 25, Article e06592. https://doi.org/10.1016/j.cscm.2026.e06592

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06592

Keywords: lightweight concrete, high-performance concrete, anti-penetration, projectile impact, protective engineering, ceramsite aggregate, steel fibers, finite element simulation, penetration depth, compressive strength, critical infrastructure, materials science

The researchers built their investigation around cellulose nanofibers, or CNF, fibrils just 4 to 20 nanometers in diameter and 1 to 3 micrometers long, packed with hydroxyl and carboxyl surface groups. These renewable, high-aspect-ratio fibrils have attracted growing interest in cement science because they can nucleate hydration products, bridge nanoscale defects and, crucially, hold onto mixing water like tiny molecular sponges. But CNF comes with a notorious drawback: its enormous surface area and hydrogen bonding make the fibrils clump together, raising the water demand of a fresh mix and threatening the flowability that UHPC placement depends on. The team therefore faced a classic trade-off question. The issue was not whether fibers at different length scales can help control cracking, but how a nanocellulose-modified matrix interacts with micro-scale aramid fibers and macro-scale steel fibers when workability, strength, shrinkage and durability all impose competing demands.

The experimental system was built on a meticulously optimized UHPC matrix containing 1092 kilograms of cement per cubic meter, supplemented with silica fume and metakaolin in a ternary binder blend, plus high-purity quartz sand, at a water-to-binder ratio of just 0.16. All mixtures included 2.0 percent by volume of steel fibers, 12 to 13 millimeters long and about 0.20 millimeters in diameter, as the macro-scale reinforcement. Variable amounts of CNF, from 0.05 to 0.25 percent of binder mass, were dispersed using a clever metakaolin-assisted pre-dispersion procedure: the nanofibers and metakaolin were first co-dispersed in the mixing water to physically separate the fibrils before the suspension entered the highly ionic cementitious environment, limiting direct CNF-to-CNF contact and the agglomeration that would otherwise turn the nanofibers into defects. Aramid fibers, roughly 12 micrometers in diameter and about one millimeter long, were added at the final mixing stage in dosages of 0.2 to 0.6 percent by volume.

The results reveal a sharp optimum. As CNF dosage rose, flowability declined monotonically, from 232 millimeters on the flow table for the reference mix down to 205 millimeters at 0.25 percent CNF, an 11.6 percent loss. At the mechanically selected dosage of 0.15 percent, however, the workability penalty was modest, only 4.7 percent, while the mechanical rewards were substantial. The 28-day compressive strength reached 149.4 megapascals, an 11.4 percent gain over the control, and the mean flexural strength climbed from 30.9 to 51.6 megapascals, a striking 67.0 percent relative difference. The authors are careful to frame this correctly: because the control already contained the same steel fiber volume, the flexural jump reflects a system-level response, in which a tougher matrix cracks less readily and activates the existing steel fibers more efficiently, rather than a direct load-carrying contribution from the nanofibers themselves.

The mechanism behind these gains is a story of water management and nanoscale crack arrest. Well-dispersed CNF acts as nanoscale bridges that restrain the initiation and propagation of microcracks, while its surface functional groups provide nucleation sites for hydration products, densifying the microstructure. Most importantly for UHPC, the hydrophilic fibrils absorb and temporarily retain part of the scarce mixing water, releasing it gradually under the internal humidity gradient as hydration proceeds. This internal curing effect directly counteracts self-desiccation, the root cause of autogenous shrinkage. Beyond 0.15 percent, the benefits reversed: agglomerates formed, free water for lubrication and hydration dwindled, compaction quality suffered, and both strength and flowability dropped. The reinforcing effect of CNF, the study makes clear, depends entirely on dispersion state and dosage.

The shrinkage data are arguably the most consequential for practice. After seven days of sealed monitoring with laser displacement sensors, the autogenous shrinkage of the 0.15 percent CNF mixture fell to approximately 1190 microstrain, down from roughly 1600 microstrain for the control, a reduction of about 25.6 percent. At 0.20 percent CNF the reduction was smaller, around 11.3 percent, confirming that the optimum for volume stability coincided with the optimum for strength. In a separate drying-shrinkage series, the micro-scale aramid fibers stole the show: a mix containing 0.4 percent aramid fiber cut 28-day drying shrinkage from about 134 to 76 microstrain per meter, a 43.3 percent reduction. Because aramid fibers have far smaller diameters and much higher number densities than steel fibers at the same volume fraction, they bridge the distributed microcracks that form during moisture loss far more effectively, providing three-dimensional internal restraint precisely where drying damage begins.

The aramid fiber series exposed a striking loading-mode trade-off. At fixed 0.15 percent CNF and 2.0 percent steel fiber, raising aramid content from 0.4 to 0.6 percent by volume increased 28-day compressive strength from 154.7 to 159.4 megapascals, an 18.9 percent gain over the original control, but simultaneously reduced flexural strength from 54.0 to 48.4 megapascals, a 10.4 percent penalty. The explanation lies in how cracks travel. Compression damages a large distributed volume, where a denser population of fine fibers restrains lateral microcrack growth. Flexure, by contrast, is governed by a single localized critical section, where fiber crowding at high dosages can disturb orientation, reduce the effective number of well-bonded fibers crossing the crack plane and introduce local defects. The team therefore recommends 0.4 percent aramid fiber as the balanced choice, reserving 0.6 percent for cases where compressive strength alone governs.

Durability testing added a further dimension. Specimens cycled between immersion in 5 percent sodium sulfate solution and oven drying retained mass and compressive strength more effectively when CNF and the hybrid fibers were present. Sulfate attack exploits the same weaknesses that shrinkage does, namely connected capillary pores and microcracks that serve as preferential transport pathways, so a matrix densified by nanocellulose and stabilized against cracking by aramid and steel fibers resists ingress more effectively. Microstructural evidence supported the picture: X-ray diffraction showed the same principal crystalline phases, portlandite, calcite, residual alite and quartz, in control and nanocellulose mixtures, indicating no new crystalline products, while scanning electron microscopy revealed a markedly more compact matrix with fewer microcracks and microvoids in the 0.15 percent CNF specimen compared with the loosely packed, crack-riddled control region.

The study, published under the DOI 10.1016/j.cscm.2026.e06532, arrives with appropriately measured language: because the experimental program did not include every possible single- and dual-fiber control, the authors describe their finding as coordinated multi-scale enhancement rather than claiming rigorous proof of synergy. Even so, the demonstrated composition, 0.15 percent CNF plus 0.4 percent aramid fiber plus 2.0 percent steel fiber, delivered the highest measured flexural strength of 54.0 megapascals, an 11.4 percent compressive gain, roughly a quarter less autogenous shrinkage, over 40 percent less drying shrinkage in the aramid series, and improved sulfate resistance, all at a workability cost of under 5 percent. As the construction industry searches for materials that combine extreme performance with lower cracking risk and longer service life in bridges, marine works and protective structures, the idea that a dash of renewable plant nanofibers, working in concert with aramid and steel fibers, can simultaneously strengthen, stabilize and toughen the strongest concrete we know how to make is a genuinely compelling advance.

Subject of Research: Multi-scale hybrid fiber reinforcement of ultra-high performance concrete using cellulose nanofibers, aramid fibers and steel fibers

Article Title: Coordinated multi-scale enhancement of UHPC by a nanocellulose-based hybrid fiber system: Mechanical performance, shrinkage mitigation and microstructural evidence

Article References: Huang, J., Wang, S., Jiang, Y., Shi, R., & Yang, Y. (2026). Coordinated multi-scale enhancement of UHPC by a nanocellulose-based hybrid fiber system: Mechanical performance, shrinkage mitigation and microstructural evidence. Case Studies in Construction Materials, 25, Article e06532. https://doi.org/10.1016/j.cscm.2026.e06532

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06532

Keywords: ultra-high performance concrete, cellulose nanofibers, aramid fibers, steel fibers, autogenous shrinkage, drying shrinkage, flexural strength, compressive strength, sulfate resistance, internal curing, microstructure, cementitious materials

The research team, led by Bowen Zhang, Zhi-Qi He, Jiatong Chen, and Wenjie Li, conducted eighteen local compression tests on UHPC specimens designed to represent prestressed anchorage zones. The specimens varied systematically across four key parameters: steel-fiber volume fractions of 0.5, 1.0, and 2.0 percent; bearing area ratios of 3, 5, and 7; three forms of indirect reinforcement, namely spiral reinforcement, reinforcing mesh, and none at all; and three anchorage plate configurations, including flat plates, embedded plates, and no plate. The main group of specimens was arranged in an L9 orthogonal experimental matrix, allowing the relative influence of the primary variables to be ranked without running a full factorial program, while targeted comparison groups isolated the effects of plate type and reinforcement ratio.

The material itself was a dense cementitious composite of ordinary Portland cement, silica fume, slag powder, fine quartz sand, water, superplasticizer, and straight copper-coated steel fibers 14 millimeters long and 0.20 millimeters in diameter, with an aspect ratio of 70. Measured axial compressive strengths ranged from 88.0 to 108.7 megapascals, with converted cube strengths between 122.2 and 151.0 megapascals. Direct-tension tests revealed mean tensile strengths of 3.89, 6.34, and 8.15 megapascals for the three fiber contents, a clear demonstration of how dramatically steel fibers transform the tensile character of an otherwise brittle matrix. All specimens were tested in a 500-tonne hydraulic compression machine under carefully controlled stepwise loading, with local plate indentation separated from overall specimen compression through a dual displacement measurement scheme.

Every specimen failed by splitting rather than compressive crushing, and the failure process unfolded in three recognizable phases: crack initiation, stable crack propagation, and rapid splitting collapse. Initial vertical hairline cracks appeared at loads ranging from roughly 30 to 90 percent of the ultimate capacity, most often in the upper or upper-middle regions of the specimens. Close-up examination of the fracture surfaces revealed that most steel fibers were pulled out of the matrix rather than ruptured, confirming that fiber bridging and pull-out resistance were the dominant mechanisms transferring tensile stress across cracks after the matrix had split. Specimens with lower fiber contents developed fewer but wider cracks, typically one or two dominant vertical splits, while higher fiber contents produced finer, more distributed cracking patterns.

Strain gauges on the indirect reinforcement told an equally revealing story about when different load-resisting mechanisms activate. At low load levels, reinforcement strains remained small, indicating that the UHPC matrix and its embedded fibers carried the transverse tensile demand on their own. Only as loads approached roughly 60 to 80 percent of ultimate capacity did reinforcement strains accelerate sharply, coinciding with visible cracking and lateral expansion. Specimens with 2.0 percent fiber volume kept reinforcement strains well below the monitoring reference even near failure, suggesting that abundant fibers reduce the confinement burden that transverse steel must carry. In other words, fibers and reinforcement work in sequence rather than in parallel: fibers manage microcracks and early post-cracking stress transfer, while spirals and meshes restrain macrocrack opening and lateral dilation later in the loading history.

Range analysis of the orthogonal test matrix produced a striking hierarchy of influence. The bearing area ratio, which controls the intensity of stress concentration and the geometry of load diffusion beneath the plate, was by far the dominant factor, with a range value of 65.8 megapascals in mean bearing stress. Fiber volume fraction ranked second at 49.0 megapascals, while the form of indirect reinforcement trailed at just 15.9 megapascals. The anchorage plate configuration mattered in nuanced ways: flat plates increased ultimate capacity by 15.4 percent in one matched comparison but simultaneously reduced initial stiffness and cracking load, whereas embedded plates consistently improved stiffness and cracking resistance, raising cracking loads by as much as 71.4 percent, without reliably increasing ultimate strength. Changing the reinforcement ratio between 1.5 and 3.0 percent produced no monotonic trend in either cracking or ultimate load, underscoring that its effect is conditional on the surrounding fiber and plate-transfer state.

To place these findings in context, the team compiled a database combining their eighteen tests with published UHPC and reactive powder concrete data, yielding 51 unreinforced and 34 reinforced records. When existing design provisions were evaluated against this database, their limitations became evident. AASHTO LRFD 2017 came closest on average, with a mean calculated-to-test ratio of 0.947 for unreinforced specimens, but the French UHPC standard NF P 18-710 tended to overestimate capacity with a mean ratio of 1.154, and the Chinese bridge code JTG 3362-2018 overestimated reinforced specimens at 1.092. The Chinese steel-fiber design standard JGJ/T 465-2019 performed worst, overestimating capacity by roughly 30 percent with standard deviations above 0.20, indicating that its fiber and confinement terms fail to capture the distinctive splitting resistance and fiber-bridging behavior of UHPC.

In response, the researchers developed two prediction models grounded in the observed mechanics. The first is a modified confinement-based model that expresses bearing capacity as the sum of a fiber-enhanced UHPC contribution and an additional confinement term from indirect reinforcement. A steel-fiber characteristic parameter captures the reinforcing efficiency of the fibers, while a regression-derived coefficient reveals that the confinement contribution of transverse steel actually decreases as fiber content rises, vanishing entirely beyond a fiber characteristic parameter of 1.38. The model achieved calculated-to-test ratios of 0.993 with a standard deviation of 0.081 for unreinforced specimens and 1.013 with a standard deviation of 0.105 for reinforced ones, outperforming existing literature models in both accuracy and scatter.

The second model takes a more theoretical route, treating local bearing failure as a plasticity problem. Using a modified Mohr-Coulomb criterion with a nonzero tensile cutoff to represent fiber-bridged post-cracking strength, the team idealized the failure zone as a wedge mechanism combining a shear-compression plastic flow zone with a tensile splitting zone. An upper-bound virtual work solution, calibrated with a reduction factor of 0.172 against 25 unreinforced database records, reproduced measured capacities with a mean ratio of 1.00 on the calibration set. The authors are candid that this wedge model, with its rigid-plastic idealization and simplified geometry, is best suited for mechanism interpretation and analytical assessment of unreinforced zones rather than direct design use, while the confinement-based model serves as the practical design tool within the calibrated parameter ranges.

The implications reach well beyond the laboratory. Prestressed bridge girders, stay-cable anchorages, and other heavily loaded UHPC components can now be designed with equations that explicitly recognize how steel fibers and transverse reinforcement share the job of resisting splitting, and how plate geometry reshapes the load-transfer path. The study also flags its own boundaries: the models are calibrated only within the tested ranges of fiber content, bearing area ratio, and reinforcement ratio, and further validation is needed for full-scale members, different anchorage systems, larger specimen sizes, and alternative fiber types. Still, by demonstrating that anchorage-zone failure is a progressive compression-splitting-confinement process, the work gives engineers a mechanically honest foundation for pushing one of construction’s strongest materials closer to its true limits, potentially enabling lighter, more slender, and more durable prestressed structures at a time when infrastructure demands are growing faster than conventional materials can comfortably answer.

Subject of Research: Experimental and analytical evaluation of the local bearing capacity of ultra-high-performance concrete prestressed anchorage zones

Article Title: Experimental study and analytical evaluation of local bearing capacity of UHPC anchorage zones

Article References: Zhang, B., He, Z.-Q., Chen, J., & Li, W. (2026). Experimental study and analytical evaluation of local bearing capacity of UHPC anchorage zones. Case Studies in Construction Materials, 25, Article e06508. https://doi.org/10.1016/j.cscm.2026.e06508

Image Credits: AI Generated

DOI: 10.1016/j.cscm.2026.e06508

Keywords: ultra-high-performance concrete, UHPC, anchorage zones, local bearing capacity, steel fibers, prestressed concrete, splitting failure, confinement, plasticity wedge model, bridge engineering, bearing area ratio, indirect reinforcement