Bi-layer reinforced concrete beams represent a structural application of the functionally graded concrete concept, where materials with different properties are distributed along the section height to improve mechanical efficiency and reduce environmental impact. This study investigates the mechanical behaviour and global warming potential (GWP) of reinforced bi-layer concrete beams compared with conventional monolithic beams. An experimental program was conducted on reinforced concrete beams tested under four-point bending. Two bi-layer configurations incorporating steel fibre-reinforced concrete in the tensile zone were examined. The structural response was analysed using digital image correlation, acoustic emission, distributed fibre optic sensing and fiber Bragg gratings to monitor crack development, deformation fields and steel strain evolution. In parallel, a cradle-to-site environmental assessment was performed, and structural optimisation strategies based on variations of the geometric ratio h/b and reinforcement ratio were analysed using Pareto optimisation. Bi-layer beams exhibit an increase in flexural capacity of approximately 16–18

Predicting the flow behavior of self-consolidating concrete (SCC) remains a persistent challenge due to the highly nonlinear interactions among material constituents and the lack of standardized experimental data. In this study, we introduce a hybrid, explainable-AI framework that integrates machine learning interpretability techniques (SHAP, LIME) with genetic programming-based model optimization to accurately predict SCC flow properties using a literature-derived database of 1,577 mixtures. The optimized models achieved coefficients of determination of R2 = 0.72 for J-ring flow diameter, 0.80 for V-funnel flow time, 0.56 for L-box ratio, and 0.79 for segregation resistance. Model interpretability analysis revealed that cement composition, superplasticizer dosage, and coarse aggregate content are the dominant factors governing SCC flow behavior. This work demonstrates that coupling explainable AI with genetic programming provides both quantitative accuracy and physical interpretability, offering a reproducible pathway toward data-driven optimization of sustainable and low-carbon concrete formulations.

In this study, we applied machine learning techniques to classify concrete into consistency classes, using real-world data provided by an industrial partner in France. The dataset includes over 480 different mixtures and more than 1,515 concrete samples, most of which incorporate supplementary cementitious materials (SCMs) such as slag and fillers. There are twelve input features and one output representing each of the five consistency classes. The machine learning models demonstrated strong performance, with the Light Gradient Boosting Machine emerging as the best performer, with an F1-score exceeding 88

This study develops an integrated, data-driven framework for designing low-carbon concrete that balances mechanical performance, environmental impact, and cost. A comprehensive database of 1,273 compressive strength tests was used to train and validate several Machine Learning (ML) and Deep Learning (DL) models, including Random Forest (RF), Decision Tree (DT), Gradient Boosting Ensemble, and Bayesian Neural Networks. The RF model achieved the best predictive performance (R2 = 0.68, RMSE = 4.85 MPa) and served as the core predictor for mix optimization. Model interpretability was addressed using SHAP (SHapley Additive exPlanations), which identified the water-to-binder ratio, Total binder content, and filler percentage as the most influential factors on compressive strength. Local SHAP analyses revealed the nonlinear effects of supplementary cementitious materials (SCMs) such as slag and filler, providing insights into optimal substitution levels. Optimized mix designs were generated via a Genetic Algorithm (GA), simultaneously targeting 28-day strength, CO2 emissions, and cost. Life Cycle Assessment (LCA) and cost analyses demonstrated that moderate slag substitution (~25%) achieves the best trade-off, reducing CO2 emissions by ~60 kg/m3 while maintaining target strength. Large-scale simulations exceeding 10 million virtual mixes enabled the creation of decision-support maps linking SCM ratios to strength classes, highlighting the interplay between mechanical, environmental, and economic objectives.This framework demonstrates that combining predictive ML models, interpretable ML, GA optimization, and LCA can effectively guide the design of sustainable, low-carbon concrete, offering actionable insights for both research and industrial applications.

The emergence of 3D printing concrete technology can revolutionize the construction industry by offering improved precision, customization, and sustainability. However, evaluating the printability of cement-based materials remains a challenge due to the complex rheological requirements they must meet. In many 3D printing applications, stiff cement-based materials with a high rate of structuration are often used in 3D printing applications to ensure the buildability and shape stability of printed elements. Yet, the inherent stiffness of these materials poses challenges of appropriate characterization methods to assess the rheological parameters related to printability. To address these challenges, this paper introduces the squeeze test as a simple yet effective test method. By applying a controlled compression load between two plates, the squeeze test replicates the conditions encountered during 3D printing processes. Printability characteristics are assessed by analyzing the material's response under various testing conditions. The paper outlines the methodology, highlighting its flexibility to simulate both the extrusion and layer-deposition stages. The study examines extrudability and buildability criteria based on squeeze test results for stiff cement-based materials, identifying key factors such as the impact of sand content on consolidation rate and plasticity, and the influence of viscosity modifying agents (VMA) on material elasticity. The findings reveal that increasing sand content reduces plastic deformations, making buckling failure more common, while VMA additions improve elasticity and plastic yield. Additionally, the squeeze test is extended to evaluate interlayer bonding by conducting tests on bi-layer samples, offering a means to fine-tune interlayer printing time. This experimental study validates the squeeze test as an effective tool for assessing printability and predicting failure mechanisms, particularly for stiff cement-based materials.

The mechanisms underlying creep in cementitious materials have been extensively studied, with pioneering contributions shedding light on its origins. However, the amplitude and precise causes of creep remain less understood, particularly in modern mixes incorporating significant proportions of supplementary cementitious materials (SCMs). Nanoindentation offers a rapid and reliable method for investigating creep at the microscale, enabling the evaluation of individual cementitious phases under small loads over short durations. In this study, the creep of five cement pastes incorporating ground-granulated blast furnace slag (GGBFS) and calcined clay was studied. The mechanisms of the observed creep on the pastes scale were discussed in the light of nanoindentation results. Gaussian mixture model (GMM) and scanning electron microscopy (SEM) imaging (GMM-SEM) were combined to assess the micromechanical properties and the volume fractions of the individual phases in these binders. Results shows that GGBFS incorporation results in the formation of low density C-S-H, but the creep of the overall paste is balanced by the high amount of unhydrated particles (of cement and GGBFS) compared to the other blended pastes. In addition, the joint incorporation of GGBFS and calcined clay enables the formation of high density C-S-H more resistant to creep. Finally, the results of GMM-SEM were supplied to a recent analytical homogenization scheme and creep of the different pastes was accurately predicted compared to microindentation measurements. The results of this study may be used to design blended concretes with high resistance to creep and facilitate micromechanical properties evaluation.

The present article introduces a novel and rapid method for processing nanoindentation data using SEM image analysis, called Auto-NI-SEM (Automatic NanoIndentation-Scanning Electron Microscopy). The originality of the method lies in the exploitation of micromechanical data, coupled with image analysis, by developing a suitable automatic deconvolution method. The method leads to a rapid phase quantification of cement paste phases along with their elastic and viscoelastic properties, essential for accurate prediction of macroscopic properties. The results were validated on cement pastes with different water-to-cement (w/c) ratios. Furthermore, the method was found to be reliable for the detection of minor phases, such as for portlandite in low water-to-cement ratios cement pastes, whose volume fractions are in line with those measured by thermogravimetric analysis. Finally, the refined properties helped to accurately predict the cement pastes creep using an analytical homogenization scheme, which was confirmed by microindentation measurements.

Self-compacting mortars and concretes, commonly used in horizontal structures, exhibit superior fluidity and homogeneity. However, these materials are prone to significant shrinkage and cracking, especially in floors and slabs due to large moisture exchange surfaces. Incorporating fibres offers an efficient alternative to traditional reinforcement methods by reducing shrinkage-induced cracking. This study evaluates the effectiveness of glass fibres, polypropylene mono-filament, and polypropylene multi-filament fibres in controlling cracking phenomena. Three-point bending tests with acoustic emission (AE) monitoring were conducted to investigate different stages of cracking mechanisms. The results indicate that while fibre addition does not significantly affect the maximum force, fibre dosage and type substantially influence post-peak mechanical behaviour, particularly the residual force and toughness index. In order to identify the source failure mechanisms, acoustic emission signals were analysed. Two types of investigations are performed: AE signature investigation and AE multi-parametric investigation. Two clustering methods: Multivariable K means non supervised and KNN supervised machine learning methods are applied to identify the contribution of each failure mechanism namely; fibre breakage, fibre-matrix sliding, and matrix cracking. Waveform analysis further revealed characteristic AE signatures for each mechanism, highlighting the potential of AE for understanding and optimizing fibre-reinforced mortar’s crack resistance. The combination of AE and machine learning clustering methods. The supervised KNN model achieved a high classification performance in terms of distinguishing the mechanisms confirming the reliability of the combined AE–machine learning approach.

Enhancing the early-age mechanical performance of low-carbon cementitious materials is essential for their wider adoption. While thermal curing is widely used in precast and prestressed concrete, and carbonation curing has recently emerged for strength improvement and CO2 sequestration, their combined effect remains unexplored. This study introduces a novel curing regime that combines thermal and carbonation curing to assess its impact on strength development and CO2 capture. Four curing conditions (humid, thermal, carbonation, and combined thermal and carbonation) were applied to cement pastes with only CEM I then with 60 % clinker replacement using binary and ternary blends of blast-furnace slag and calcined clay. Results show that carbonation curing achieved the highest CO2 uptake (up to 14 % of slag binder mass), while thermal curing significantly enhanced early compressive strength. Notably, the combined thermal-carbonation curing led to the highest 48 h strength gain, particularly for the ternary binder, and improved micromechanical properties, as evidenced by higher indentation and creep modulus. These findings highlight the potential of this innovative curing approach to optimize both mechanical performance and carbon sequestration in sustainable cementitious materials.

The use of calcined clays in cementitious systems has lately gained extensive attention, especially regarding the current context of environmental issues. Several studies focused on the hydration aspects of cement binders in the presence of pure calcined clays, but few are the studies that were conducted on the mechanical aspects of these same systems with low-grade calcined clays. In this work, four cementitious systems incorporating slag, low-grade and high-grade calcined clay are studied, and their hydration and mechanical properties are compared. The incorporation of slag enabled highlighting the synergetic effect between slag and low-grade calcined clay and the ability of this ternary binder to be equivalent to the binary binder with high-grade calcined clay in terms of: 1-hydration speed between 7 and 28 days and 2-mechanical properties at 28 days. Finally, a novel deconvolution method of grid nanoindentation results is presented, coupling XRD/Rietveld analysis and classic Gaussian Mixture Model deconvolution method; and the advantages and disadvantages of this method are discussed.

Low-clinker concretes offer potential for reducing CO2 emissions of concrete and construction industries. However, it is important to consider the increased risk of corrosion resulting from carbonation. This article aims to assess the effect of accelerated carbonation on the electrical resistivity of concrete, which is closely related to the risk of corrosion propagation. An experimental study was conducted on six concretes based on binary binders (Clinker—Limestone and Clinker—Slag) and ternary binders (Clinker—Slag—Limestone) to characterize their resistivity and porosity in both non-carbonated and carbonated states. This study is complemented by microstructural characterization of cement pastes with the same water-to-cement ratio and containing the same binders, in both non-carbonated and carbonated states, using techniques such as TGA and XRD. The results show that carbonation significantly affects the electrical resistivity of concrete, with variations up to + 305

The common phenomenon observed for concrete in aggressive water is leaching, which involves the dissolution of cement hydration products. Many studies have focused on leaching in demineralised water or acid attacks, but mineral water still deserves further investigation. In most standards, the aggressiveness of a given water body is determined by its pH and not its composition. The effect of the calcium content of the water on degradation is yet to be determined. In this paper, the leaching of Portland cement-based mortar was induced by two types of drinking water with different calcium contents and buffer capacity in controlled conditions. The Langelier saturation index (LSI) was used to describe water aggressiveness based on the calco-carbonic equilibrium. The studied waters had the same pH but LSIs of +0.5 and −1.0 corresponding to scaling with respect to aggressive water; demineralised water was used as a reference. Microstructural damage was checked by TGA and X-ray microtomography. Macroscopic measurements were used to monitor global degradation. The soft water caused a 53% deeper deterioration of the mortar sample than the hard water. Soft water-induced leaching was found to be similar yet slower to leaching via demineralised water (with a mass loss of −2.01% and −2.16% after 200 days, respectively). In contrast, hard water induced strongly time-dependent leaching, and the damage was located close to the surface. The roughness of leached specimens was 18% higher in hard water than in soft water. The formation of calcite on the sample surface not only affects the leaching rate by creating a protective surface layer, but it could also act as a calcium ion pump.

This study focuses on the use of alkali-activated materials and geopolymer grouts in deep soilmixing. Three types of grouts, incorporating metakaolin and/or slag and activated with sodium silicate solution, were characterized at different scales to understand the development of their local structure and macroscopic properties. The performance of the soilmix was assessed by using combinations of the grouts and model soils with different clay contents. Feret’s approach was used to understand the development of compressive strength at different water-to-solid ratios ranging from 0.65 to 1. The results suggested that incorporating calcium reduced the water sensitivity of the materials, which is crucial in soilmixing. Adding soils to grouts resulted in improved mechanical properties, due to the influence of the granular skeleton. Based on strength results, binary soilmix mixtures containing 75% of metakaolin and 25% of slag, with H2O/Na2O ratios ranging from 28 to 42 demonstrated potential use for soilmixing due to the synergistic reactivity of metakaolin and slag. The optimization of compositions is necessary for achieving the desired properties of soil mixtures with higher H2O/Na2O ratios.

The present article introduces a novel and rapid method for processing nanoindentation data using SEM image analysis, called Auto-NI-SEM (Automatic NanoIndentation-Scanning Electron Microscopy). The originality of the method lies in the exploitation of micromechanical data, coupled with image analysis, by developing a suitable automatic deconvolution method. The method leads to a rapid phase quantification of cement paste phases along their elastic and viscoelastic properties, essential for accurate prediction of macroscopic properties. The results were validated on cement pastes with different water-to-cement (w/c) ratios. Furthermore, the method was find to be reliable for the detection of minor phases, such as for portlandite in low water-to-cement ratios cement pastes, whose volume fractions are in line with those measured by thermogravimetric analysis. Finally, the refined properties helped to accurately predict the cement pastes creep using an analytical homogenization scheme, which was confirmed by microindentation measurements.

Cements made with pozzolanic materials and/or latent hydraulic materials are developed to replace Portland cement and decrease the impact of cement production on the CO2 emissions. Those cements need to be qualified regarding their performances and durability, knowing they exhibit good behavior when exposed to external sulphate attack. Pozzolanic materials offer a specific composition of hydrated cement phases with less portlandite and more C-A-S-H than Portland cement. Moreover, they contain a reduced proportion of clinker, hence a reduced share of C3A. All these characteristics explain the unique response of pozzolanic cement to external sulphate attack. However, the pozzolanic reaction is more than often incomplete, with some remaining portlandite in the hydrated cement paste. In order to investigate the role played by this remaining portlandite fraction during external sulphate attack, the behavior of two CEM III and two CEM IV cement-based mortar samples were studied and compared to two CEM I- type cements, one SR and one non-SR. The pozzolanic nature of these cements has been investigated by monitoring the portlandite content of cement paste over time by thermogravimetric analysis. In addition, the behavior of synthetic C(-A)-S-H containing portlandite and in suspension in sulphate solutions has been assessed by X-ray diffraction and scanning electron microscopy. Results have confirmed the sulphate resistance of the studied blended cements compared with a sulphate-resistant Portland cement (CEM I SR3), although the cements still containing portlandite did undergo a slight expansion. The test on C(-A)-S-H showed that the presence of portlandite encourages the precipitation of ettringite within the C-A-S-H nano porosity, even in the absence of pH regulation.

Shear failure is one of the most catastrophic mode of failure in reinforced concrete elements. In order to prevent shear failure, the factors responsible for shear capacity and shear transfer mechanisms should be completely understood. These mechanisms involve aggregate interlocking, the compression zone, the dowel action and the residual tensile stresses of the concrete. Further, these mechanisms are accompanied and sometimes driven by complex patterns of shear cracking where the shear crack inclination angle plays an important role. In this study, we analyzed the different shear transfer mechanisms and their role in the event of shear failure. Different theoretical models coupled with experimental measurements are used in this work. Secondly, cracking growth were monitored by simultaneously applying acoustic emission based b-value analysis and the digital image correlation technique. A new insight is provided regarding the correlation between the acoustic emission parameters, b-value, the crack propagation and the stress drops. Three sizes of beams are investigated in this study. The results show that, for large and medium beams, shear stress drops quantified by the DIC technique is well captured by the b-value parameter. Crack sliding and Crack openings are important process during shear failure. The b-value characterizes the scaling of the magnitude distribution of AE events. This parameter is very sensitive to stress drops and stress redistribution. The b-value seems also related to the dimensionality of the volume available for cracking.

The current context of global warming requires urgent action to reduce carbon emissions from all industries, including the concrete production. To meet this challenge, the European cement standard, EN 197–5, has been implemented to introduce CEM II/C-M low-carbon Portland cements incorporating a higher percentage of mineral additions (up to 50%), and a new family CEM VI cements with a very low clinker content, ranging from 35 to 49%. This work focuses on these new cements, mainly ternary mixtures based on clinker, slag and limestone filler for high clinker replacement rates. This choice is based on the exploitation of the synergy between limestone and mineral additions containing aluminates which has been established in the literature. This study is focused on concretes based on these new ternary clinker-slag-limestone cements, to compare them with existing binary cements in terms of durability with respect to corrosion induced by carbonation. Two main tests were performed: accelerated carbonation that reveals corrosion initiation time, and electrical resistivity that provides insights into corrosion propagation time. Additional microstructural characterization was also carried out to understand the link between the macroscopic durability properties and the microstructure. A service life prediction model was used to quantify the service life of the concrete cover from material properties. Concretes were then compared on the basis of two carbon efficiency indicators corresponding respectively to the carbon cost of 1 MPa and one year of service life. Opposite trends are observed on the carbonation resistance and electrical resistivity. Good durability can be expected for low carbon cements, as they have high electrical resistivity and therefore a long corrosion propagation time, despite their relatively low carbonation resistance,. The carbon footprint indicator related to service life is more relevant than the one related to mechanical performance since it takes into account an important function of concrete which is durability.

Recycled concrete aggregates (RCA) lead to higher shrinkage of concrete than natural aggregates because of two main reasons: lower elastic modulus and possible shrinkage of RCA. The study aims at quantifying the effects of both phenomena, thus Trisphere model has been chosen as it takes into account the influence of aggregates modulus and shrinkage on the properties of con-crete. The existing model has been adapted to determine the elastic modulus and shrinkage of recycled aggregates from the modulus and shrinkage of concrete. The shrinkage of coarse RCA was much lower than the shrinkage of fine RCA, thus the increase in the shrinkage of concrete can be mainly attributed to the reduction in aggregate modulus and variations of composition pa-rameters (W/C, volume of paste). The model also allows estimating the delayed elastic modulus of concrete. The obtained values are compared with the numerical determination of viscoelastic-aging Young's modulus from restrained shrinkage tests (ring tests). The long-term moduli pro-vided by both approaches are consistent for most of studied materials. The relaxation increased with the content of coarse recycled aggregates, but the variation was lower than reported in creep studies. This could be due to lower relative stress level as recycled aggregates also induce lower elastic modulus thus lower stress for a given restraint.