Glass fiber reinforced cement is utilized in construction field. Interfacial modification with coupling agents and functional polymers represents a critical strategy to enhance interfacial bonding. However, molecular mechanisms remain insufficient. In this study, molecular dynamics were used to evaluate interfacial responses of C–S–H/glass fiber modified with KH-570 and poly (vinylpyrrolidone-co-acrylic acid) (PVP–co–PAA). The results demonstrated that KH-570 induced marginal increase, while PVP–co–PAA led to remarkable enhancement. The analyses of coordination number, bond lifetime decay decoded interfacial bonding networks before and after modification. This study could reveal the distinct failure pathways and molecular mechanisms of interfacial bonding.

All the MD simulations were performed based on a package named LAMMPS. Firstly, all the systems were relaxed under NVT ensemble last for 5 ns, with timestep 1 fs and 300 K. Temperature control was achieved by Nose-Hoover thermostat and damping parameter 0.1 ps. The bottom Ca layers of C–S–H were fixed throughout the simulation to prevent model rotation. An additional 2 ns relaxation was conducted and trajectories were recorded every 1 ps. During the shear simulations, spring forces were applied on atoms located at the top of SiO₂ slab.

The molecular dynamics (MD) simulations are employed to evaluate shear mechanical properties of C–S–H/SiO₂, C–S–H/KH-570 and C–S–H/PVP–co–PAA interfaces. The molecular mechanisms of interfacial modification are thoroughly revealed via analyzing the types and dynamic behaviors of interfacial bonding networks. During dynamic failure, PVP–co–PAA exhibits a distinct behavior compared with rigid SiO₂. Flexible polymer networks undergo conformational changes, leading to an increased shear displacement. The density analyses of bonding networks indicates that interfacial bonding networks can reorganize from O_SiO2—Ca—O_C–S–H into O_KH-570/O_{PVP–co–PAA}—Ca—O_C–S–H. Moreover, interfacial failure is decided by ionic bonding of O_SiO2/O_KH-570/O_{PVP–co–PAA}—Ca. The coordination number follows an increasing order of O_SiO2<O_KH-570<O_{PVP–co–PAA}, explaining the differences of interfacial mechanical strength.

The MD simulation revealed that PVP–co–PAA modification showed the most enhancement, with an improvement of 302.7%. The coordination number followed an increasing order of O_SiO2<O_KH-570<O_{PVP–co–PAA}. The time correlation function (TCF) also revealed that interfacial stability was enhanced after modification. Based on the coordination number and the TCF results, the differences of interfacial mechanical strength were explained. These findings could provide fundamental insights to guide the functional design of fiber-cement interfaces.

Predicting the compressive strength of Steel Fiber Reinforced Concrete (SFRC) is challenging due to the complex, coupled influences of numerous mix design parameters. Traditional experimental methods are often costly and time-consuming for exploring this high-dimensional factor space. While machine learning (ML) models offer a powerful alternative, their practical application is hindered by two major limitations: the scarcity of comprehensive, high-quality experimental data which restricts model performance and generalizability, and the inherent "black-box" nature of advanced ML algorithms which obscures the underlying mechanisms linking mix proportions to mechanical properties. This study aims to address these critical gaps by developing a hybrid framework that integrates ML-based prediction, data augmentation via Generative Adversarial Networks (GAN), and post-hoc model interpretability using SHapley Additive exPlanations (SHAP). The objective is not only to achieve high-precision strength prediction but also to enhance model robustness with limited data and, crucially, to uncover the influential mechanisms of key components on SFRC compressive strength.

A database of 141 mix designs with 11 input features—Cement (C), Fly Ash (FA), Silica Fume (SF), Water (W), Coarse Sand (CS), Water Reducer (WR), Recycled Aggregate (RA), Natural Aggregate (NA), Fiber Length (FL), Fiber Diameter (FD), and Fiber Content Ratio (FCR)—and the corresponding compressive strength was compiled from peer-reviewed literature. Four tree-based ML algorithms (Random Forest, XGBoost, Stacking, and AdaBoost) were employed for regression modeling. To mitigate the data scarcity issue, a GAN architecture was implemented to generate 282 synthetic but physically plausible samples, expanding the training set. The quality of the GAN-generated data was validated using t-Distributed Stochastic Neighbor Embedding (t-SNE) visualization and statistical moment analysis. The model trained on the augmented dataset was evaluated using the coefficient of determination (R²) and Mean Squared Error (MSE). Finally, the optimal model (Random Forest) was interpreted using the SHAP framework to quantify and visualize the contribution and interaction effects of each input feature on the model's predictions.

Using the original dataset, all ML models demonstrated competent predictive capability. The Random Forest model achieved the best performance with an R² of 0.93 and an MSE of 15.6 MPa², followed closely by XGBoost (R²=0.92, MSE=16.6 MPa²). The introduction of GAN-generated data led to a remarkable and consistent improvement in all models' performance. For Random Forest, the R² increased to 0.98 and the MSE decreased by approximately 56% to 6.9 MPa². The Stacking model achieved an R² of 0.99. This significant enhancement confirms that GAN-based data augmentation effectively enriches the feature space and improves the models' ability to generalize from limited experimental data.

The SHAP analysis provided profound insights into the governing factors of compressive strength. The summary plot identified cement, water, and silica fume as the three most influential features. Cement and silica fume exhibited strong positive SHAP values, highlighting their fundamental role in forming a dense cementitious matrix. The impact of water displayed a distinct non-linear relationship: optimal content promoted hydration and strength, while excess water reduced strength, likely by increasing porosity. The interaction SHAP dependence plots revealed nuanced coupling effects. For instance, the positive contribution of a high WR dosage was more pronounced and stable under conditions of lower water content, emphasizing the critical coupling between water and chemical admixtures in determining the water-to-cementitious materials ratio. In contrast, parameters related to fiber geometry (FL, FD) and content (FCR) showed relatively low and concentrated SHAP values near zero, indicating their secondary role in influencing compressive strength, which aligns with their primary function of improving toughness rather than pure compressive capacity. The SHAP force plots and Sankey diagram further illustrated how specific combinations of high cement/silica fume and optimal water content drive predictions toward high strength, whereas high coarse sand content or inappropriate water levels contribute to lower strength outcomes.

This work successfully developed an integrated framework combining ML, GAN, and SHAP for predicting and interpreting the compressive strength of SFRC. The main conclusions are as follows:1) Among the evaluated ML models trained on the original dataset, Random Forest provided the most accurate predictions for SFRC compressive strength.2) GAN-based data augmentation proved highly effective, substantially improving the predictive accuracy and generalization ability of all models, with performance metrics (R² and MSE) showing dramatic improvement.3) SHAP interpretability analysis quantitatively identified cement, silica fume, and water as the most critical factors affecting compressive strength, with water exhibiting a non-linear effect. The analysis also elucidated important feature interactions, such as that between water and water reducer. Fiber parameters were found to have a minor influence on compressive strength prediction.4) The proposed framework offers a powerful, data-driven tool that moves beyond mere "black-box" prediction. It enhances prediction robustness from small datasets and provides actionable, mechanism-based insights for optimizing SFRC mix design, thereby bridging the gap between empirical data mining and fundamental materials science understanding.

It is vital to maintain the durability of construction materials since sulfate attack is a primary factor influencing the lifespan of concrete in marine setting. Sulfate ions (SO₄^2–) degrade concrete via attacking the calcium silicate hydrate (C-S-H) that is the primary hydration product of cement and the binding phase in concrete. This process primarily occurs through calcium leaching, where Ca²⁺ reacts with sulfates to form gypsum, altering pore structure and weakening mechanical properties. C-S-H exhibits a “long-range disorder, short-range order” molecular structure, consisting of nano-scale particles. It is indicated that sulfate-induced degradation involves a competition between SO₄^2– and silicate ions (SiO₃^2–) for Ca²⁺, leading to a material destabilization. The simulations by molecular dynamics (MD) and quantum chemistry (QC) provide insights into this phenomenon, but they have limitations, i.e., MD lacks chemical reaction accuracy, while QC is restricted in a scale. Understanding the molecular interactions governing Ca²⁺ transfer from SiO₃^2– to SO₄^2– is essential for decoding sulfate-induced degradation. However, the existing experimental and computational methods struggle to capture this process comprehensively. A more effective approach is required to accurately simulate chemical reactions and address the dynamics of Ca²⁺ transport in C-S-H degradation.

This study was to construct a C-S-H model based on tobermorite’s unit cell structure, optimizing elemental balance to achieve a Ca/Si ratio of 1.5. The energy minimization was performed by the GFN-xTB method, with the model embedded within a 13 Å water sphere to simulate aqueous conditions. The simulations by Born-Oppenheimer Molecular Dynamics (BOMD) employing the GFN-xTB method investigated the C-S-H decalcification under sulfate attacks. Two 100 ps NVT ensemble simulations at 298.15 K were conducted, i.e., one in pure water and another in sulfate solution (with one SO₄^2– molecule analyzed for Ca²⁺ detachment). The atomic structures were visualized using VMD. The calculations based on Complementary Density Functional Theory (DFT) (Gaussian 16) focused on critical decalcification steps, isolating structural regions (SO₄^2–, desorbed Ca²⁺, and Si—O tetrahedra). The PBE0 hybrid functional and def2-TZVP basis set were used for single-point energy calculations, with DFT-D3 dispersion correction addressing weak interactions.

This study investigates the decalcification kinetics and structural evolution of C-S-H under sulfate attack using molecular dynamics simulations. In pure water, C-S-H experiences minor structural changes, with surface calcium ions partially dissociating but remaining adsorbed. However, in a sulfate solution, calcium ions desorb rapidly upon sulfate introduction, leading to a structural fragmentation, including Q² to Q¹ transitions and severe silicate chain distortions. The analysis by reduced density gradient (RDG) reveals sulfate’s stronger interaction with Ca²⁺, compared to silicate, facilitating calcium extraction. The analysis by electron localization function (ELF) shows a weak Ca—O_Si bonding, while Ca—O_Sul exhibits a stronger electronic localization, promoting calcium displacement. The analysis by bond critical point (BCP) indicates non-covalent interactions drive decalcification, with competitive binding between sulfate and silicate.

The analysis by electrostatic potential (ESP) indicates an increased C-S-H electronegativity post-decalcification, destabilizing the structure. The energy barrier calculations show that sulfate significantly lowers decalcification and silicate chain fracture barriers, weakening C-S-H integrity. Water further amplifies sulfate-induced degradation, enhancing polarity and self-expansion effects. These findings demonstrate that sulfate corrosion accelerates C-S-H structural failure through enhanced calcium extraction, increased electronegativity, and silicate chain fracture, ultimately compromising mechanical properties.

This study explored C-S-H decalcification and degradation under sulfate attack using quantum chemistry and DFT. Sulfate ions could extract calcium via the Coulomb interactions, confirmed by RDG and BCP analysis. The decalcification could destabilize C-S-H, increasing silicate chain electronegativity, expanding Si—O bonds, and reducing dissociation energy. Silicate chain fractures caused C-S-H gel collapse, weakening cementitious materials. Water could facilitate a degradation via lowering energy barriers and enhancing hydrogen bonding with silicate chains, accelerating structural breakdown. These findings could reveal the fundamental mechanisms of sulfate-induced C-S-H deterioration, thus providing insights into material performance under aggressive environments.