Compressive Strength Prediction in Cement-Based Materials

The compressive strength of cement-based materials constitutes a primary measure of structural performance and durability in diverse civil engineering applications. Predicting this parameter efficiently has long been a core challenge, guiding mixture design, quality control and service-life estimation while reducing reliance on protracted laboratory tests. Traditional methods rely on empirical correlations between strength, water-to-cement ratio and curing age, but recent advances in statistical methods, multiscale modelling and machine learning have expanded the toolkit substantially. Regression analyses, decision-tree models and ensemble approaches facilitate rapid estimation across varied mix designs, while deep learning frameworks leverage large datasets to capture nonlinear interactions among admixtures, supplementary cementitious materials and curing regimes. These predictive capabilities underpin optimised material formulations that enhance resource efficiency and lower the carbon footprint of cement production. Moreover, by integrating microstructural descriptors and time-dependent hydration kinetics, contemporary models bridge fundamental science with practical implementation. The evolving landscape of compressive strength prediction thus supports global efforts to improve infrastructure resilience, accelerate innovation in sustainable cementitious composites and meet ambitious environmental targets.

Research from Nature Portfolio

No recent Nature Portfolio content available.

Research from all publishers

Recent investigations have harnessed advanced computational approaches to predict the compressive strength of cementitious systems across diverse formulations. A novel hybrid deep learning model combining one-dimensional convolutional neural networks with long short-term memory layers has been applied to permeable and porous concretes incorporating high volumes of fly ash, achieving prediction accuracies exceeding 95 per cent. By encoding parameters such as water-to-cement ratio, fly-ash content and curing time, the model demonstrates the potential of sequential feature extraction in forecasting strength development alongside carbonation depth. In another study, multiple regression techniques—including linear, nonlinear and M5P-tree models—were assessed for high-strength concretes modified with metakaolin. Here, the tree-based framework outperformed traditional regressions in capturing nonlinear influences of binder composition and curing age, offering a cost-effective predictive tool for mix optimisation. Complementing these advances, statistical and soft-computing techniques have been employed to evaluate green concretes containing ground granulated blast furnace slag under varying temperatures. Comparative analyses of quadratic models, full-quadratic formulations and artificial neural networks revealed that multilayer perceptron architectures best account for complex interactions among binder ratios, aggregate proportions and thermal histories in strength estimation.

Compressive Strength Prediction in Cement-Based Materials publication trend

The graph below shows the total number of articles in compressive strength prediction in cement-based materials across all publications each year (not limited to Nature Index journals).

Technical terms

Compressive strength: The maximum axial stress that a material can withstand under crushing load.

Water-to-cement ratio (w/c): The mass ratio of mixing water to cement, a primary determinant of hydration and matrix porosity.

Artificial neural network (ANN): A computational model inspired by biological neurons, used to learn complex input–output relationships for prediction.

Convolutional neural network (CNN): A class of ANN that employs convolutional layers to extract hierarchical spatial features from input data.

Long short-term memory (LSTM): A recurrent neural network architecture designed to capture long-range temporal dependencies in sequential data.

Pozzolanic reaction: A chemical interaction between silica- or alumina-rich materials and calcium hydroxide that forms cementitious compounds.

Ground granulated blast furnace slag (GGBFS): A by-product of iron smelting used as a supplementary cementitious material to enhance concrete properties.

References

1. A hybrid model based on convolution neural network and long short-term memory for qualitative assessment of permeable and porous concrete. Case Studies in Construction Materials (2023).
2. Mathematical modeling techniques to predict the compressive strength of high-strength concrete incorporated metakaolin with multiple mix proportions. Cleaner Materials (2022).
3. Predicting the Compressive Strength of Green Concrete at Various Temperature Ranges Using Different Soft Computing Techniques. Sustainability (2023).