Compressive strength evaluation of cement-based materials in sulphate environment using optimized deep learning technology

Strength serves as a vital performance metric for assessing long-term durability of cement-based materials. Nevertheless, there is a scarcity of models available for predicting residual strength of in-situ structures made of cement-based materials exposed to sulphate conditions. To address this challenge, this study presents a novel approach using deep learning to predict the degradation of compressive strength of cement-based materials under marine environments. Specifically, a deep convolutional neural network (DCNN) is established, consisting of two convolutional layers, one pooling layer, and two fully connected layers. In this innovative model, contents of cement, water-to-cement ratio, sand, sulphate concentration and exposure temperature are selected as inputs, while the output is strength of cement-based materials subjected to sulphate deterioration. To improve the forecast capability, particle swarm optimization is adopted for optimizing hyperparameters of DCNN, which can be implemented by reducing the discrepancy between model prediction and measured strength. Finally, experimental data are used to establish and evaluate proposed method. The results show that the proposed deep learning-based predictive model has the best performance of strength degradation prediction of cement-based materials suffering from sulphate attack via a comparison with other commonly used models. The outcome of this research offers a potential solution for predicting remaining strength of cement-based materials that undergo practical sulphate attack.