AI-based prediction of the optimal incorporation rate of dredged sediments in concrete: Mechanical performance analysis

The management of marine dredged sediments is a critical environmental and economic issue, particularly in port cities where dredging is a necessary activity to maintain navigability. These sediments are typically viewed as waste products and often require costly and environmentally challenging disposal methods. However, repurposing dredged sediments as a component in concrete production presents a promising solution for both waste management and the creation of sustainable construction materials. Despite this potential, determining the optimal percentage of sediment incorporation and accurately predicting the mechanical properties, such as compressive strength, remain significant challenges. This study proposes an artificial intelligence (AI)-based approach to predict the optimal incorporation percentage of marine dredged sediments from Moroccan ports into concrete and to forecast the resulting compressive strength. A dataset consisting of 104 samples, including dune sand and port sediments from JEBHA, was used. The data includes key properties such as granulometry, cleanliness, fineness modulus, and the compressive strength of the concrete mixtures. These experimental data were employed to train and validate several machine learning models, including linear regression, Random Forest, Gradient Boosting, and XGBoost, chosen for their ability to model complex, non-linear relationships between sediment characteristics and concrete performance. The performance of these models was evaluated using two key metrics: the coefficient of determination (R²) and the root mean square error (RMSE). Among the models tested, the Random Forest Regressor delivered the best results, with an R² value greater than 0.98 and an RMSE of less than 0.20 MPa, indicating highly accurate predictions of both the optimal sediment incorporation rate and the compressive strength of the concrete. This model’s exceptional performance underscores its potential as a reliable tool for optimizing the use of dredged sediments in concrete production. The findings of this study demonstrate the considerable potential of AI in optimizing the incorporation of marine dredged sediments into concrete. By accurately predicting the mechanical properties of the resulting material, this approach enables the development of more sustainable construction practices while reducing the environmental burden associated with sediment disposal. Moreover, this work illustrates the broader applicability of AI in addressing environmental challenges, offering a pathway to valorize waste materials in the construction industry. The study not only advances our understanding of sediment utilization in concrete but also contributes to the growing field of sustainable material science, offering promising avenues for future research and development. Nevertheless, further research is needed to validate the model’s scalability to other sediment types and assess practical limitations in industrial applications.