Machine Learning and Deep Learning Methods for Evaluating Mechanical Properties of Fiber-Reinforced Concrete Containing Crumb Rubber: A Review of recent developments and future prospects

Fiber-reinforced concrete (FRC) that has crumb rubber (CR) is stronger and more flexible than simple concrete. Furthermore, FRC is stronger and manages impacts with greater durability. Using CR can sometimes lead to lower strength and flexibility in materials. This problem can be improved by adding steel fibers. The mechanical performance of rubberized fiber-reinforced concrete presents difficulties that conventional experimental and computational techniques often fail to address adequately. Recent developments in artificial intelligence (AI), machine learning (ML), and deep learning (DL) have shown potentially effective modeling and predicting the mechanical properties of FRC-CR. Research utilizing ML algorithms, including random forests and decision trees, alongside DL methodologies such as Long Short-Term Memory (LSTM) networks, has successfully predicted compressive strength and other significant properties. These approaches utilize input variables such as the water-cement ratio, the percentage of CR replacement, and the age of the concrete to enhance performance metrics like Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). Despite these advancements, challenges persist in adapting models to various material features, ensuring the interpretability of models, and managing uncertainties that are intrinsic to data-driven methodologies. Future research should focus on standardizing approaches, incorporating natural fibers into FRC-CR, and investigating advanced ML/DL models, such as ensemble techniques and gated recurrent units, to improve the accuracy and reliability of predictions regarding concrete properties. This review brings together current studies, points out their weaknesses, and suggests new research ideas to improve the use of machine learning (ML) and deep learning (DL) in understanding fiber-reinforced concrete with crumb rubber (FRC-CR). This study aims to advocate for the implementation of sophisticated artificial intelligence algorithms to enhance the forecasting of eco-friendly concrete and elevate the standards of infrastructure quality.

کلیدواژه ها:

Mechanical properties ، Fiber-reinforced concrete ، Crumb rubber ، Machine learning ، Deep learning