Title:
Application of Artificial Intelligence for Accurate Chloride Permeability Predictions in Concrete Structures (Prepublished)
Author(s):
Mohamad Kharseh and Fayez Moutassem
Publication:
Materials Journal
Volume:
Issue:
Appears on pages(s):
Keywords:
artificial neural networks (ANN); chloride permeability; concrete structures; durability; sustainability
DOI:
10.14359/51749256
Date:
10/8/2025
Abstract:
The durability of reinforced concrete is often compromised by chloride penetration, leading to corrosion of reinforcing steel and reduced structural strength. To improve the sustainability and longevity of concrete structures, it is crucial to model and predict chloride permeability (CP) accurately, thereby minimizing the time and resources required for extensive experimental testing. This paper presents a proof-of-concept study applying Artificial Neural Networks (ANN) to predict CP in concrete structures. The model was trained on a small but carefully controlled experimental dataset of 10 concrete mixtures, considering four key parameters: water-to-cementing materials ratio, silica fume content, cementing materials content, and air content. Despite the limited dataset size, which constrains generalizability and statistical robustness, the ANN captured nonlinear relationships among the input parameters and CP. The comparison between experimental and simulated CP values showed reasonable agreement, with errors ranging between –242 and 420 Coulombs. These results establish the trustworthiness and reliability of the proposed model, providing a valuable tool for predicting CP and informing the design of durable and sustainable concrete structures.