Models for chloride diffusion of concrete employing fuzzy-based algorithms

Chloride ion exposure is one of the obstacles that reinforced concrete structures may be required to endure over the course of their lifetime. Concrete constructions may become less durable and degrade as a result of this exposure, especially in coastal locations. Artificial intelligence (AI) may be used to develop models that can reliably forecast the chloride diffusion coefficient (D_C) of unsteady state apparent concrete over an extended period of time by using practical field data. This method may enhance the assessment of a concrete construction’s longevity by identifying the variables that matter most. The current study demonstrates the utilization of an Adaptive neuro-fuzzy inference system (ANFIS) for the prediction of concrete’s D_C under various exposure scenarios. After being trained on a dataset including 216 data points, the forecast models were enhanced by using the Harris Hawks optimization (HHO) and Chimp optimization algorithm (COA). The results show that the ANF(COA) and ANF(HHO) models show great potential for accurately predicting the D_C of concrete in various exposure scenarios while preserving appropriate coefficient of determination (R²) values. A thorough index using a variety of metrics shows that the objective function (OF) value for the ANF(COA) was about 30% lower at 0.5235 than it was for the ANF(HHO) 0.7738. By providing reliable D_C predictions, the models can inform maintenance strategies, guide material selection, and support the design of more durable concrete structures, ultimately improving infrastructure resilience and reducing long-term maintenance costs.