Optimization of effective chloride migration in mortar containing nano titanium dioxide using bio-inspired metaheuristics

Abstract

This research has investigated the impact of nano-TiO₂ (NT) as a partial replacement for cement (0–3%) on chloride migration in cement mortar and has established a predictive framework utilizing advanced metaheuristic algorithms. Experimental results have indicated that an increase in NT content has markedly delayed the onset of steady-state diffusion, with the duration required to reach steady state having increased by approximately 45% at 3% NT, primarily as a result of pore refinement and matrix densification. The chloride diffusion coefficient has consistently been decreased with a lower water–cement ratio and a higher NT content, indicating enhanced resistance to chloride ingress. To accurately model this nonlinear behavior, ten metaheuristic algorithms have been employed: Bat Algorithm, Cuckoo Search, Differential Evolution, Harris Hawks Optimization, Grey Wolf Optimizer, Firefly Algorithm, Jaya, Moth–Flame Optimization, Multi-Verse Optimizer, and Particle Swarm Optimization, with the objective of minimizing mean squared error. Differential Evolution, Multi-Verse Optimizer, and Particle Swarm Optimization have achieved the fastest convergence, reducing the mean squared error to below 0.5 within 20 iterations, whereas Cuckoo Search, Grey Wolf Optimizer, Harris Hawks Optimization, and Jaya have demonstrated consistent and reliable performance. Thus, the integrated experimental and computational approach has provided a dependable data-driven framework for forecasting chloride transport in NT-modified mortar and has offered valuable insights into the design of durable and sustainable concrete.