Bayesian-Guided Development and Experimental Validation of a Hybrid Self-Healing Nanoencapsulated Epoxy Coating for Reinforcing Steel Applications

Abstract

This study combines formulation selection, experimental validation based on the standards, uncertainty analysis, reinforcing-bar bond, and green-cost analysis of a hybrid self-healing nanoencapsulated epoxy system with the aid of a Bayesian approach. Four coated systems (E1 – E4 consisting of 5 independent coated coupons per formulation) were compared to an uncoated reference system (B0). E4 contained graphene-oxide diffusion-barrier reinforcement, IPDI-type microcapsules, pH-responsive ZIF-8@Ca-type microfillers, MCM-41/PEI/PSS nanocontainers, and a dynamic hydrogen-bonding component. The confirmatory tests included electrochemical impedance spectroscopy tests for intact, damaged, and healed conditions; polarization-based corrosion assessment; 720 h neutral salt-fog exposure; pull-off adhesion; microcrack closure; 90-day storage-preservation assessment; and SEM/TGA characterization. Each group had a number of test specimens per coating/size group (n = 3, for each group), and were tested against galvanized reinforcing bars in accordance with ASTM A944. E4 produced log10|Z|0.01Hz = 11.59 ± 0.08 and i_corr = 0.065 ± 0.004 μA cm−2, corresponding to a calculated corrosion rate of 7.54 × 10−4 mm y−1 and a 97.6% reduction relative to conventional epoxy E1. E4 showed underfilm corrosion ≤ 0.35 ± 0.04 mm after 720 h of salt fog, whereas it still had 95 ± 2% adhesion and 93 ± 3% microcrack closure after 24 h. Its DQI was 0.91 (with a 0.88-0.94 uncertainty interval), and it was ranked 1st in 94% of ±20% weight-perturbation runs, and 91% of the 10,000 Monte Carlo draws. Two-way ANOVA for the reinforcing bar dataset showed that there was a significant difference in failure load for bar size (p = 0.0015) but not for surface condition (E4 vs. galvanized) (p = 0.169). The central plant scenario resulted in a net value of USD 32 t−1 and an annual projected gain of USD 5.568 million for                                     174,000 t y−1 of protected product. The results from all sources indicate that E4 is the recommended formulation and offer a traceable pathway for making experimental to industrial decisions.