Accelerated versus field corrosion of defective painted steel: Acceleration factors and a transferable blister–corrosion relationship

Accelerated corrosion tests qualify protective coatings for steel bridges, yet their quantitative correspondence to field exposure is poorly established, especially for localized degradation at coating defects. We use 3D laser profilometry to quantify coating blistering over time and the terminal substrate corrosion of defective painted steel in four environments, comprising two field sites (Choshi, Pacific-coast; Miyakojima, subtropical marine) and two accelerated tests (ISO 16539 Method B and a composite cyclic test), for four coating systems. The acceleration factor (accelerated/field blister volume at comparable exposure) is large (≈13–200×), strongly coating-dependent, and site-dependent; the composite cyclic test far exceeds ISO 16539, and both greatly exceed the field. The coating ranking is preserved across environments, but accelerated tests inflate the spread between best and worst coatings, so a single acceleration factor cannot serve for quantitative life prediction. Cumulative time-of-wetness only weakly explains field degradation, and the two sites do not share a common dose–response curve; adding measured airborne sea-salt (≈1.8× higher at Miyakojima) as a second driver improves the fit and explains much of the inter-site difference. By contrast, the non-destructive blister volume predicts the post-blast substrate corrosion through a single transferable power law (corrosion volume is proportional to blister volume^0.60, R 2 = 0.89; corrosion depth, R 2 = 0.80) holding across all environments and the three coatings with measurable corrosion. The blister–corrosion mechanism is thus consistent across the environments studied though the rate is not, giving a non-destructive basis for estimating hidden corrosion beneath intact coatings and a foundation for data-driven cross-environment prediction.