Ström, M. and Ström, G., "A Statistically Designed Study of Atmospheric Corrosion Simulating Automotive Field Conditions Under Laboratory Conditions - Final Volvo Report on the AISI Cosmetic Corrosion Set of Materials," SAE Technical Paper 932338, 1993, https://doi.org/10.4271/932338.
An extensive atmospheric corrosion test program to simulate automotive field conditions has been successfully completed. This paper focuses on the corrosion results from the widely exposed AISI set of correlation panels for cosmetic corrosion. Eight factorially designed 12-week tests have been performed in the laboratory, using high performance test equipment, capable of simulating diverse outdoor conditions. The results have been compared with those of outdoor scab exposures and reference panels on vehicles, running in Canada and Sweden.The influences of six corrosive test variables on the response creep-back from scribe on the painted panels are demonstrated, based on a statistic evaluation of the test matrix. All higher settings of the introduced accelerating test variables have each resulted in a decrease in the test correlation with on-vehicle exposures. A high degree of correlation regarding on-vehicle exposure was reached when the test conditions were comparatively mild (acceleration factor < 10). However, to achieve a high degree of correlation to the Volvo outdoor scab test, the test conditions corresponded to the high settings of the variables. The results of the test program clearly suggest that the predictability of a test is generally improved when simulating outdoor conditions under mild acceleration combined with linear extrapolation instead of forcing the corrosion process in order to establish the postulated damage within a comparatively short time. From the results of the statistical design in the high performance equipment, simplified test methods, suitable as standards for the small scale laboratory, have been worked out. The results of the AISI set from two such tests are presented. One provides excellent correlation to on-vehicle behaviour.The scribe creep propagation for zinc and zinc-iron coated materials are convincingly shown to be inversely related to the zinc-coating thickness, both in the eight Volvo tests and in field conditions on vehicle. Hence, the paint undercutting is directly determined by the progress of the dissolution front of the zinc coating and the anodic consumption rate of zinc is constant. A plausible explanation for this effect is that the overall corrosion rate is limited by constant-rate oxygen reduction, either in the scribe (at limited undercutting) or by a stabilised cathodic zone tailing the anodic dissolution front (at extensive creep-back from scribe).