Predicting Coating Failure using the Central Limit Theorem and Physical Modeling

Abstract

In service, coatings are assaulted by huge numbers of ultraviolet photons, moisture molecules, pollutant molecules, thermal or other stresses with additional, occasional insults. Individually, these arrive at indeterminate times and locations so either Monte Carlo simulations or the statistics of random processes can be used to estimate the accumulation of damage. These damage results can then be used to calculate the deterioration in coating properties using well-known and robust physical models. In this way, not only can appearance properties like gloss and yellowing be modelled, but also the coating integrity via its fracture toughness or barrier properties, via electrochemical impedance. Additional physical detail or processes may be included, at several points, to produce a more complete predictive simulation or interpretative scheme. Simple calculations of how accelerated weathering affects coating service lifetime also demonstrate how inconsistencies may easily arise between accelerated testing and natural exposure.