Development of mathematical models to predict the atmospheric corrosion rate of carbon steel in fragmented subtropical environments

Mathematical modelling of atmospheric corrosion based on the aggressiveness categories defined by standard ISO 9223 greatly fail to predict the actual corrosion rates of metals in subtropical environments. Therefore, new concepts for modelling are required as to adequately predict corrosion rates from environmental factors such as the deposition rate of chemical agents (namely chloride and sulphur dioxide), climate effects (such as moisture and time of wetness), and the duration of metal exposure. The novel methodology is based on the definition of a set of qualitative variables to distribute locations exhibiting distinctive initial characteristics towards metal corrosion. The validity of the method was checked by using data obtained during three years of exposure of carbon steel in 74 stations distributed along the seven main islands of the Canary Islands (Spain). A definite evaluation of the impact of environmental factors on the extent of corrosion was achieved, and good results were defined in terms of fitting quality.