No abstract

To understand the corrosion characteristic of metal in atmospheric environment in Taiwan, Taichung Harbor was used as an example in this study, and ISO 9223 standard was followed to conduct an atmospheric corrosion deterioration factor survey and metallic exposure test for four types of screw metals: carbon steel, zinc, copper, and aluminum. The main objective of this study was to understand the classification and ranking of metallic corrosion, conduct surface analysis, examine trend change, and set up a forecasting model of metallic corrosion speed. According to the classification of ISO 9223 atmospheric corrosion of 1‐year‐period deterioration factor (time of wetness, chloride, and sulfur dioxide), it was found from the results that all four metals were of C4 ranks. However, for the one‐year period exposure test of metal, carbon steel, copper, and aluminum can reach the C5 rank. In the trend change of 2‐year metallic corrosion, for all four metals, a higher corrosion rate can be seen in the initial stage; for aluminum and zinc, a stable corrosion rate can be seen in the second month. For carbon steel and copper, a stable corrosion rate can be seen in the sixth month. In the process of setting up an artificial neural forecasting model, carbon steel and copper showed good forecasting capabilities in three models. The ANN model was suitable for zinc, while the PCA‐ANN model was suitable for aluminum.

Abstract-This study employed artificial neural network (ANN) to develop a regional forecasting model to predict atmospheric corrosion rates of carbon steels within general industrial zones and coastal industrial zones in Taiwan