Development of Data Driven Corrosion Environmental Severity Scoring System

Abstract The severity of a given geographic location with respect to its tendency to induce corrosion is dependent on the material being corroded. A diversified severity index could be developed to rank the environmental corrosion severity of locations in a manner less dependent on a single material. Aluminum and steel mass loss coupons and corrosion sensors were exposed for 1 year at 25 locations across the world. The data collected from this exposure was used to define an environmental corrosion severity index (ESI) from the perspective of aluminum mass loss, steel mass loss, and corrosion sensor readings using Principal Component Analysis. The results are presented alongside an analysis of the sensitivity of the ESI to various experimental factors. A methodology for predicting ESI using a subset of the required data sets and the associated uncertainty is also presented.