Spatial Nonhomogeneous Poisson Process in Corrosion Management

Abstract

A method to test the assumption of nonhomogeneous Poisson point processes is implemented to analyze corrosion pit patterns. The method is calibrated with three artificially generated patterns and manages to accurately assess whether a pattern distribution is random, regular or clustered. The interevent and the nearest-neighbor statistics are employed to check the method's performance. Three empirical corrosion patterns are studied. The outcome of this investigation suggests that maximum pit depths are generally encountered where pit clusters are detected. This result is in agreement with previous studies.