Predicting the risk of pitting corrosion initiation of stainless steels using a Markov chain model. Issue 3 (28th September 2017)

• Record Type: Journal Article
• Authors: Brenna, Andrea; Bolzoni, Fabio; Lazzari, Luciano; Ormellese, Marco
• Is Part Of: Materials and corrosion. Volume 69:Issue 3(2018)
• Page Start: 348
• Page End: 357
• Publication Date: 2017-09-28
• Subjects: chloride -- crevice -- Markov chain -- pitting -- stainless steel -- transition probabilities; Materials -- Periodicals; Metals -- Periodicals; Corrosion and anti-corrosives -- Periodicals; 620.1122305
• DOI: 10.1002/maco.201709753
• Languages: English
• ISSNs: 0947-5117

Abstract

Abstract : A probabilistic approach based on Markov chains for the assessment of pitting and crevice corrosion initiation is proposed. A Markov chain is a stochastic process that undergoes transitions from one state to another through a finite number of possible states, until a so‐called "absorbing state" from which the system has no tendency to evolve is attained. The proposed model calculates the probability to have pitting or to maintain a stable passive condition involving a large number of operating parameters related to both metal and environment: steel chemical composition (namely, the PREN index), chlorides content ([Cl − ]), chloride critical threshold ([Cl − ]cr ), temperature, pH, fluid velocity, electrochemical steel potential, E, presence of crevices. Model and algorithms are mainly based on corrosion data collected from literature. Although the proposed model needs experimental confirmation, the simulations here presented seem acceptable, considering the engineering application of stainless steels. Abstract : An approach based on Markov chains for the assessment of localized corrosion is proposed. The model assumes the existence of five states: an initial metastable condition, two final absorbing, and two transient states. Three transition probabilities are considered, according to the possibility to move from a state to another. Metal and environment parameters are introduced in any transition probabilities by ad hoc algorithms.