Probability of stainless steel corrosion : experimental validation of the Markov's chains mathematical model

Stainless steels are widely used in industrial application for their corrosion resistance; nevertheless, in presence of chloride ions, they may be subject of localized corrosion attack, as pitting and crevice, which, once initiated, rapidly propagates and leads to equipment failure. Due to the stochastic nature of pitting phenomenon and, in particular, of its initiation, a statistical approach is mandatory for initiation time prediction and, then, for materials selection in engineering application. In this work is proposed a Markov chain’s mathematical model which assumes that pitting retains no memory of its past history and is characterized by two absorbing conditions (pitting and repassivation) and three transitional states (metastable, metapassive, metapitting); the model takes into account metallurgical and environmental factors, as material composition (PREN), chlorides concentration, pH, temperature, fluidodynamic condition and oxidant power of the system, with the aim to calculate the corrosion probability of stainless steels in different industrial and natural environments. The proposed equations, which relate input parameters and transitional probabilities were, initially, empirical expressions based on engineering knowledge of corrosion behavior of stainless steels; aim of this thesis is to validate and confirm these mathematical expressions: electrochemical tests (potentiostatic, potentiodynamic and linear polarization resistance tests) were carried out and the experimental results, in addition of literature corrosion data, were carefully elaborated in order to propose new revised equations of Markov model.