Comparing Modelling and Experiments for Prediction of Atmospheric Corrosion Under Controlled Dynamic Thin Film and Droplet Electrolytes

Abstract

Corrosion prediction by multiscale modelling aims to achieve the required paradigm shift in the lifetime prediction of metals. Modelling of atmospheric corrosion, for example, based on Finite Element Modelling (FEM) requires an exact description of the boundary region between metal and electrolyte. This requires a radically different research approach for the electrolyte in contact with the metal surface. The critical issue is that most corrosion experiments in fundamental research are performed “in solution” under the assumption that the electrolyte layer is thick enough to approximate it as “infinite”, ignoring significant local effects introduced by dynamic electrolyte dimensions caused by droplets and puddles. In other words, a major part of the research in academia is not relevant to mimic the actual conditions encountered on the metal surface. To create accurate numerical models, we first and foremost needed to abandon this "infinite" and “stationary” approximation and consider dynamic and finite electrolyte dimensions. During atmospheric corrosion of metals, the evolution of the thin-film electrolyte thickness is the primary factor determining the corrosion rate. A full understanding the evolution of the electrolyte thickness and properties as a function of the key environmental parameters is yet to be achieved. A novel methodology has been developed to conduct experiments under a regulated environment to measure the film thickness evolution or individual droplets on an undisturbed metal surface [1]. Experimental results are compared with FEM models predicting both the electrolyte formation as well as the related corrosion. The heat transfer coefficient is noted as a critical parameter influencing both the film characteristics and the resulting corrosion rate. This model was tested in case of cut-edge corrosion of galvanized steel. Combining the in-house MiTReM (Multi ion Transfer Reaction electrochemical model) with the electrolyte film thickness model, we could see that the corrosion rates become very critical in the range of 20 µm [2]. The next step was to go into the dimension of droplets. A thorough literature review [3] was made and published around modelling of droplet formation and corrosion, forming a base for further progress and improvements. The literature study also highlighted that the topic of droplet formation is relatively new within the corrosion research community. A new approach [4] was proposed to numerically predict and study atmospheric corrosion under droplet size distributions. The proposed methodology allows for a corrosion prediction based on observed droplet size distributions and droplet contact angles. A mechanistic finite element model, including oxygen transport and Butler-Volmer kinetics, is solved to obtain the current density for the droplet geometry. This is done for a range of both droplet radii and contact angles. The computed corrosion current densities are then adapted onto imposed droplet distributions for a calculation of overall corrosion current. This allows for a calculated material loss estimation for different distributions and electrolyte configurations and shows the extent of the impact of the droplet distribution on atmospheric corrosion. This model is validated by comparing with experiments introducing droplet-induced corrosion in a lab scale reactor that allows monitoring in parallel droplet geometry and corrosion as function of induced thermal cycles in a controlled atmosphere. Keywords: Atmospheric corrosion, FEM modelling, thin electrolyte films, droplets