Electrochemical tomography as a novel non-destructive approach for corrosion rate measurements in localized corrosion

Background

Being able to measure the rate at which corrosion occurs is crucial for research and practice. The most dangerous and also a very common type of corrosion is so-called localized corrosion or galvanic corrosion, i.e. the anodic and cathodic areas are spatially separated. Examples are chloride-induced corrosion of steel in concrete or corrosion in water utility systems. In both examples, a direct visual inspection of the steel is not possible.

Non-destructive electrochemical measurement procedures exist to assess the corrosion state of the steel, without the need to remove concrete or soil. The most standardly applied method, Potential mapping, allows just for a prediction of the probability of corrosion and does not supply any information about the corrosion rate. Other electrochemical measurements that try to assess the corrosion rate, rely on the so-called polarization resistance. This is based on the mixed-potential theory by Wagner and Traud, which, however, is not applicable to localized (galvanic) corrosion [1]. This is one of the numerous reasons for the large uncertainties associated with common methods for the measurement of corrosion rates in localized corrosion. 

Aims and objectives

This project seeks to finally deliver a fundamentally sound method for measuring corrosion rates in localized corrosion, and thus to overcome the limitations of the commonly applied non-destructive electrochemical methods. We develop a novel approach, termed electrochemical tomography (ECT), that is to use inverse methods to estimate corrosion locations and rates from temporospatial potential measurements at the surface of the electrolyte upon polarization. 

Methodology

A numerical approach is used to model the corrosion situation at the metal surface and the resulting electrical potential field throughout the electrolyte. Potential measurements at the surface are used to invert for the location and rate of the corrosion. The key innovation is to explore data recorded both over time and at different points in space during external polarization of the galvanic cell.

Both stochastic (Bayesian) and deterministic inverse methods will be used, to first of all explore the inverse model itself and secondly optimize the inversion in terms of speed. Validation in the lab will allow us to identify the main corrosion processes involved, to test the accuracy of our model and to further optimize the method.