Stochastic reduced order models for uncertainty quantification of intergranular corrosion rates

• A reduced-order model for uncertainty quantification in corrosive systems is shown.
• Uncertainty in current density due to randomness in electrode sites is determined.
• The proposed model achieves convergence with far fewer samples than Monte-Carlo.
• Computation of correlation between different electrochemical quantities is demonstrated.

We present a stochastic reduced order model (SROM) approach for quantifying uncertainty in systems undergoing corrosion. A SROM is a simple random element with a small number of samples that approximates the statistics of another target random element. The parameters of a SROM are selected through an optimization problem. SROMs can be used to propagate uncertainty through a mathematical model of a corroding system in the same way as in Monte Carlo methods. We use SROMs to estimate the statistics of corrosion current density, considering randomness in anode–cathode sizes. We compare the performance of SROMs against the more common Monte-Carlo approach.