Prediction of concrete corrosion in sewers with hybrid Gaussian processes regression model

Abstract

Concrete corrosion is a major concern for sewer authorities due to the significantly shortened service life, which is governed by the corrosion rate and the corrosion initiation time. This paper proposes a hybrid Gaussian Processes Regression (GPR) model to approach the evolution of the corrosion rate and corrosion initiation time, thereby supporting the calculation of service life of sewers. A major challenge in practice is the limited availability of reliable corrosion data obtained in well-defined sewer environments. To enhance the predictability of the hybrid GPR model, an interpolation technique was implemented to extend the limited dataset. The trained model was able to estimate the corrosion initiation time and corrosion rates very close to those measured in Australian sewers.