Predictive Analytics for Sewer Corrosion

November 4th, 2016


Sulphide induced corrosion of reinforced concrete sewers is a significant problem in waste water systems worldwide, particularly in countries with a warm climate like Australia. Data61 collaborates with water utilities to model sewer corrosion, by focusing on data-driven approaches. The work intends to help water utilities to reduce the uncertainty of corrosion factors (e.g. H2S concentration) and pipe condition over time, leading to significant cost saving (see Figure 1).


This project is performed in following phases:

  • Incorporating prior knowledge about corrosion processes, recorded historical corrosion and observed environmental data to develop a methodology to assign areas with different corrosion risk levels (e.g. low, medium, and high) to the sewer network.
  • Incorporating microbial activity data besides common used factors i.e. temperature, humidity and H2S concentration, to reduce the uncertainty in corrosion prediction.
  • Specifying the type and location of extra sensors to be added to the sewer network to reduce the uncertainty in corrosion prediction.


The benefit of this project includes but not limited to:

  • Data-driven/nonparametric approach
  • Dynamic modelling of corrosion factors
  • Flexible data input/incomplete data
  • Adaptive maintenance with decision support
  • Operational cost saving.


Collaboraters Logos- Sydney Water , data61 , University of New Castle, UTS, Melbourne Water, Water Corporation and SA Water