Data

SUCCESS IN DATA ANALYTICS  
Sydney Water and Data61 collaboration
D Vitanage, C Doolan L Maunsell, B Cameron, F Chen, Y Wang, Z Li
Publication Date (Web): 22 December 2017
DOI: https://doi.org/10.21139/wej.2018.002


A Sydney Water and Data61 collaboration is researching advanced analytics approaches to solve water industry challenges, including water pipe failure prediction, customer segmentation demand analysis, sewer corrosion prediction, optimising water quality, predicting sewer chokes, and prioritising active leakage detection areas, to achieve better outcomes for customers and to deliver world class network performance. Both organisations have partnered to understand complex data sets that can be translated into knowledge. These numbers add insight. They help to see further, understand deeper and see it sooner. Within this partnership we have developed skills on how to think about it, how to use it, and how to value it. This paper outlines how Sydney Water has progressed on predictive analytics to develop capabilities using machine learning to develop tools of value to operations, shareholders and customers. 

The collaborative effort on data analytics in these projects has used machine learning to predict a number of core requirements on pipes or processes for Sydney Water. The focus of the research is to learn from the current operation data and identify previously unknown or unconfirmed relationships. The aim of doing this is to improve the prediction of the required needs. In these projects, integrating current knowledge and expertise with data analytics has demonstrated promising values in predicting asset performance. The six collaborative projects are detailed as below.

  1. Improving the prediction of the likelihood of failure for critical water pipes and small reticulation pipes
    Data61 and Sydney Water have developed an advanced machine-learning technique-based conceptual model for Sydney Water which improves probabilistic prediction of high-risk failures on critical water pipes.
  2. Customer segmentation and demand analysis 
    Better prediction of water demand in the next few years will provide significant value to the price making decision and supply security planning at Sydney Water. In this project, we used datasets available from Sydney Water to discover the consumption patterns from the properties.
  3. Predicting critical factors related to preventing corrosion in concrete sewers 
    Predicting sewer corrosion is a critical task for water utilities worldwide in order to improve efficiency and save costs in chemical dosing, sewer pipe rehabilitation and sensor deployment. We are developing a new and reliable toolkit to enable spatiotemporal estimation of H2S within the sewer network.
  4. Optimising water quality in delivery systems - a case study
    We conducted data analysis to determine optimal network operation strategies, including chemical dosing strategies and operational reservoir protocols, which can in turn help reduce energy use and improve water quality throughout the network. 
  5. Predicting sewer chokes
    We analysed related data to gain a better understanding of the causes of sewer chokes and further developed an analytical tool to predict the likelihood of future chokes. 
  6. Prioritising active leakage detection areas 
    We developed methodologies to improve the efficiency of the active leak detection (ALD) program by segmenting large pressure zones into smaller segments with different leakage behaviours and better prioritisation of zones/segments.

Click here to read the full paper

Please note: You need to login to access this member-only content.
Not a member? You can purchase the paper in our Online Document Library.