Best Practices for Predicting Water Main Breaks
This research program is a collaboration with the National Water and Wastewater Benchmarking Initiative (NWWBI). Fourteen NWWBI participating utilities have shared their water main inventories and historical records of main breaks. By sharing data, models, and experiences across Canadian cities, the project seeks to facilitate the broader application of water main break predictive models.
Framework for Water Main Break Data Collection
A variety of factors can affect pipe breakage and previous research has focused on different subsets of data. This project seeks to analyse data currently collected by Canadian water utilities in order to identify top factors contributing to main breaks and develop a framework for collecting pipe data.
Machine Learning Models for Predicting Water Main Deterioration
This project compares the most powerful machine learning algorithms and their accuracy in predicting rate of failure, remaining useful life and probability of failure for water mains. Algorithms include Artificial Neural Networks, Tree-Based algorithms (e.g. Random Forest, XGBOOST, Gradient Boosting), and Support Vector Machines.
Impact of Weather on Water Main Breaks
This project focuses on the impact of weather attributes on main breaks, specifically daily temperature and precipitation. Historical main break and weather data for 4 Canadian cities is being analysed.
This project is developing approaches for detecting leaks using machine learning models, available SCADA data and hydraulic models. The methods are being tested on a hypothetical network and dataset. Initial applications of support vector machines have lead to high accuracies in defining a radius of leakage.
Energy Recovery in Water Distribution Systems
This project is developing a method to optimize the location and selection of Pumps as Turbines (PATs) in water distribution systems. It considers the full lifecycle costs of installing and operating the PATs and extends upon available hydraulic models.
Energy Efficiency Metrics
A series of projects have focused on the assessment of water distribution energy efficiency and development of methods to enable the identification and selection of energy efficient solutions.
Energy Conservation in Small-Medium Water Distribution Systems in Ontario
This project is a collaboration with the Canadian Urban Institute (CUI) and the Ontario Clean Water Agency (OCWA). An energy efficiency assessment tool is being developed for water distribution systems. The tool automatically calculates energy metrics for all network compoments and maps them. It is being applied to four pilot systems.
Advancing Energy Efficient Water Services in Toronto
This project was completed in 2016 as a collaboration between the University of Toronto, the Canadian Urban Institute (CUI), and the City of Toronto. A spreadsheet tool was developed to calculate energy metrics and create inputs for mapping. These were applied in identifying and comparing efficient solutions, such as changes to pump operations and installation of VFDs.
Enhancing Decision Support for Sustainable Water Planning with Municipal Data
This project was completed in 2016 as a collaboration between the University of Toronto, the Canadian Urban Institute (CUI). One of the outcomes was the development of energy metrics to assess the energy efficiency of water distribution systems. This was applied to a case study of the City of Toronto.