71 - Framework for Predicting Water Main Breaks in the Face of Climate Change

Water distribution systems are a vital element of sustainable communities. As these systems age, and their operation and environment changes, risk of failure can increase. Such failures result in significant problems, including water loss, sanitation issues, damages to other infrastructure, and service interruptions. Climate change can exacerbate the risk of water main failure, as various weather parameters and patterns are altered, including precipitation, as well as warm and cold spells. Through a comprehensive review of existing literature on water main failure prediction, this research underscores the pivotal role of climate-related factors such as temperature variations, rainfall deficit, and freezing index. This study aims to develop a robust predictive framework for water main breaks, accounting for climate change. Deep learning algorithms, particularly Long Short-Term Memory (LSTM) networks are applied. Different types of data comprising water main inventory and break records, along with climate data encompassing temperature and rainfall records are utilized. To assess the impact of climate factors on water main failure, climate-related covariates, e.g. freezing degree days, are also considered. The sensitivity of results to various climate change scenarios is also explored. The methodology is validated with a case study of Saskatoon, Canada, serving as a real-world testbed for evaluating its effectiveness. Results of this study will provide insights for decision-makers regarding the effective management of water distribution networks in the face of climate change.