Leveraging Deep Learning to Understand Climate Change Effects on Urban Water Main Failures

Abstract Submission: The optimal annual replacement of deteriorating water mains is a critical task for water utilities. Pipe failure prediction models serve as essential tools in the strategic planning of rehabilitation efforts for urban water distribution infrastructure. However, climate change introduces new challenges to this process which may significantly increase the risk of water main failures. As climate patterns shift, predicting how these changes may impact water main integrity is crucial for sustainable urban water management. Current failure prediction methods typically overlook the impact of long-term climate change scenarios, limiting their accuracy and applicability. Therefore, there is a pressing need for advanced predictive models that can incorporate both historical data and climate projections to better forecast water main failures. This study aims to develop a novel deep learning approach for predicting the probability of failure for individual pipes within a water distribution system, considering historical trends and the potential impacts of climate change. The research utilizes three key datasets from Saskatoon, Canada: pipe inventory data, historical break records, and weather information. A long short-term memory (LSTM) neural network model is developed to capture the temporal dependencies and non-linear interactions in the data. Three climate change scenarios based on Shared Socioeconomic Pathways (SSPs) including low, medium, and high emissions are used to generate future weather projections up to 2035. The findings will contribute to our understanding of how climate change affects urban water infrastructure and can inform proactive asset management strategies and climate adaptation efforts for water utilities.

Learning Objectives/Expected Outcome (Optional) :