Harnessing AI for Cleaner Water: A Game-Changing Approach to Water Quality Forecasting

Water utilities face a constant challenge: ensuring the quality of raw water supplies amidst fluctuating environmental conditions. Factors like weather patterns, upstream pollution, and seasonal changes can dramatically affect water quality, making it difficult to anticipate and address issues before they impact treatment processes or public health. Traditional methods of water quality monitoring often rely on reactive measures, which can be costly and inefficient.

Enter the UTS Data Science team, who have developed AI-driven tools to tackle this problem head-on. By leveraging machine learning, they’ve created a system capable of predicting raw water quality from one day to one month ahead, giving water utilities a powerful tool to stay ahead of potential issues. The team are now optimising a production level software to optimising performance across the fleet of Sydney water treatment plants

How It Works: AI-Powered Precision

The UTS solution integrates real-time monitoring, weather data, and upstream environmental information to forecast water quality with remarkable accuracy. Here’s a breakdown of how it works:

• Real-Time Data Integration: The system pulls in live data from sensors monitoring water sources, capturing variables like turbidity, pH, and contaminant levels.

• Weather and Upstream Inputs: By incorporating weather forecasts and upstream data, the AI can predict how events like heavy rainfall or industrial runoff might affect water quality.

• Machine Learning Models: Advanced algorithms analyze this data to identify patterns and predict future water quality trends, offering forecasts ranging from daily to monthly timelines.

• Actionable Insights: The predictions enable water utilities to optimize treatment processes, allocate resources efficiently, and mitigate risks before they escalate.

This proactive approach marks a significant shift from traditional reactive strategies, allowing utilities to anticipate challenges and act swiftly.

The Impact: Efficiency, Cost Savings, and Sustainability

The benefits of this AI-driven system are far-reaching. For water utilities, the ability to predict water quality issues translates to:

• The Swift model predicts chemical doses for changed raw water quality

• Operators can rapidly respond and dynamically optimise the plant during a water quality event

• Save many Jar tests, manual optimisation and delayed response leading to plant shut down

• Using the Swift model, the Sydney water staff at the Nepean WFP made 35 dose changes in afew days during a rapidly changing water quality event in 2022

• Cost Efficiency: By anticipating water quality changes, utilities can optimize treatment processes, reducing the need for expensive last-minute interventions. Due to rapid ability to adjust dose rates to conditions, chemical usage has dropped some 30-50% across the treatment process. By being able to

• Improved Public Health: Early detection of potential contaminants ensures safer drinking water for communities.

• Environmental Sustainability: Better resource management minimizes waste and energy use, aligning with global sustainability goals.

• Resilience to Climate Change: With weather patterns becoming increasingly unpredictable, AI forecasting helps utilities adapt to changing conditions.

The UTS case study demonstrates that these tools are not just theoretical—they’re already delivering tangible results across the Sydney Water fleet. Water utilities adopting this technology have reported improved operational efficiency and enhanced ability to manage water quality risks.

A Broader Vision: AI in Environmental Management

This project is part of a larger trend of using AI to address environmental challenges. From smarter transport systems to energy-efficient housing, UTS is at the forefront of applying data science to real-world problems. The water quality forecasting tools align with other initiatives, such as those exploring AI in urban studies and environmental management, signaling a growing role for technology in building resilient, sustainable systems.

Moreover, the UTS approach has potential applications beyond water utilities. Similar AI-driven forecasting could be adapted for other environmental management tasks, such as predicting air quality or monitoring ecosystem health. As research continues, collaborations like the PhD opportunity in Explainable AI (XAI) for water management at the University of Sydney highlight the growing interest in scaling these solutions.

The Future of Water Management

The UTS case study is a testament to the transformative power of AI in addressing one of humanity’s most pressing challenges: access to clean water. By combining real-time data, predictive analytics, and machine learning, this technology empowers water utilities to make informed decisions, save costs, and protect public health. As climate change continues to strain our resources, innovations like these will be critical to ensuring a sustainable future.

For water utilities, policymakers, and communities, the message is clear: AI isn’t just a tool for the future—it’s driving change today. To learn more about this groundbreaking work, check out the full case study at UTS Case Studies.