In the world of water management, artificial intelligence (AI) and digital twins are transforming treatment plants, enabling process optimization and improving sustainability. In this interview, Marta Fraile, a water specialist at Idrica, explores how AI is changing the paradigm in water and wastewater treatment plants. The conversation covered topics such as the maturity of digital twins, the importance of ensuring data accuracy, and the impact of regulations and demands for increased efficiency. This interview provides an in-depth insight into the future that AI holds for water treatment.
Could you tell us about Idrica’s AI-enabled solutions for digital transformation at water and wastewater treatment plants?
Idrica is an international pioneer in water data management, analytics and smart-water solutions. Our AI-enabled solutions are designed to drive the digital transformation of water and wastewater treatment plants, enhancing operational efficiency, sustainability and regulatory compliance through advanced analytics and real-time data integration.
Xylem Vue powered by GoAigua is our flagship product. It is an integrated water cycle management solution that leverages AI to optimize operations in water and wastewater treatment plants. The platform integrates data from various sources, including SCADA systems, IoT sensors and historical records, to create a unified view of plant operations. The platform brings together different algorithms and AI-driven models to simulate and optimize different parts of DWTPs and WWTPs.
Our solutions have brought proven results in the utilities where they have been implemented. These include operational efficiency, enhanced decision-making, regulatory compliance and sustainability, which reduces environmental impact.
How can utilities improve operational performance and plan new investments?
We have developed an open set of functionalities which adapt to the specific needs of each utility, thus enabling them to save on consumable costs, as well as reducing the energy and resources required for plant management.
First, we conduct a detailed study of the entire water treatment process in the plant, identifying possible areas for improvement. Once the critical points have been identified, we deploy machine–learning models that apply artificial intelligence to find the optimal point for treatment operations. This empowers utilities to save costs while improving the quality of treated water.
A relevant case study of the application of AI models to DWTP management is the water supply facility serving the city of Valencia and its metropolitan area. Water quality prediction models were implemented at its most critical points to anticipate events that may disrupt operations. Models for recommending optimal chemical doses have also been developed and deployed, enabling savings in operating consumables, and the generation of specific recommendations focused on preventing deterioration that could imply extra maintenance costs. Results include 18% savings in chemical consumption and a 16% reduction in energy use.
How do you adapt an operational digital twin to the AI era? What are the major opportunities and challenges in this area?
Adapting operational digital twins to the AI era involves enhancing the previously described models with AI-driven capabilities to maximize their effectiveness in real-time decision-making, predictive analytics and process optimization.
Some of the major opportunities that we have seen on this journey are:
- Intelligent operational optimization:
AI algorithms can continuously enhance processes in digital twins, from adjusting chemical dosing in water treatment to optimizing energy use. This results in significant cost savings, improved efficiency and better compliance with regulatory standards.
- Enhanced predictive maintenance:
AI-powered digital twins can predict equipment failures more accurately by analyzing patterns in data that might be missed by traditional methods. This leads to more effective maintenance strategies, reduces downtime and extends the life of critical assets.
- Improved decision-making:
Digital twins can provide actionable insights in real-time thanks to AI, enabling faster and better informed decision-making processes.
Some of the challenges include:
- Variable quality of data from different sources. This is a determining factor for the effectiveness of digital twins.
- Low level of process instrumentalization that creates a weak data model for predicting and recommending actions.
- Complex implementation of the models and a need to ensure they are up to date with the latest features.
- The changes required in organizational culture before the new digital twin systems can be incorporated into the operators’ daily routines.
Where will the focus for innovation be in this area?
The combination of AI and digital twins opens up a world of opportunities for improving water management and optimizing processes. We believe that innovations will focus on several key areas that are set to bring significant progress in operational efficiency, sustainability and scalability. Some of these areas include:
- Predictive analytics and intelligent automation: digital twins will be capable of optimizing operations thanks to smart decision automation based on different scenarios.
- Enhanced process control through AI-augmented simulation: process control will continue to improve by combining AI with real-time simulations within the digital twin.
How can AI enhance digital twin maturity, calculate ROI and ensure data accuracy through performance metrics?
AI plays a critical role in making digital twins more mature by adding intelligence, adaptability, and learning capabilities to these models. As digital twins evolve, AI can contribute to maturity thanks to continuous learning and adaptation as well as autonomous decision-making.
The return on investment (ROI) of a digital twin can be calculated by weighing up the benefits gained from its implementation on one hand and the costs involved on the other. These include the development and integration costs of its software and hardware with legacy systems, and operational costs such as maintaining and updating the digital twin.
The benefits encompass cost savings in maintenance, energy consumption and downtime due to predictive maintenance and process optimization. There are also productivity gains and time savings thanks to process automation and enhanced decision-making. Therefore, ROI should be calculated by quantifying the benefits and costs over a specific period of time. This simple calculation will demonstrate the value of the digital twin.
AI is critical to ensure the accuracy and reliability of the data that drives the digital twin. Specifically, it oversees the processes of data validation and cleaning, anomaly detection and feedback loops. This enables the digital twin to compare predictions with actual outcomes, and to learn from any discrepancies.
There are several metrics for evaluating the data accuracy and performance of a digital twin, such as prediction accuracy, response time, uptime and availability. There are many others that depend on the features to be measured and on the approach to the digital twin’s performance.
What is the expected market penetration of AI in treatment plants over the next few years?
Estimating the market penetration of AI in water treatment plants implies analyzing technological advancements, regulatory drivers and industry forecasts. While exact numbers are very difficult to predict due to a number of different factors, we can provide an estimate based on available data and industry insights.
The adoption level of AI in treatment plants is in its early stages. Some larger utilities in digitally mature regions have started to build AI into their operations. However, widespread adoption in smaller and less technologically advanced utilities remains limited. The estimated penetration rate to date stands at approximately 10-15% of the world’s treatment plants.
y 2025, the penetration of AI in treatment plants is expected to increase significantly due to cost reductions, proven ROI and financial incentives. Therefore, we can say that the estimated global penetration rate by the end of next year is expected to be about 25-30%.
AI is set to be a mainstream technology in the water treatment sector by 2035, due to the technological maturity of digital twins, their widespread benefits, and industry standards. AI is likely to become a mainstream component of WTP operations for ensuring safety and compliance. In the next decade, penetration could reach up to 70-80% of WTPs globally, and most utilities will likely integrate AI into their core operations.
