High flying engineer develops intelligent new system for the water sector

As water scarcity becomes a key issue for humanity, water efficiency is essential. But globally, almost 90 billion litres of water are lost each day – nearly half of all water pumped around the world – through leaks and burst water mains. When critical mains fail, the consequences can be catastrophic, yet preventative replacement is simply financially feasible for utilities. 

For water companies, being able to accurately predict when water pipes are damaged, and in need of repair, is key to protecting this precious natural resource.  

It’s a challenge that UK company Datatecnics set out to tackle, through a recent Knowledge Transfer Partnership (KTP). Working with experts from the University of Exeter’s Centre for Water Systems, the company wanted to develop a machine learning solution that could, more accurately than ever, reveal a water main’s true health and predict when it might fail. 

But was this possible? What causes water mains to burst, and could an intelligent system really learn how to prevent that from happening? 

At the outset, one thing was for sure. A solution would have far-reaching benefits, giving the industry a tool for proactively managing the water network and saving millions of litres of lost water each year.  

To kickstart the project in 2021, the team hired Milad Latifi as the KTP Associate. A highly talented water engineer with a PhD and 12 years’ experience in the industry, Milad brought his in-depth knowledge to the project, acting as the bridge between Exeter’s academic experts and the company. 

Over the three-year project, Milad worked with Professors Akbar Javadi and Raziyeh Farmani to develop the advanced machine learning tool. To start with, this involved creating a physical model of a water pipe in their Exeter laboratory. Here, they ran several experiments to replicate the different strains and pressures that water pipes will experience in the real world, using sensors to collect information about the different scenarios that would cause a pipe to break. 

Because physical experiments like this are incredibly costly, the team used this core data to develop a machine learning model – known as Finite Element Model (FEA) – that could run through thousands of different scenarios.  

Then, they used the scenario results to shape a self-learning machine learning tool, capable of predicting the soil-pipe behaviour under various conditions. This is where the new tool came into its own. By churning through these scenarios, the model soon learnt to predict exactly where pipes would break – in any given scenario – with an unprecedented level of accuracy. 

“We have created a best-in-class technology for monitoring pipes. No other company in the in the world offers this kind of technology”

Dr Milad Latifi, KTP Associate

These outstanding results have allowed Datatecnics to take the service to market, with water companies in the UK and mainland Europe now using their AI models and sensors to monitor their water networks. As Milad says, this early success is down to one thing: “We have created a best-in-class technology for monitoring pipes. No other company in the in the world offers this kind of technology.” 

Such progress brings big benefits. The company have experienced rapid growth in their revenues and workforce over the course of the KTP project. As the tool reaches new markets and contractors, the company expect to grow further still. 

“The KTP partnership has been critical for product development at Datatecnics, fostering deep knowledge, ambition, and close collaboration.”

Suhayl Zulfiquar, CEO, Datatecnics

Here’s what Datatecnics’ CEO, Suhayl Zulfiquar, had to say about the impact of the KTP: “The KTP partnership has been critical for product development at Datatecnics. Through the deep knowledge, ambition, and close collaboration fostered in the KTP, we have been able to build solutions that give water utilities an unprecedented level of prediction accuracy and modelling capability for their critical pipelines.” 

The innovation has also strengthened the UK’s knowledge economy, with the University of Exeter team publishing four peer-reviewed journal articles about the model. Following all this success, Dr Milad Latifi won the ‘Bright Future Award: Early Career Research Impact’ for his work on the project, at the Knowledge Exchange Awards 2024.

But the biggest impact will be felt by consumers and communities in the years to come. As more water companies become efficient at predicting bursts before they happen, they will save millions – if not billions – of litres of water each year.  

At the 2024 Knowledge Exchange Awards, Dr Milad Latifi won the ‘Bright Future Award: Early Career Research Impact’ for his work on the KTP project.

More Information

This KTP project was funded by Innovate UK to support business-led innovation and access to university expertise. The KTP Associate was Milad Latifi. The University of Exeter’s academic supervisors were Prof Raziyeh Farmani, Associate Professor of Water Engineering, and Prof Akbar Javadi, Professor of Geotechnical Engineering.  

Headquartered in Media City, Manchester, Datatecnics builds prediction models and visualisation tools for clean and wastewater networks, helping utilities dramatically reduce water loss. 

Knowledge Transfer Partnerships (KTPs) with the University of Exeter give organisations the opportunity to access new expertise, develop new products and services, and unlock industry-leading solutions. Our KTP team have delivered over 100 successful projects to date, and were voted National KTP Award winners for ‘Business Impact and Transformation’ and ‘Technical Excellence’ in 2022.  Learn more here.

Written by Ben Dickenson Bampton