Italian Utilities using Geospatial AI Analytics

Faced with a high level of water leakage and legislation demanding improvement, two Italian utilities turned to space-based observation to help target their repairs, replacements and resources.

Tackling water waste

It is estimated that more than 40 per cent of all the drinking water produced in Italy is wasted. According to some estimates, leakage in the country is so bad that it costs the nation's consumers around €4 billion a year.

“Italian utility companies have been exploring new ways to predict pipeline failures to direct their limited resources more effectively.”

Recognising the impact on the sustainability of its water network and the significant cost implications, in 2017, the government introduced a resolution (917/2017) obligating suppliers to reduce water leakage and provide a more secure water supply.

In response, Italian utility companies have been exploring new ways to predict pipeline failures to direct their limited resources more effectively and make any repair or replacement programmes more efficient.

Tuscany turns to satellites

Utility company Acquedotto del Fiora SpA (AdF) has embraced geospatial artificial intelligence (AI) in a bid to reduce its pipeline losses. Tuscany’s largest water company covering 55 municipalities, including all 28 in the province of Grosseto and 27 in the province of Siena, the AdF pipeline network stretches some 8,000 km.

"Our network is very extended, but with only low numbers of customers, so it is a problem to find water losses quickly," Alessio Giunti, head of water balance and metering equipment, protection of the water resource, Acquedotto del Fiora S.p.A.

"We needed to analyse our network for leak detection and how we utilise our workforce in the best, most efficient way. We have a big network, and so we wanted to assist with intelligence to help us find water losses quickly and to decrease the time needed to find those failures.”

AdF recognised the potential benefits of an AI-led solution as part of its leak prevention toolbox. As Piero Ferrari, Chief Operating Officer at AdF, explained in a statement: "Our goal is to drive down leakage and intelligently inform our capital improvement programme. Although we're a smaller utility, this [geospatial AI approach] allows us to replicate the operations of much larger utilities, with much larger budgets.”

"Our goal is to drive down leakage and intelligently inform our capital improvement programme. Although we're a smaller utility, this allows us to replicate the operations of much larger utilities."

"Our network is very extended, but with only low numbers of customers, so it is a problem to find water losses quickly. We needed to analyse our network for leak detection and how we utilise our workforce in the best, most efficient way. We have a big network, and so we wanted to assist with intelligence to help us find water losses quickly and to decrease the time needed to find those failures,” adds Giunti.

Working with ISOIL, which partners with UK headquartered company Rezatec, the utility launched an initial deployment of Rezatec’s Pipeline Risk product across some 674 km early this year. The pipeline risk assessment tool combines satellite data with artificial intelligence algorithms to produce risk maps.

“The initial results since January are very interesting. We can see a lot of information about the network but also every characteristic of the network. The algorithm provides a better consideration of losses than before,” says Giunti.

New tools for leaks at HERA

Another Italian company deploying geospatial AI is multi-utility HERA Group SpA. Headquartered in Bologna in the northeast of the country, HERA manages a water distribution network of more than 35,000 km of network and more than 400 water treatment plants.

Keen to address its water losses, HERA began exploring new approaches to find and repair leaks even before the introduction of the government resolution with a tie-in with the University of Bologna.

The collaborative research project was designed to assess the various risk factors associated with pipeline failure. Whereas it is generally thought that failures largely depend on the age of a pipeline, in reality, there are more significant factors associated with pipeline failure.

"We discovered that there are interesting factors that aren't only the diameter or the material of our pipelines, but we realised that other factors like the soil and the temperature had an influence too."

As Maurizia Brunetti, Coordinamento Acquedotto (water supply technical coordination manager) at HERA explains: “We discovered that there are interesting factors that aren’t only the diameter or the material of our pipelines, but we realised that other factors like the soil and the temperature had an influence too. We found we also had a correlation with other factors. For example, the soil is a variable that must be considered, and there is also the groundwater level.”

This programme soon led to working with ISOIL Industria S.p.A and Rezatec to help reduce non-revenue water using water leakage control software based on remote sensing and artificial intelligence.

Identifying the likelihood of failure

The tool identifies the Likelihood of Failure (LOF) of the network but also the Consequence of Failure (COF). This takes into account the number and type of customers affected by a failure as well as any third-party liabilities, for instance.

The combination of both gives a Pipeline Risk Value that allows the utility to target its repair and replacement budget on the top 20 per cent of failure risk, which corresponds to around two thirds of actual pipeline failures.

The assessment of the network began with a retrospective analysis of data from 2016-2018 for 490 km section of HERA’s water system with the objective of finding 70 per cent of failures.

"The predictive algorithm can be used to better orient HERA renewals.”

Following on from the initial pilot programme, HERA signed up to cover 2800 km of its network earlier this year. The service now covers the entire province of Rimini and 400 km of the network in Forlì- Cesena province.

“The predictive algorithm can be used to better orient HERA renewals, which are budgeted at more than more than €30 million per year, or for active leakage detection,” Brunetti notes, adding: “This was interesting for us because we could focus our investments and activities.”

Under pressure to deliver better services

As the results in Italy show, the risk of water pipeline failure is not simply a function of age or the pipeline material. A combination of material, age, soil pH, ground heave and even soil temperature all offer a significant correlation with the risk of failure.

By using machine learning to more accurately identify those areas more likely to fail, utilities like HERA and AdF are able to more effectively target their renewals programmes and deployment of their scarce resources.

All utilities are under pressure to reduce water losses, save money and deliver a better service. Digitally enabled tools such as artificial intelligence are now delivering proven benefits allowing progressive companies to work smarter, not just harder.