Resources > Latest News > From big data to big trees smart approaches to sewer maintenance

From big data to big trees: smart approaches to sewer maintenance

As utilities continue on the path to digitisation, smart data-driven approaches to asset management are finding their way underground, offering up new and exciting solutions to the age-old issue of blocked or damaged sewer pipes.

One company setting the pace in the adoption of smart technologies in the water sector is Veolia.

Quentin Bechet, Smart Water Manager with Veolia, said taking a smart approach to sewer network maintenance can help with a range of expensive and time-consuming challenges.

“When it comes to the sewer network, there are generally two areas of concern for utilities and operators: service problems and structural problems,” he said.

“Service issues are created when sewer pipes get blocked, and this happens very often. There’s also usually an impact on customers in these situations, and there are environmental issues, as well. 

“In terms of more traditional management of sewer networks, most of these blockages are only detected when customers are already impacted and ring to notify the water utility that something’s gone wrong.”

Aside from blockages caused by the infiltration of tree roots and other unwanted items — wet wipes, for example — Bechet said structural issues with pipes can lead to stormwater infiltration into sewer networks.

“When the underground water table rises due to big rain events, it saturates the ground around the pipes. Even though sewer networks are designed to be completely waterproof, structural issues can sometimes result in a lot of water getting into the network,” he said.

“These situations can create huge overflows of sewage and stormwater that can end up on the street or in the environment.

“If wastewater treatment plants are expecting a peak flow rate, and if there’s more water coming through the system due to excessive stormwater infiltrating the network, it can lead to an overflow, where untreated wastewater ends up in the waterway.”

Bechet said water utilities have always needed to try to avoid these problems, even more so today as community concern about the environment grows: “And to minimise blockages, you need to make sure your network is in the best structural condition possible”.

Old pipes, new perspective

Traditional approaches to sewer network management tend to focus on cleaning and maintenance programs. Bechet said that while this will always be required for physical assets, smart approaches can help operators detect small issues before they become big problems.

“Every year, there are pipes targeted for maintenance and cleaning. We work with utility operators to remove tree roots, jet the pipes, or reline sections of pipe if that’s needed,” he said.

Despite this, Bechet said developing smart approaches to sewer management offers utilities and network operators the opportunity to become far more predictive, which is why it’s become a priority for Veolia.

“Smart sewer management is really all about trying to refine our estimation of which pipes are going to get blocked, or need special structural maintenance,” he said.

“Preventing blockages is a key area that adopting smart sewer management practices can help. Usually, cleaning and maintenance programs are designed from human intuition or experience. It’s not the most accurate way to predict upcoming issues.”

New-age insights

One of the smart technology approaches currently being used in sewer network management is the installation of level sensors, Bechet said, which helps operators keep an eye on water volumes within the sewer network.

“We now have lots of different sensors that can be placed in the network to detect issues associated with blockages. By installing level sensors, changes in the volume of water in the network can be detected very easily,” he said.

“If the network is in good shape, you won’t have a massive increase in the water level. We are able to use this information to detect anomalies and signal when there may be infiltration upstream, or a blockage downstream.

“By keeping an eye on network-level anomalies, our operators are then able to inspect and address the causes before they become bigger issues.”

While sensors are one approach to retrieving more information about what’s going on underground, Bechet said Veolia has been working on a big-data algorithm that can predict future blockages without sensors at all.

“At Veolia, we have been looking into a different approach using historical data that already exists. We have developed an AI algorithm that is able to predict the location of the next blockage based on historical big data, from the past decade,” he said.

“Operators will always have a valuable insight into network conditions, based on experience and expertise. But human beings can’t be expected to keep track of thousands of kilometres of mains from memory. This is where AI and big data can be very useful.”

Bechet said the algorithm combines the historical data of blockages in the network, as well as a range of other characteristics, including age, materials and external factors, like nearby trees, roads or soil type.

“With this information, the AI is able to determine the set of conditions or patterns that lead to a blockage and even predict the likelihood of blockage for all the pipes in the network,” he said.

“The operator can then use this information to design a targeted approach to pipe cleaning and maintenance every year, addressing looming issues before they result in sewage overflows, disruptions to the community or problems for the environment.

“The operational benefits speak for themselves, and it's cheaper to send a crew on planned works, rather than sending them to fix a problem that pops up unexpectedly. It also minimises the public health and environmental risks involved in sewage overflows.”

Future focus

While smart approaches to sewer condition management have come a long way in recent years, Bechet said there’s plenty more expected in future as technologies and data analytics continue to progress.

“One new approach to sewer management that’s almost mature now is the use of satellite imagery to assess potential structural anomalies,” he said.

“One of the main reasons for blockages is tree roots. If you have a crack in a pipe, and there are trees in the area, you can be sure the tree’s roots will seek out the abundant supply of water and nutrients that are leaking from the pipe underground.

“As soon as they have access to these supplies, their roots grow much bigger. If a tree has its roots in a sewer pipe, it's potentially going to grow much faster. This is something Veolia’s technology can track by analysing satellite images, which is a simple but really cool new way of approaching an old problem.”

Using images available for free from high-resolution satellites that fly over Australia every three to five days, Bechet said the approach essentially uses the trees as anomaly sensors.

“By monitoring the growth of trees, the trees themselves become sensors of sewer network issues. You certainly have to run a few algorithms, it’s not something you can easily track by eye, but we can now definitely monitor tree growth around any area of the network,” he said.

“Veolia completed a trial with our partner D-CAT last year, which showed that the approach works pretty well. It’s a good indicator of tree root issues and it’s relatively inexpensive compared to sensor-based approaches, because trees are your sensors!”

Aside from using satellite-generated data, Bechet said drones are becoming a big focus in all areas of the sewer operations, too, with hopes that smart drones will soon be able to manage large-scale network inspections.

“This idea still seems pretty sci-fi right now, but the use of drones has the potential to help generate huge amounts of conditional data very easily,” he said. 

“These days we have operators inspecting pipes with remote-controlled robots. But we are also starting to see efforts toward AI algorithms capable of analysing pipes in real-time.

“If drones have the ability to collect and analyse the information as they go, automating what usually requires an operator, we’ll be able to inspect networks much more quickly and at a significantly lower cost. That means we can gain much more insight into their condition, which will help with proactive maintenance strategies as well.

"We think this new deployment of machine learning and digital tools to water networks is equivalent to when the global internet was developed, it changed communication for good.

“We think these digital water tools create a real-life internet of assets which can communicate at the edge, in real-time, optimising performance and delivering game-changing data streams."