Recently, 2021 EU Green Capital Lahti has focused its efforts on the issue of inflow and infiltration water in the sewer system. Machine learning became a part of their toolkit, as Lahti Aqua adopted the Neuroflux network management service. It quickly became clear that the service was not only useful for locating extraneous water in the sewer system, but was also a great aid in finding problems at an early stage.
When the Neuroflux service was adopted in Lahti in January of 2021, the main objective was to examine infiltration and inflow water in the sewer and to locate potentially vulnerable areas. The users quickly found that the service became a natural part of their daily routine: using machine learning to keep up with real-time data turned out to be an intuitive way to spot other problems in the sewer as well.
Thoughtful visualization makes data faster to process: “Neuroflux is very approachable compared to many other systems: nobody wants to comb through a hundred rows in an Excel file. The information is much nicer to look at in the form of clear charts,” describes water network engineer Mikael Bragge from Lahti Aqua, who was in charge of the pumping station monitoring for the first year.
Automated Alerts Became Part of the Daily Routine
Bragge explains that opening the Neuroflux tool became a natural part of his daily tasks. “Almost by accident, I realized I was checking the alerts every morning. If something seems amiss, we go look at it in the field.” The alerts are a way to get automatic notifications of measurements that differ from the predicted behavior. To help interpret these alerts, Bragge can rely on many years of experience with the Lahti network.
The alerts have helped spot problems: “We have located blockages in their early stages in places that have previously been difficult to monitor. For example, we have locations with three joining sewer lines, and in these cases, one of those lines was starting to clog. Other systems had not necessarily picked up on the problem, because water had still been flowing from the other two directions. With Neuroflux, we have already located a number of these cases.”
Focusing on Extraneous Water
In the fall of 2021, the water utility and Neuroflux began a project where the wastewater flow at every pumping station was analyzed and divided into components. The components consist of billed water, storm water and base flow, describing the proportions of different types of water in the sewer. The key objective was to find out how extraneous water enters into the sewer network in different areas. The first phase of the project was about assessing the quality of data to ensure reliable results from the analysis.
“There is always that doubt in the back of your mind when it comes to tank level measurement,” Bragge says. However, the data quality analysis proved that in many cases tank level measurements function well in different conditions. The analysis was conducted with every pumping station’s flow measurement in order to assess the reliability of the measurement and where it could be improved.
The main goal in Lahti is to further decrease the portion of inflow and infiltration water in the sewer. “We want to take care of everything as best we can. The more inflow and infiltration there is, the more treatment plants need chemicals and electricity, which results in higher costs.”
By locating and fixing vulnerable spots, they expect to see the results in the utility’s key figures as well. “Inflow and infiltration in the sewer is a challenge: the catchment areas are large and stormwater, for example, can seep in through tiny holes. In the end, the total amount may be quite large.”
In Lahti, measurement data is used in multiple ways to keep track of the network, and it helps prioritize renovation and maintenance needs. Wastewater components offer a new criterion for their work.
More Preventative Maintenance
When asked about the most important feature of the Neuroflux service, Bragge brings up the alerts: “For me, it has been finding blockages at an early stage. They have not had time to develop far enough to affect our customers or to cause an overflow. It has helped us to react preventatively.” Bragge also mentions the wastewater components, pointing out that they are still new to their system. Their goal will be to use it to find areas that need to be investigated further. “There will certainly be a lot of use for it on the wastewater side.”