Locating Infiltration & Inflow with Pumping Station Data

Locating Infiltration & Inflow with Pumping Station Data

If the water level in the sewers regularly exceeds capacity, the first assumed solution is often to increase capacity. This is done with bigger wastewater treatment plants, which are worth hundreds of millions of euros. If the capacity is not increased, wastewater treatment plants are constantly working at full capacity, requiring more energy and other resources. Inflow and infiltration dilutes wastewater, making the treatment processes more complicated. At times, the treatment process must be bypassed entirely, forcing the utility to discharge untreated wastewater into the environment.

Where does the excess water in the network come from? Water might end up in the network from lakes, seas, or groundwater areas. Other possible sources include unauthorized connections and heavy rain. A rough estimate of inflow and infiltration can be calculated by comparing the amount of sold water to the amount of treated wastewater. That does not, however, help with locating the vulnerable parts of the network. A careful analysis of network measurements could provide some answers.

Data analytics can also take other factors into account: network measurements can be combined with weather information, electricity consumption, or even sea levels. When months’ or years’ worth of data is analyzed, long-term trends become visible. Last year, the Helsinki Region Environmental Services Authority HSY noticed that the water flow of certain pumping stations correlated with the changes in the sea level. A quick assessment proved that seawater was flowing into the pumping stations at certain times. During the following months, the pumping stations were fixed and the waterflow returned to normal. The changes were noticed at the treatment plant as well, where the high levels of saltwater had occasionally caused difficulties.

Renovation projects are often prioritized by the age of the pipes. Data analysis provides ways to locate parts of the network that are not necessarily the oldest, but are still worth fixing first. Successful prioritizing builds resilience and saves money. As the example above shows, it is very important to have data about the functions of each pumping station: how has the water flow developed over the years and what does it look like under normal conditions? With the help of machine learning, it is possible to create a model using past measurements and other data sources, resulting in reliable information about each pumping station’s normal activity.

The network measurements collected by water utilities are a significant data asset that makes knowledge-based decisions possible. Infiltration and inflow cause problems for water utilities around the world. With climate change, population growth, and aging infrastructure, related problems are likely to only increase in the future. The resilience of water infrastructure is best protected by locating and fixing any vulnerable parts of the network, and the first step is to begin the systematic collection of network measurements.