A network of low-cost temperature sensors for real-time monitoring of combined sewer overflow

The H2020 innovation project digital-water.city (DWC) aims at boosting the integrated management of
water systems in five major European cities – Berlin, Copenhagen, Milan, Paris and Sofia – by leveraging
the potential of data and digital technologies. The goal is to quantify the benefits of a panel of 15
innovative digital solutions and achieve their long-term uptake and successful integration in the existing
digital systems and governance processes. One of these promising technologies is a new generation of
sensors for measuring combined sewer overflow occurrence, developed by ICRA and IoTsens.
Recent EU regulations have correctly identified CSOs as an important source of contamination and
promote appropriate monitoring of all CSO structures in order to control and avoid the detrimental effects
on receiving waters. Traditionally there has been a lack of reliable data on the occurrence of CSOs, with
the main limitations being: i) the high number of CSO structures per municipality or catchment and ii) the
high cost of the flow-monitoring equipment available on the market to measure CSO events. These two
factors and the technical constraints of accessing and installing monitoring equipment in some CSO
structures have delayed the implementation of extensive monitoring of CSOs. As a result, utilities lack
information about the behaviour of the network and potential impacts on the local water bodies.

The new sensor technology developed by ICRA and IoTsens provides a simple yet robust method for CSO
detection based on the deployment of a network of innovative low-cost temperature sensors. The
technology reduces CAPEX and OPEX for CSO monitoring, compared to classical flow or water level
measurements, and allows utilities to monitor their network extensively. The sensors are installed at the
overflows crest and measure air temperature during dry-weather conditions and water temperature when
the overflow crest is submerged in case of a CSO event. A CSO event and its duration can be detected by
a shift in observed temperature, thanks to the temperature difference between the air and the water
phase. Artificial intelligence algorithms further help to convert the continuous measurements into binary
information on CSO occurrence. The sensors can quantify the CSO occurrence and duration and remotely
provide real-time overflow information through LoRaWAN/2G communication protocols.
The solution is being deployed since October 2020 in the cities of Sofia, Bulgaria, and Berlin, Germany,
with 10 offline sensors installed in each city to improve knowledge on CSO emissions. Further 36 (Sofia)
and 9 (Berlin) online sensors will follow this winter. Besides its main goal of improving knowledge on CSO
emissions, data in Sofia will also be used to identify suspected dry-weather overflows due to blockages. In
Berlin, data will be used to improve the accuracy of an existing hydrodynamic sewer model for resilience
analysis, flood forecasting and efficient investment in stormwater management measures. First results
show a good detection accuracy of CSO events with the offline version of the technology. As
measurements are ongoing and further sensors will be added, an enhanced set of results is going to be
presented at the conference.