RIWWER

Reduction of the Impact of untreated WasteWater on the Environment in case of torrential RainReduction of the Impact of untreated WasteWater on the Environment in case of torrential Rain

RIWWER is developing an intelligent, automated and resilient measurement and control system for municipal wastewater treatment plants which is based on cloud/edge technologies. With the aid of this system, water volumes can be intelligently controlled, especially during heavy rainfall. The system distributes water in the sewer system and rain retention basins in a manner that prevents or minimises overflows of wastewater into watercourses or bodies of water. RIWWER is thus establishing a reference project for smart city applications that responds to the changing weather conditions caused by climate change. The results of the research project are expected to be transferable to similar projects. The project is the first to integrate different AI and edge methods in a water management application that improves the management of wastewater systems while minimising pollutant discharges into watercourses.

Market perspectives and product claims
Wastewater management, including discharge and treatment, is a key responsibility of cities and municipalities which is mainly carried out by owner-operated municipal enterprises. RIWWER is therefore developing a solution, primarily for the municipal capital goods market in Germany, which will allow municipalities to adapt their wastewater systems to the effects of climate change while simultaneously improving the efficiency of their wastewater management. Both measures will help stabilise wastewater charges in the long term. At present, the maintenance costs of wastewater systems temporarily increase following damaging events, such as heavy rainfall or flooding, in order to compensate for the incurred damage. With RIWWER, the adverse effects of heavy rainfall on humans and the environment will be avoided, or at least mitigated.
RIWWER provides municipalities with the following competitive advantages: Flooding, leading to flooded basements, is minimised, along with the associated consequences such as interruptions to supply chains and production downtimes. Improved water pollution control is also ensured, since fewer pollutants are released into rivers and canals. In the event of flood disasters, companies are better protected from being compelled to discharge untreated wastewater into watercourses due to heavy rainfall and flooding, thus avoiding potential liability for this action. Furthermore, RIWWER is a smart city application that cities and municipalities can integrate efficiently into their smart city strategy. By leveraging digitalisation and sustainability in this way, they can differentiate themselves in competition with other municipalities.

Challenge and innovation
In Germany, untreated wastewater must not be discharged into rivers and lakes. Wastewater from households, commerce, and industry is therefore directed to sewage treatment plants through the sewer system. In addition to these five billion cubic metres of wastewater, there are around three billion cubic metres of rainfall every year that cannot drain away through roads and surfaces. Leaks in the sewer network also allow considerable amounts of infiltration water to enter the sewage system. During periods of flooding or heavy rainfall, mixed water from rain and wastewater is collected in rain retention basins so that it can later be directed to sewage treatment plants. If these basins overflow due to their incapacity to contain the water volumes, untreated wastewater is discharged into watercourses. This leads to hydraulic stress in rivers and lakes, i.e. damage to the typical aquatic communities, disruption to the natural water flow, an increase in the concentration of pollutants, and oxygen depletion.

Extreme weather events such as heavy rainfall and flooding are expected to increase in number as a result of climate change. RIWWER is taking the first step to address this challenge by digitising the wastewater system. This involves fitting IT components to key points of the sewer system and stormwater tanks. The municipal wastewater system is also equipped with additional sensors and, if required, actuators, enabling online monitoring and digital analysis of key process data as well as control of wastewater flows. This capability allows water levels in sewers and overflow basins to be detected in real time and linked to weather data.

AI models can be used in this way to calculate precise forecasts of water levels over the course of weather events. This capability, in turn, facilitates intelligent control of the wastewater system. By strategically opening and closing sluices and valves, excess water can be directed into underutilised areas of the wastewater system. This control is made possible by an edge AI system currently being developed as part of the RIWWER project. The system allows for training of AI models based on sensor data at the cloud level and also enables decentralised control of the wastewater system through a sensor actuator system at the edge level.

Use cases
RIWWER’s digital edge AI wastewater system is to be tested and evaluated in sewers and wastewater treatment plants at a demonstration plant in Duisburg-Vierlinden and Duisburg-Hochfeld. The proof of concept is designed to allow the system to be transferred to other sewer systems. The intelligent wastewater demonstration system comprises three levels:
At the top level, there is a cloud-based control centre. This control centre is connected to the subsystems. System processes are mapped at this level. The system processes also form the training basis for the edge level, which is arranged below them.
At the middle level, there are local subsystems at central and special water management structures. These subsystems consist of several sensors that provide data on water levels and water quality in sewer networks and overflow basins. The subsystem is wirelessly connected to the control centre. At the lowest level (field level), there are individual sensors capable of recommending a change in system behaviour through control optimisation. These sensors measure the water level and precipitation and are wirelessly connected to the local subsystem.
The distinctive feature of this approach is ensuring that operation is maintained even in the event of a communication link failure. This is achieved by transferring the AI models trained in the cloud to the edge level and constantly updating them. If the connection fails, the edge AI continues to control the sub-area using the latest forecast data and local data.

Consortium
Fraunhofer IMS (consortium leader), University of Duisburg-Essen (UDE), The Association of German Engineers (VDI), KROHNE, Okeanos, HST Systemtechnik, RWTH Aachen University, Berlin University of Applied Sciences and Technology (BHT)

Benefits