Conference Agenda

Certain application of urban drainage models require highly-performant models. Such applications include, inter alia, their use within Digital Twins and within real time control frameworks, in particular when model-based predictive control is envisaged. Also other modelling applications involving the evaluation of long time series and/or of a multitude of scenarios require highly performant models.

The paper presents the application of a procedure for automated sewer model transformation to a hydrological model for the rather complex sewer system of the city of Hannover, Germany, which is the third-longest sewer network in Germany. However, as is also the case in the Hannover system, the system contains significant parts subject to hydrodynamic effects such as surcharge and backwater effects, thus rendering the use of purely hydrological models infeasible. Therefore, as a second step of the simplification procedure, a hydrodynamic model for that part of the drainage system subject to significant hydrodynamic effects is combined, within the same model, with a hydrological model of the remaining parts of the network. As results are encouraging, the implementation of that model as a core element of a Digital Twin of the Hannover system is underway. The automated procedure is applicable also to other networks.

The Digital Twin platform developed for the “ZwillE” research project in Germany serves different purposes, covering a wide range of urban water management tasks of the wastewater utility of Hannover in Germany. Fundamental concepts, such as model integration issues and different use cases with different levels of complexity are tackled within the project. This contribution focuses on a new approach for the fast estimation of maximum water levels for use in relatively flat urban environments. The approach combines pre-calculated flood maps based on design rainfall for fixed return periods with online radar measurements and radar nowcasts, including uncertainty information. As case study, a convective heavy rainfall event in Hannover is evaluated with observed flooding and a high number of fire brigade calls registered. Results of the new approach for fast flooding estimation are evaluated in comparison to fire brigade calls and to a coupled 1D sewer and 2D surface runoff simulation using the model Hystem-Extran/Hystem-Extran2D. The fast estimation shows good results of the maximum water level in comparison to the reference simulation, with absolute differences mostly below 0.1 m or 0.2 m respectively. The approach is suitable for real-time application in the newly developed digital Twin for the city Hannover.

Urban areas are prone to occurrence of pluvial flooding, which has the potential to inflict significant damage on urban infrastructure. At the same time, the real-time prediction of such events remains challenging. Therefore, there is a necessity for the development of advanced predictive models to mitigate potential risks and enhance urban resilience. In the present study, we show a test case for a model chain for predicting urban flooding in the city of Hanover, Germany. The model chain consists of a radar-based rainfall nowcasting model and a data-driven water level prediction model. The training data base for the water level prediction model was generated with a detailed hydrodynamic model. The objective of this study is to evaluate the predictive capabilities of the entire model chain, rather than merely those of its constituent parts. Therefore, we have utilised a historical event from 2021 to evaluate the model chain.

Urban drainage systems (UDSs) are facing increasing challenges due to aging infrastructure and external factors, requiring innovative decision support solutions. Digital Twins (DT) of the real-world systems offer a promising decision-support tool for UDS management. These digital replicas, updated in real-time through sensor data integration, can assist with energy efficiency, resource allocation, and scenario analysis for contingency planning. DTs require up-to-date simulation model, and integration of the sensor data into the model is perceived as a critical component. This research introduces an innovative data assimilation method, utilizing Proportional-Integrative-Derivative (PID) controllers to update UDS model states based on sensor data. Proposed approach, tested with PySWMM on a synthetic dataset, demonstrates the potential for improving UDS performance and reducing uncertainty through continuous updates.

Urban drainage systems (UDS) are crucial for managing stormwater in cities. The failure of UDS can lead to severe flooding, causing damage to infrastructure, property, and environment. Their design and implementation often face significant challenges and in coastal lowland cities, these challenges are compounded by the influence of sea tides. Designing these systems in areas with tidal influence requires careful consideration of factors such as sea level rise, tidal patterns, and the topography, making the use of simplified hydraulic equations, such as Manning, unfeasible, as it does not represent the backwater effect. The complexities of these designs and the use of unreliable methods often result in increased costs and technical difficulties, making it a critical area of urban planning and engineering. This paper aims to present the design process of the urban drainage system in a coastal lowland city using the multilayer quasi-2D model MODCEL. The simulation was developed for a study area located in Maricá, Rio de Janeiro, Brazil. The area has a disordered urban occupation aggravating the challenge of project. The design process with hydrodynamic modelling resulted in a functional drainage network, capable of draining a 10-years storm considering backwater effects of high tide and river stage rise.