TY - BOOK

T1 - A Lagrangean model for pollutant transport in urban hydrology

T2 - Application in pluvial flood events

AU - Sämann, Robert

N1 - Doctoral thesis

PY - 2021

Y1 - 2021

N2 - The quality of drainage water in urban areas affects the health of humans and nature. Accumulated pollutants on surfaces and accidentally emitted hazardous substances are washed off to the drainage system during rain events. In case of very strong rain event, the capacity of the pipe system can be exceeded and complex flow paths between the surface and the drainage system distribute the hazardous substances in the urban catchment. A prediction of potential transport paths is necessary to take actions before and during a pollution spill. Current models are not sufficiently capable of combining the pipe and surface domain for transport of pollutants. This thesis introduces a transport model for solutes in urban drainage for the bi-directional coupled pipe system and the surface, built for use in urban heavy rainfall conditions. The applied Lagrangean particle approach diminishes the effect of numerical dissipation. The so called Random-Walk motion of particles allows to define physical dispersion in the transport. An explicit Euler scheme with an iterative design to allow for large time steps is introduced for the 1D and 2D calculation domain. The particle routing and tracking is aware of boundary collision and sinks. The particle based transport model allows for an easy balancing between fast and accurate calculation. At the same time it is well suited for exact tracking of transport paths, calculation of mass and concentration. The model is implemented in the created software framework “GULLI”. It is designed for very fast calculation of transport paths on large scale domains with low processor and memory usage. Flow fields can be imported from commercial and free hydrodynamic software models. The use is easy due to a graphical user interface. No specialised hardware is required. A forecast of a spill takes less than one minute for a forecast of three hours with 100 000 particles on a standard desktop PC. The design is intended for use of point source spill events during heavy rainfall events. An application for the use of diffusive spread pollutants is also demonstrated. The influence of rain intensity on the wash-off and transport to the outlets of different catchments is analysed. The model can forecast concentration peaks at the outlets that can be used to optimize actively controlled wastewater treatment facilities. For the use in pollution forecast systems during pluvial floods, the selection of precalculated flow fields is required to serve as a basis of the particle transport model. A Nearest-Neighbour selection approach is introduced, which takes the precipitation pattern into account. A database of 944 events is used to search for flow fields that fit to an upcoming rain event. Good results could be achieved with the selection, if the response time of the pipe system is taken into account. The workflow of the methods for flow field prediction and pollutant spread forecast is combined and analysed regarding the total processing time to generate real-time warnings.

AB - The quality of drainage water in urban areas affects the health of humans and nature. Accumulated pollutants on surfaces and accidentally emitted hazardous substances are washed off to the drainage system during rain events. In case of very strong rain event, the capacity of the pipe system can be exceeded and complex flow paths between the surface and the drainage system distribute the hazardous substances in the urban catchment. A prediction of potential transport paths is necessary to take actions before and during a pollution spill. Current models are not sufficiently capable of combining the pipe and surface domain for transport of pollutants. This thesis introduces a transport model for solutes in urban drainage for the bi-directional coupled pipe system and the surface, built for use in urban heavy rainfall conditions. The applied Lagrangean particle approach diminishes the effect of numerical dissipation. The so called Random-Walk motion of particles allows to define physical dispersion in the transport. An explicit Euler scheme with an iterative design to allow for large time steps is introduced for the 1D and 2D calculation domain. The particle routing and tracking is aware of boundary collision and sinks. The particle based transport model allows for an easy balancing between fast and accurate calculation. At the same time it is well suited for exact tracking of transport paths, calculation of mass and concentration. The model is implemented in the created software framework “GULLI”. It is designed for very fast calculation of transport paths on large scale domains with low processor and memory usage. Flow fields can be imported from commercial and free hydrodynamic software models. The use is easy due to a graphical user interface. No specialised hardware is required. A forecast of a spill takes less than one minute for a forecast of three hours with 100 000 particles on a standard desktop PC. The design is intended for use of point source spill events during heavy rainfall events. An application for the use of diffusive spread pollutants is also demonstrated. The influence of rain intensity on the wash-off and transport to the outlets of different catchments is analysed. The model can forecast concentration peaks at the outlets that can be used to optimize actively controlled wastewater treatment facilities. For the use in pollution forecast systems during pluvial floods, the selection of precalculated flow fields is required to serve as a basis of the particle transport model. A Nearest-Neighbour selection approach is introduced, which takes the precipitation pattern into account. A database of 944 events is used to search for flow fields that fit to an upcoming rain event. Good results could be achieved with the selection, if the response time of the pipe system is taken into account. The workflow of the methods for flow field prediction and pollutant spread forecast is combined and analysed regarding the total processing time to generate real-time warnings.

M3 - Doctoral thesis

T3 - Bericht

CY - Hannover

ER -