| dc.description.abstract | Sewer networks are designed to collect and transport stormwater runoff. The capacity of these systems can be exceeded during extreme rainfall events, which can lead to flooding. Computational modelling is used to understand the behaviour and capacity of these networks, and to determine possible flood locations.
Traditional sewer models, thatwhich can be coupled with water quality or catchment hydrological models, are typically computationally expensive,. This which limits their use for real-time modelling during an event. Conceptual models that solve less complex numerical algorithms can be used for faster modelling. However, the conceptual models developed so far have often been less accurate.
In this study, two conceptual sewer simulators have been developed based upon Cellular Automata (CA) principles, which have low computation times in comparison to recognised benchmark models. CA models represent the region being simulated by a grid of cells, and simple rules are used to change the cell states.
These models have been tested using three case studies (one hypothetical and two real world cases). The accuracy was determined in the case studies by performing a visual and statistical analysis of the results. The statistical analysis included measures such as the Root Mean Square Error, the Nash-Sutcliffe Efficiency, and the Index of Agreement. From this and looking at the computation times of the modelsIt it has been demonstrated that these new simulators are both fast and accurate. | en_GB |
