A Novel Riemann Solver for Modelling Partial Barriers to Flow within 2-D Hydrodynamic Models

Flood modelling is a vital component of flood risk management practice, as the results produced by flood models are used to inform future investment. Accurate flood modelling is therefore vital to effective flood risk management. Within catchments, linear features, such as bridge structures, can act as partial barriers to flow, significantly influencing the behaviour of flood flows. However, despite the established importance of accurately modelling flood flows for effective flood risk management, no flood model currently has the capacity to explicitly model the interaction between flood flows and partial barriers to flow. Within contemporary industry practice such features are often modelled using coarse approximations, empirically based methods or by omitting such features entirely. The development of a novel solver, capable of explicitly modelling the interaction between flood flows and partial barriers to flow is therefore proposed. Development of such a solver would promise more effective flood risk management including advances in the structural analysis of bridges, infrastructure resilience modelling and tsunami and dam break scenarios, among others.
Development of the novel solver is to be achieved by first developing a benchmark model within a suitable programming language, using existing solvers utilised by industry standard flood models. Following this, experiments within a state-of-the-art multipurpose flume are to be conducted to study the physical behaviour and educate the development of the novel solver. The development of the solver is to be conducted in a somewhat iterative manner: by comparing the predictions of the numerical model and the observed phenomena recorded during the experimentation process, an understanding of the physically logical reasons for the observed differences can be developed and consequently necessary changes to the solver can be identified and implemented. Once sufficient accuracy has been achieved and the code has been optimised, the experimental scenarios conducted within the multi-purpose flume are to be replicated using the novel solver, benchmark model and open-source 3D computational fluid dynamics analysis software and the results compared. Through comparison with the experimental results the predictive capacity of the respective models can be validated and via comparison between the different models the relative accuracy and value of the novel solver can be determined.
The objectives for the study can therefore be summarised as:
1. Develop a benchmark model using existing solvers.
2. Conduct experiments within a state-of-the-art multipurpose flume.
3. Develop and implement a novel solver capable of explicitly modelling the interaction between flood flows and partial barriers to flow.
4. Validate the predictive capacity of the novel solver against experimental data.
5. Compare the results produced by the novel solver to other commercially available solutions.