Composite approach to modelling of extreme hydraulic flows

The design of hydraulic infrastructure such as reservoir spillways, weirs and other critical structures is an area of significant importance from a safety and cost perspective. There is a large programme of work to upgrade existing infrastructure to ensure they are resilient in the face of changing climate and extreme weather events. Physical models have been used for many years as the prime approach to predict flow behaviour at full-scale, however in recent times CFD free-surface models have been used and present strong potential to become a robust tool to aid in the design of these complex hydraulic structures. To date CFD models have only been validated in limited number of geometries and flow conditions, and there is a need to demonstrate the reliability and applicability of such models. Industry require validation cases as exemplars to build an evidence base for their applicability (and limitations) for different application. Areas of significant challenge to existing computational models include: reliably including the effects of air entrainment (that can significantly impact water depths and wall pressures); along with uncertainty around appropriate and efficient ways to consider turbulence in these large transient simulations -such that there is the ability to predict key parameters (depths, velocities, wave profiles) within required tolerances.

For the design of a physical model there is the need to establish similarity between the model and the prototype. A physical model is identical to the prototype if geometric, kinematic and dynamic similitudes between prototype and model are accomplished (mechanical similarity). This is not possible to achieve with a scaled physical hydraulic model, and hence the most relevant force ratio is selected and matched in the prototype and model. In hydraulic free surface flows, gravity effects are highly relevant and hence the Froude number similarity is chosen. This leads to scale effects due to other force ratios potentially having discrepancies between model and prototype. This results in turbulence effects and air entrainment in the physical model being under-predicted.  Recent work at Leeds has demonstrated that using physical models alongside CFD (a form of Composite modelling) can allow the uncertainties in the physical model to be quantified through use of CFD models at both physical and prototype scale.
This PhD work will look to understand how the composite approach can be developed, ensuring that there is improved understanding of predictive capability of the numerical models and extending/updating the models as required to appropriately consider the key physics (e.g. air entrainment etc.)