The application of numerical modeling of open channel flows is fairly recent in the course of river engineering history. This can be attributed, in part, to conservatism in the civil engineering design community and the relatively slow uptake of computer applications as a means of problem solving in the industry. Another factor to account for the slow introduction of numerical hydraulics and fluid mechanics in civil engineering is the complexity of the problem and the level of uncertainty in defining boundary conditions.

This resulted in civil engineering taking an initially different route from the other engineering disciplines to enable the implementation of numerical models of river flows. While the other disciplines were able to adopt fluid mechanics treatment of the Navier-Stokes equations in simple small-scale geometries civil engineering initially had to formulate the Navier-Stokes equations in a simpler way that would take into account the variability and complexity of the natural solution domain. Moreover, of initial concern to most civil engineers were the predictions or water levels and discharges to be used for the design of flood protection work, the construction of structures and the assessment of flood risk. These did not require the use of detailed fluid mechanics, but rather the ability to present a large-scale picture of the flow, which the St Venant equations provided. These equations solved in 1- and 2-D still form the basis for large scale river and flood flow modeling. However only 3-D Navier-Stokes model can offer the description level necessary to understand turbulence and complex vortical structures found in flood flows and their consequences on the river system, and model flow around and past hydraulic structures.

Work carried out at the University of Nottingham encompasses all aspects. CFD codes such as ANSYS CFX have been used to consider flow structures in experimental channels such as the Flood Channel Facility (FCF; Morvan, 2001; Morvan et al., 2002; Morvan, 2005) and the Riprap Test Facility (RTF; Wright and Morvan, 2006), but also in natural river systems such as the Rivers Ribble and Severn, for kilometer reach scales (Pender et al., 2005). Ongoing work involves the study of compound by way of LES in order to inform the 3 parameters necessary to implement the large scale 1-D model championed by the UK Environmental Agency (EA).

With urbanization and the increased frequency in extreme flood events, the behavior of flood defenses under extreme loads is an unknown that needs to be investigated. Work carried out by the writer and his team have investigated the risk of embankment failure by overtopping (Morvan and Zelmar, 2006) using detailed CFD. The position of the free surface was predicted but, more importantly, the pressure and shear stress values along the faces of the embankment used to assess the stability of the structure. Ongoing structural hydraulics work involves the computation of hydraulic loads, afflux, erosion risk on bridges when they are subjected to flood flow. CFD could prove a very useful tool to be used for engineering design, and the work carried out here aims to evaluate the level of accuracy of CFD compared to small scale experimental data and the methodology of Delleur and Biery.

Flow over a Weir.
This simulation was carried out for 1:1 existing weir using CFX homogeneous model. The reach is about 600 m long.

Embankment Overtopping (Zelmar, 2005).
This is for a scaled model of a 1:2 earth embankment that is overtopped.