Authors: D. Bonin, S. Campbell, J. Groeneveld
Canadian Dam Association Conference, Winnipeg, September 2008
Abstract
This paper discusses an integral approach of using both physical modelling and Computational
Fluid Dynamics (CFD) modelling that was used in the hydraulic design (spillway, powerhouse
intake, river management) of the Keeyask Generating Station (GS) project. It is argued that
although the potential is undoubtedly great and CFD is thus increasingly being used in the water
resources industry for the assessment of design, screening and optimization of hydraulic
structures, there are still significant areas in fluid mechanics which are poorly understood and need to be addressed through physical modeling. The hydraulic design of a water resources project can sometimes present the designer with unique and difficult challenges. In the past, the designer has usually resorted to the use of physical models to provide guidance on critical design issues that arise. However, with recent advancements in numerical modelling techniques and computing power, designers now have additional tools at their disposal to provide this guidance. For many decades, the standard hydraulic laboratory tool, the physical model, was used to study flow phenomena. Empirical design rules have largely been formulated on the basis of the theoretical understanding aided by experimental data obtained in the field or the laboratory. However, computing power and the simultaneous development of sophisticated numerical techniques have reached the point where it is often claimed (or at least thought) that such 'oldfashioned' and 'surely expensive?' tools like physical models are redundant. Cannot everything now be predicted by the computer?
Whether numerical modelling will ever completely replace physical modelling depends on the
complexity of the situation under consideration. However, it is important to recognize that CFD is not just an alternative method; in fact it brings a wide range of sophistication that can be used to enhance physical modelling, through increased interpretation and understanding.
This paper
addresses how an integral approach using both physical modelling and computational modelling
can be beneficial in addressing the full range of design issues. Comparative results are provided
where available.
