Go with the flow

Scientists and engineers concerned with making hydro power turbines more efficient will meet in Sweden next year. Their goal is to determine the best type of turbulence model available for the simulation of flow on elbow-type draft tubes, as well as the errors one can expect from such simulations. With this information it will be possible to eliminate some of the expensive experimental testing that normally takes place during the design or modification of turbine installations.

Participants at the Turbine ‘99 Workshop will share their experiences of the challenge of creating draft tube flow computations. Detailed measurements of mean velocities and turbulence quantities will enable comparisons of the performance of different turbulence models and numerical methods.

Kinetic energy

‘Much of the potential for improvement of a modern low head hydro power installation lies in better draft tubes,’ says Professor Rolf Karlsson of Vattenfall Utveckling in Sweden. Without the draft tube, all the kinetic energy that the water possesses after running through a turbine will be lost. The purpose of the draft tube is to gradually slow down the water flow and convert the kinetic energy to pressure energy. This will lead to increased pressure differences over the turbine, which will improve its performance. However, if the expansion in the draft tube is too rapid, streamline separation will occur and much of the kinetic energy will be transformed into turbulent eddies. Most importantly, the flow in draft tubes is dominated by turbulence, swirl and separation.

The detailed description of turbulence is still proving to be a white spot in fluid dynamics. The exact equations for turbulence flow are known but are too complex to solve in a real engineering problem. Much effort has therefore been spent on developing simplified models of turbulence.

These models are normally calibrated on simple test cases such as pipe flow and flat plate boundary layers. The hope is that the models should possess some degree of universality but it is far from clear that this is this case. It is important that the models are tested on a particular class of problems, in this case the draft tubes, before large scale computations can be trusted for design purposes.

Laser doppler velocimetry (LDV) has been used to obtain data in a scale model (1:11) of a Kaplan turbine at the Älvkarleby laboratory of Vattenfall Utveckling. LDV is an optical method for non-intrusive velocity measurements. The basic principle behind it is to let two laser beams intersect at a well determined point in space. An interference pattern in a volume with a sub-millimetre width will then be formed. When a particle, eg a naturally occurring dust particle or air bubble, crosses the interference pattern it will reflect light that can be detected with a photo-multiplier tube.

Subsequent signal processing will finally yield the velocity of the particle, ie the velocity of the fluid. In the Vattenfall experiment, laser light with different colours was used to measure several velocity components at the same time.

Professor Håkan Gustavsson from Luleå University of Technology in Sweden says that to predict complex phenomena, such as turbulence, swirl and separation, requires state-of-the-art turbulence modelling. ‘However, there is even uncertainty about whether the best models are able to predict the phenomena in draft tube flow.’ Gustavsson says. ‘This is why we are organising the workshop so that we can determine the state-of-the-art,’ he added.

Workshop

The turbine workshop is organised jointly by Luleå University of Technology and Vattenfall Utveckling. ‘When the workshop takes place, it will be ten years since a similar workshop was organised in Switzer-land. The improvements of computational power, turbulence modelling and software sophistication in the interval since then gives us reason to believe that much better predictions of the draft tube flow should be possible,’ speculates Associate Professor Rikard Gebart.

‘We hope to see some of the leading research groups in the world at the workshop and expect that the knowledge about simulation of complex flows in draft tubes will take a significant step forward,’ Professor Gustavsson said.