Design and operation of a Francis Turbine with Variable Speed Capabilities
MetadataShow full item record
Statkraft recently decided to develop 1000 MW of wind energy in the middle of Norway. Wind power is an unreliable energy source. Hydropower is therefore needed to balance energy production supplied to the grid. However, it is limited how much hydropower production can vary because the turbines are designed for a certain operating range. Currently, the most limited variable in hydropower is synchronous speed. It is believed that variable speed operation can yield higher overall efficiencies in a larger operating range. The objective of this thesis is to optimize a design for a runner intended for variable speed operation for a particular hydropower plant and perform simulations to determine the hill chart for this runner. The possible efficiency gain with variable speed operation is to be determined. This thesis presents simulations of a design similar to the Tokke prototype turbine, located in Telemark County. Then the original design is altered to make it more suitable for variable speed operation. A design that resembles the Tokke prototype and model have been created and simulated in ANSYS CFX. This design is referred to as design 1 in the thesis. The hill diagram obtained from this design has been compared with the experimental hill diagram of the model turbine in the laboratory. The two hill diagrams do not correlate. However, for n =375 when efficiency is plotted against power output, the shape of the two curves are quite similar. The shape of the curve also correlates well with historical data from the Tokke prototype. For n =375 the average deviation of efficiency between design 1 and experimental model is approximately 2.25%. It seems like neither the model nor the design 1 would have an efficiency gain if the speed could have been adjusted. From the reviewed literature it comes forth that pump turbines show a higher efficiency increase when operated at variable speed. Pump turbines are designed with higher values of u1, than regular Francis turbines. A second simulation design was made, design 2, where u1 were changed from 0.72 to 0.80. u1 was the only parameter changed from design 1. This change leads to a different runner geometry and inlet dimensions. The comparison between design 1 and 2 show that design 1 has higher efficiencies in the normal operating ranges, as well as high load operational areas. Design 2 gives higher efficiencies in part load operational areas. If the speed could have been adjusted, design 2 shows a maximum efficiency increase of 1.2%. Due to several simplifications both in the design stage, numerical setup and the choice of steady state solver the numerical results cannot be trusted. Thus, design and numerical setup must be further improved and validated with experimental results. However, the results in this thesis show a promising possible trend for variable speed operation.