Stationary and transient numerical simulation of a bulb turbine

In this paper, the commercial 3D Navier-Stokes CFD-solver Ansys CFX was used to investigate the flow through a horizontal shaft bulb turbine in a stationary and transient way. The draft tube is one of the most critical parts for the performance of bulb turbines and highly influences the efficiency of the whole configuration with a level of several percent of the total efficiency. A calculation of the existing situation was the basis for a modification and adaptation of the draft tube geometry to realize a shorter machine construction and also to compare the results with hydraulic model tests from a closed loop turbine test stand. For all influences on the draft tube rotor stator effects are decisive - in reality as well as in numerical simulations. In a first part, several modifications were analyzed, e.g. inlet variations, as those cause different space requirements of the generator as far as the flow rate is concerned, the operating points differ more than 300%. For the flow through the runner the meridional velocity was analyzed on 5 different planes around the runner. In front and also after the runner the flow over the radius was nearly on an equal level - this means that the hydraulic is good and that no fluid transport in radial direction is needed. In a second part, an investigation of different influences of rotor stator interactions was carried out, based on a large full 360 degrees-model of the whole machine configuration of this 3-blade runner and 16 guide vane turbine. The simulations were done after tests with different grids, turbulence models, grid interfaces and settings. To compare the simulation results with experimental test results, the histogram pressure method for the sigma calculation developed by the Astroe-team was used. For the flow through the runner the meridional velocity was analyzed on different planes around the runner. In front and also after the runner the flow over the radius was nearly on an equal level - this means that the hydraulic is good and that no fluid transport in radial direction is needed. After the hub a zone of very low velocities was estimated. For best efficiency in the draft tube a good balance between separation and reattachment of the flow is of importance. For stationary calculations remarkable discrepancies in the draft tube flow distribution are known, which was also proved and analyzed by means of a transient simulation. There could be found separations for a short draft tube. We could verify and check the test data with the simulation for the client and give him a good impression of the complex physics of the draft tube flow and also of the other components of the machine. A further optimization of the whole geometry and more complex transient simulations are possible, as well as an adaptation of the turbine for different operation points and installation schemes. The model runner diameter was D=340 mm with a specific speed of approx. n(q)=210 min(-1) near the best point for the 3-blade runner configuration with 16 guide vanes.