Abstract:
In the field of hydropower, the designing of different components which is cost effective is a challenging work. For proper design of these components are need to analyse critically as the flow is turbulent as well as three dimensional in behaviour. As there are many components in hydropower plant, but the most important component is turbine which converts hydro energy into mechanical energy. There are many factors which affect the turbine life as well its efficiency so there is need to analyse these factors to increase the efficiency and the life.
As in earlier times model testing was used to test performance of turbines which was costly and time consuming. Large hydro power plants have high initial cost and model testing will also be costly. Computational fluid dynamics (CFD) is an alternative to model testing, it takes less time and low in cost. Performance predicted by CFD is also accurate and very near to model testing if used properly. Cavitation is a big problem in reaction turbines and this causes pitting on various region of turbines which leads to decrease in efficiency of turbines. This problem can be reduced by checking blade geometry, blade material whether suitable or not against problem of cavitation in CFD.
In the present study, the Francis turbine of rated capacity of 255MW is modelled in a CAD software and attempted to do CFD analysis in ANSYS 15 to investigate the effect of the cavitation on hydro turbine before and after the use of coating on turbine runner. Flow analysis of the Francis turbine is done in ANSYS FLUENT. The flow simulation has been carried out in cavitation flow condition to find out a suitable coating to prevent runner from cavitation and zones of runner affected due to cavitation.
Finally it is concluded that the modelled Francis turbine has maximum efficiency of 82.4% at rated discharge of 112.22 m3/s at guide vane opening of 21.6o. The best coating material out of the materials taken is found to be cavitec because of lower volume erosion rate as compared to other materials. It is also concluded that if turbine model is very big then smoothing should be low and mesh should be unstructured and method used should be tetrahedral patch conforming.
