Numerical prediction of hill charts of Francis turbines

This present work compares numerically predicted hill chart to experimental measurements of a Francis turbine. The main objective is to create a model for recreating hill charts using computational fluid dynamics (CFD). Accurate prediction of hill charts are useful in the design stage of production and may result in a more efficient runner. The primary focus is the prediction of efficiency and investigation of possible simplifications without loss in accuracy. By using steady-state simulations, preliminary tests were made on four different meshes, and two different turbulence models, namely the standard k – ε model and the shear stress transport model. Simplifications of geometry have been tested to investigate if the simulation time can be reduced without sacrificing accuracy. Numerical simulations of 132 operating points were carried out. The efficiency was predicted with the maximal difference from measured values of 6.93%. (paper)