Development of an integrated design system (IDS) for hydraulic machinery

An integrated computer-aided system for the initial design and optimisation of hydraulic machinery components, i.e. bladings of centrifugal pumps and fans as well as of water turbines, is presented. The IDS developed at the Institute of Fluidmechanics consists of an artificial neural network (ANN) for the initial design of components, a finite volume CFD-code based on the Reynolds-averaged Navier-Stokes equations and a finite element based computational structural analysis (CSA) code, as well as of an efficient module for the interactive modification of geometry of components. The design and optimisation process is demonstrated for a Francis turbine runner blading. The process starts with an initial design taken from the ANN which is in the next step analysed by means of our inhouse Navier-Stokes solver NS3D with respect to the performances to be guaranteed. The turbine stage is modelled as a single blade passage of the rotor and a stator blading. The CFD-results may be analysed by tailor-made post-processing. Based on the numerical simulation results the initial design may be interactively improved in an iterative process. The design targets in this stage of the process are efficiency and cavitation behaviour of the runner. Finally, the CFD-results are used as boundary conditions for a steady state CSA simulation yielding the distribution and magnitude of stresses within the runner, especially between the blades and the runner band (shroud). In addition to the time-independant CFD- and CSA-simulations a transient Fluid-Structur-Interaction (FSI) analysis may be simulated taking into account all blade passages in the rotor as well as in the stator. This simulation model yields a much more accurate loading of the structure due to unsteady pressure distributions. In this step the design can be scrutinised with respect to its fatigue resistance. After generating a design satisfying all requirements and boundary conditions the optimised geometry is fed back into the ANN to be used for similar applications. After having carried out design and optimisation processes for a number of characteristic hydraulic machinery a higher level of efficiency and cavitation behaviour may be reached.