Abstract
In this study, an optimization design was performed to improve the hydrodynamic performances of an ultra-small turbine based on a 3-D Reynolds-averaged Navier-Stokes (RANS) analysis. The runner of the turbine was selected as the computational domain and optimization object. The locations of the inlet, outlet and external interface of the computational domain were set to 3 times, 6 times and 3 times the diameter of the runner, respectively. For the turbulent closure problem in the RANS analysis, the shear stress transport (SST) turbulence model was used, and a grid dependency test was performed by applying the grid convergence index (GCI) based on the Richardson extrapolation method considering the grid discretization error. Four design variables related to the geometry of the runner blade were selected to maximize the power coefficient. The radial basis neural network (RBNN) was used as the surrogate model and trained to improve prediction accuracy. The optimization results obtained by using machine learning method showed highly accurate prediction values, and compared with the reference design, the optimum design provided improved hydraulic performances.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
