Uma Metodologia Hibrida De Otimizacao - Aplicada As P As De Turbinas Hidr Aulicas Axiais
This thesis presents a methodology as optimization method for the blades of axial turbine runners. The aim is to determine the optimal shape for the blade, considering a specific condition and operation point, improving this turbine performance. The motivation behind this work is based on search of the "best"design, increasingly, in a short time, resulting in e orts to new optimization methods and design cycles. The optimization method combines two heuristic optimization techniques: Genetics algorithms (GA) and particle swarm optimization (PSO). The addition of these two techniques in one algorithm (GA-PSO) allows an optimization method more robust and ecient for the purpose of optimal runner's shape. The objective function to be optimized is the performance generated by the rotor.
To computation of the performance by the turbine blades, a commercial CFD (Computational Fluid Dynamics) code is used. This methodology is applied to a hidrokinetic turbine and a bulb turbine. The performances of the hidrokinetic turbine, for di erent operating conditions, of the optimized rotor with GA-PSO were compared to results of previous optimization methods. Proving that the hybrid GA-PSO algorithm has the best performance, being applied to optimization of an bulb runner shape. The results showed a gain in power output (hidrokinetic turbine) and eciency (bulb turbine) , beyond a improvement in the the processing time, which demonstrated the viability of the use of GA-PSO algorithm as a good design procedure for shape optimization of axial hydraulic turbines.