In this paper, the three dimensional flow field of a diagnosis flow circulating waterpump for a 600MW steam turbine-generating set was simulated numerically. Based onnumerical computations, the optimal design of the impeller was performed then,coupling with some advanced algorithms such as genetic algorithm and artificial neuralnetwork, which brought out various samples of impeller. The effects of the three freeparameters (Gamma, Beta1 and Beta2) on the global performance were analyzed, whichwould be of help to the pump designing of the same kind. The efficiency of the impellerwas improved by 5.6%. A series of CFD computations were performed under seven different flow rates andthe global performance curves were obtained. Some primary parameters reflecting theinternal flow field were also analyzed in the paper, the comparison of which showednoticeable improvement of the impeller and the pump. The optimized pump model has been applied into practical design presently.
|