School of Energy and Power Engineering, Jiangsu University, Zhenjiang, Jiangsu, 212013, China
Research Center of Fluid Machinery Engineering and Technology, Jiangsu University, Zhenjiang, Jiangsu, 212013, China
A mixed flow pump with guide vanes was chosen as research model in this study, and eight parameters of the impeller were selected as optimization variables, including blade outlet inclination angle, blade wrap angle at hub, blade inlet angle and outlet angle at middle stream line, blade outlet width, front shroud inclination angle, hub inclination angle and vane number. Firstly, orthogonal experimental method and CFD numerical simulation method were used to produce samples, then the RBF neural network was adopted to establish the performance prediction model as the objective function, multi-island genetic algorithm was used for solving the objective function at last. Based on all the above, a method of multi-parameter optimization method on energy performance of mixed flow pump without changing the nominal diameter of impeller outlet was proposed and then verified by experiments. By this method, the pump head and efficiency at the design point of the model pump were increased by 11.5% and 4.32%, respectively. Meanwhile, the peak value of pressure pulsation coefficient at pump inlet, impeller outlet, guide vane outlet and pump outlet all decreased obviously, by a maximum decrease of 62.9%. Compared to the original model, the static pressure in the optimization model increased by 30kPa and the gradient of static pressure distribution after optimization becomes larger and more uniform. The turbulent energy intensity at the impeller outlet was reduced by 0.2m2/s2. The pressures at the 60% blade position and 80% blade position both increased by nearly 65kPa and the pressure decreased by 50kPa at the blade pressure side.