Optimization design of tandem pump based on surrogate method

In this paper, we propose a global design optimization method based on a surrogate model for an axial-flow tandem pump, in order to overcome the disadvantages of conventional optimization methods, including their long design period and lower efficiency. We discuss the influences of the primary design parameters on pump performances, and used fixed blade angles in both the front and rear impellers, as well as the phase angle, as design variables. We selected efficiency and the minimum average pressure coefficient on the axial sectional surface as the objective functions, which represent the energy and cavitation performances, respectively. Surrogate models were constructed based on various methods. In addition, we used global sensitivity analysis and Pareto front methods to further analyze the design parameters and optimum point. The results show that the optimization results can enhance tradeoff performances well, with respect to both efficiency and cavitation performance, in the vicinity of the design flow of the tandem pump. The influence of phase angle on performance can be neglected relative to the influence of the other two design variables. The impact ratio on efficiency of a fixed blade angle in the two impellers is the same as in their designed loading distributions. A fixed blade angle in the front impeller has a great influence on cavitation performance.