Multi-objective Optimization Design for Reversible Counter-rotating Axial Flow Pump Based on the Adaptive Modified Genetic Algorithm

Conventional hydraulic design methodologies for reversible counter-rotating axial flow pumps predominantly adhere to traditional axial flow pump design principles. However, existing optimization approaches inadequately analyze the individual design parameters, their synergistic interactions, and relative contributions, leading to suboptimal pump efficiency and compromised bi-directional multi-objective operational performance. To overcome these limitations, this study proposes a surrogate model-based framework integrated with an automated optimization platform developed in Isight software. A comprehensive evaluation of 120 simulated sample points is conducted to quantify the effects of geometric parameters on hydraulic performance. The adaptive modified genetic algorithm (AMGA) is employed to maximize operational efficiency through systematic impeller parameter optimization within the design space, while satisfying predefined pump specifications. Sensitivity analysis identified the angle of attack as the most influential parameter, accounting for 51.28% and 56.2% of the total variance in head and efficiency, respectively. Post-optimization results demonstrated a 3% increase in simulated efficiency at the design operating point, accompanied by a 14.5 kW reduction in shaft power consumption. These findings establish a robust foundation for advancing the multi-objective design optimization of reversible counter-rotating axial flow pumps.