Multi-Objective Optimization of Double-Suction Centrifugal Pumps for Erosion Resistance and Efficiency Enhancement Based on Theoretical Model Constraints and Shapley-Value Coordination

Double-suction centrifugal pumps operating in sediment-laden irrigation districts of the Yellow River Basin are prone to severe sediment-induced wear, while achieving erosion resistance and efficiency enhancement under hydraulic performance constraints remains challenging. This study proposes a multi-objective optimization method integrating theoretical model constraints with Shapley-value-based coordinated evaluation. Flow rate, head, and sediment particle size are used as inputs, while material, rotational speed, and impeller outlet diameter are treated as design variables. An optimization model considering efficiency loss, wear rate, and disk friction loss ratio is established, and dynamic Shapley-value weighting is applied to adaptively select design-variable combinations. Full-operating-range optimization shows that increasing head is mainly accommodated by higher rotational speed, whereas under coarse-sediment conditions, a reduced-speed, increased-diameter strategy is selected in approximately 43% of comparable flow-rate–head combinations to mitigate wear. Engineering validation shows that the optimized pump achieves an increase of 3.4 percentage points in peak efficiency, with the calculated wear rate of 0.87 mm/year. Long-term operation from 2021 to 2024 remained within ±3% fluctuation, and vibration indices met industry safety standards. The proposed method provides a practical basis for erosion-resistant, efficiency-oriented pump design and adaptive parameter selection under sediment-laden conditions.