Diagnosis of energy conversion efficiency degradation in a mixed-flow pump under rotational-speed and flow-rate mismatch: Implications for marine propulsion

Mixed-flow pumps are widely used in hydropower, large-scale pumping stations, and waterjet propulsion. These pumps are typically engineered for peak efficiency under design conditions. These pumps are typically engineered for peak efficiency under design conditions. However, practical marine requirements often compel a pump to operate under off-design conditions where the rotational speed and the incoming flow velocity (determined by ship speed) become mismatched. Such mismatch markedly degrades energy conversion efficiency. To systematically diagnose this degradation, this study integrates high-fidelity computational fluid dynamics with entropy production theory, establishing a quantitative thermodynamic framework for irreversible loss analysis in a mixed-flow pump under this specific off-design scenario. The numerical model is rigorously validated through high-speed imaging and pressure fluctuation measurements. A key contribution of this work is the quantitative causal linkage established between specific off-design flow structures and their entropy generation signatures. Results reveal that detrimental flow structures govern energy dissipation, with dominant mechanisms shifting across mismatch regimes. Turbulent dissipation is the primary entropy source. Total entropy production increases nonlinearly with rotational speed, exhibiting heightened sensitivity under low flow rates. A distinctive “double-saddle” distribution of entropy production versus incoming flow velocity is identified, whose characteristic extrema migrate linearly with rotational speed ( R 2 > 0.99). The optimal efficiency point is not stationary; linear regression shows that optimal rotational speed increases linearly with flow rate ( R 2 = 0.979), providing a practical criterion for energy-efficient speed selection. These findings provide a mechanistic and quantitative basis for performance prediction and operational optimization under extreme off-design conditions.