Cavitation Flow Mechanisms and Fluid Dynamics Optimization Trends in Small Water Pumps

Shi-da Zhang, Yan-Zuo Chang, Zi-An Wang, Yu-Xuan Chen, Zhen-Ming Chen, Xiao-Feng Zhu, Shu-Wei Huang, Gao-Xiong Xie, Jun-Cai Xie

fluid dynamics; small pump; cavitation flow; hydraulic loss; CFD numerical simulation; power engineering

Small centrifugal pumps are core basic equipment for fluid transport, waste heat circulation, and cooling systems in energy and power engineering. Consequently, internal cavitation flow represents the most prominent fluid dynamics problem restricting their efficiency, stability, and service life. More precisely, cavitation is a typical gas-liquid two-phase unsteady flow phenomenon involving complex physical mechanisms such as liquid vaporization, bubble growth, shedding, collapse, and the evolution of multi-scale vortex structures. It is evident from industrial statistical data that approximately 30% of small pump failures are directly or indirectly related to cavitation, efficiency reductions caused by cavitation can reach 20%-40%, and severe cavitation erosion can shorten impeller life by more than 50%. This paper therefore provides a systematic and rigorous review of the current application status of fluid dynamics theory and numerical simulation methods in cavitation mechanism analysis, flow loss identification, and structural optimization. From this review, several important conclusions can be drawn naturally and appropriately: hybrid RANS-LES models can improve flow resolution accuracy in cavitation regions by 30%-50% compared with conventional URANS models in strongly curved flow passages; deep learning-based vibration/acoustic signal recognition methods have already achieved diagnostic accuracy above 95% for incipient cavitation; and a lightweight Vision Transformer model achieves 100% recognition accuracy for eight cavitation states under multi-noise environments, with a single inference time of only 15.4 milliseconds. However, it is undeniable that clear and prominent research bottlenecks still remain in this field: turbulence-cavitation coupling models have insufficient capability in capturing cross-scale vortex structures, multi-field coupled simulation systems are not yet mature, the cavitation evolution law under variable operating conditions lacks quantitative description, and the intelligence level of optimization design needs to be improved. Accordingly, the future development trends are very clear: cross-scale refined CFD simulation, construction of multi-field coupling systems, precise optimization for energy saving and loss reduction based on entropy production theory, and intelligent simulation optimization with deep integration of fluid dynamics and artificial intelligence. This paper provides excellent theoretical reference value for hydraulic performance optimization, cavitation suppression design, and high-efficiency energy-saving renovation of small fluid power equipment.

Received: 29 Apr 2026, Received in revised form: 28 May 2026, Accepted: 03 Jun 2026, Available online: 09 Jun 2026

10.22161/ijaers.136.4