叶轮
熵产生
机械
消散
物理
分类
熵(时间箭头)
生物系统
绝热过程
湍流
数学优化
统计物理学
圆角(机械)
体积流量
替代模型
对数
流量(数学)
机械工程
多目标优化
布朗运动
算法
离心泵
作者
Hongyin Zhang,Li Cheng,Wentao Xu
摘要
Cavitation-induced irreversible energy loss limits the efficient and stable operation of mixed flow pumps. This study proposes an entropy production driven Multi-Objective Bayesian Optimization–Non-dominated Sorting Genetic Algorithm III (NSGA-III) method for designing a biomimetic serrated leading-edge impeller. The prototype impeller's principal hydraulic and structural parameters are retained, with serration height and tooth width as design variables. Three-dimensional numerical simulations are validated using model tests. Entropy production theory quantifies internal energy dissipation, and a multi-objective framework minimizes the total entropy production rate and low-pressure-region volume fraction under constraints of hydraulic efficiency and critical net positive suction pressure. A Kriging surrogate model, the expected hypervolume improvement criterion, and NSGA-III are integrated for global exploration and local refinement. The recommended serrated impeller has a tooth height of 7.78 mm and a tooth width of 4.32 mm. Relative to the prototype, the total entropy production rate decreases from 0.90 to 0.72 W/K (a 20.4% reduction) and low-pressure-region volume fraction from 1.81% to 0.68% (approximately 62.6% reduction), without appreciable hydraulic-performance deterioration. The serrations induce localized leading-edge turbulence modulation that redistributes near-wall momentum and blade loading, suppressing downstream cavity growth, separation, and intense shear. Rear-region entropy production decreases by 32.4%, 2.04 times the added dissipation in the front and middle regions. These results reveal a spatial redistribution mechanism of controlled leading-edge dissipation and dominant rear-region loss suppression, guiding cavitation-resistant, low-loss biomimetic design of mixed flow pumps.
科研通智能强力驱动
Strongly Powered by AbleSci AI