Flow in a centrifugal pump impeller at design and off-design conditions - Part II: Large eddy simulations

Flow in a centrifugal pump impeller at design and off-design conditions - Part II: Large eddy simulations
复制标题

DOI:
10.1115/1.1524586
复制
发表时间:
2003-01-01
影响因子:
2
通讯作者:
Pedersen, N
Pedersen, N
中科院分区:
工程技术4区
文献类型:
--
作者:
Byskov, RK;Jacobsen, CB;Pedersen, N

文献摘要

被引文献

相似文献

采用大涡模拟(LES)对一台带护罩的六叶离心泵叶轮内的流场进行了研究。通过在商业CFD软件FINE/Turbo中实施的局部动态Smagorinsky模型对亚格子尺度的影响进行了建模。详细分析了设计工况(流量\(Q = Q_{d}\))以及严重偏离设计工况(四分之一负荷,\(Q = 0.25Q_{d}\))下LES的结果。在设计工况下,LES显示流场状况良好,无明显分离。在四分之一负荷时,相邻叶轮通道之间存在显著差异。在一个通道入口处观察到稳定的非旋转失速现象,并且在通道的其余部分形成了一个相对涡。失速使相邻通道畅通,该通道呈现出以旋转效应为主的流动。将LES预测的速度以及基于Baldwin - Lomax和Chien \(k - \epsilon\)湍流模型的稳态雷诺平均纳维 - 斯托克斯(RANS)模拟结果与粒子图像测速(PTV)获得的实验数据进行了比较。LES所观察到的复杂的双通道现象通过PIV得到了令人满意的验证。然而,发现两种RANS模型无法重现四分之一负荷时观察到的失速现象,并且无法检测两个通道之间的差异。
The flow field in a shrouded six-bladed centrifugal pump impeller has been investigated using large eddy simulation (LES). The effect of the subgrid scales has been modeled through a localized dynamic Smagorinsky model implemented in the commercial CFD code FINE/Turbo. A detailed analysis of the results of LES at design load, Q=Q(d), and severe off-design conditions, at quarter-load Q=0.25Q(d), is presented. At design load LES reveals a well-behaved flow field with no significant separation. At quarter-load significant differences between adjacent impeller passages are revealed. A steady nonrotating stall phenomenon is observed in the entrance of one passage and a relative eddy develops in the remaining part of the passage. The stall unblocks the adjacent passage which exhibits a flow dominated by rotational effects. Velocities predicted by LES and steady-state Reynolds averaged Navier-Stokes (RANS) simulations based on the Baldwin-Lomax and Chien k-epsilon turbulence models are compared with experimental data obtained from particle image velocimetry (PTV). The complex two-channel phenomena observed by LES is with satisfactory agreement confirmed by PIV However it is found that the two RANS models do not reproduce the stall phenomenon observed at quarter-load and are incapable of detecting the differences between the two passages.