NUMERICAL INVESTIGATIONS OF A CENTRIFUGAL BLOOD PUMP

2004 ◽  
Vol 04 (03) ◽  
pp. 237-255 ◽  
Author(s):  
W. K. CHAN ◽  
Y. W. WONG ◽  
Y. DING

This paper presents computational fluid dynamics (CFD) studies of a centrifugal blood pump. 3-D models of five different blade geometries are investigated numerically using CFX-TASCflow. The impellers were designed using an inverse design technique where the swirl distributions were prescribed. The results showed the flow in the impeller passages is highly dependent on the impeller blade profiles. The flow in the radial blade impeller is unsatisfactory as flow separates at the leading edge of the suction side. Flow is confined mainly to the pressure side. Design 2, with an inlet angle of 6.7° and outlet angle of 30°, offers the greatest potential as only a small region of flow reversal is detected. Further optimization is necessary to completely eliminate regions of flow reversals. The highest scalar shear stress in both designs is 240 Pa and 120 Pa respectively. In addition, this paper demonstrates that the use of inverse design can help the designer to better design and analyze the flow field in centrifugal blood pumps.

2012 ◽  
Vol 12 (03) ◽  
pp. 1250045 ◽  
Author(s):  
JIAXING QI ◽  
YANHONG ZHOU ◽  
DONGFANG WANG ◽  
LIANG ZHONG

Computational fluid dynamics (CFD) simulations of the flow in an axial blood pump with different blade heights (BH150, BH200 and BH250) were performed in the present study. The flow in the pump was assumed as steady and turbulent, and blood was treated as incompressible and Newtonian fluid. The flow rate increased with the rise in blade heights. At the impeller rotating speed of 20,000 rpm and a pressure of 100 mm Hg, the pump produces a flow rate up to 5 L/min in BH200 and BH250 models. The reverse flow and vortices have been identified in the BH150 and BH200 models in the outlet regions, but not for BH250 model. The high shear stress of the flow in the pump mainly occurred at the blade tips. The BH200 model achieved an expected flow rate up to 5 L/min with 90% of the shear stresses less than 500 Pa and the exposure time less than 22 ms, which has the acceptable shear stress level in the literature.


2019 ◽  
Vol 142 (2) ◽  
Author(s):  
Bo Qian ◽  
Peng Wu ◽  
Bin Huang ◽  
Kai Zhang ◽  
Shiyang Li ◽  
...  

Abstract The vibration performance of centrifugal impellers is important for pumps and hydraulic excitation is a key source of vibration. The complex internal secondary flow in the centrifugal impeller brings degradation on vibration performances. An attempt of optimization by controlling the thickness distribution of centrifugal impeller blade is given to repress the internal secondary flow and alleviating vibration. The usual method of modifying an impeller on vibration performance is applying splitter blades. In this study, an ordinarily designed impeller is improved by the optimization attempt and the optimized impeller (OPT) is compared with the prototype impeller (PRT) with traditional splitter blades. The vibration performances of the impellers, the PRT, the ordinary impeller (ODN), and the OPT, are investigated numerically and experimentally. Meanwhile, further study on the influence of the thickness distribution optimization on vibration is conducted. There is a relative velocity gradient from suction side (SS) to pressure side (PS) in impeller ODN, causing nonuniformity of energy distribution. By means of thickness distribution optimization, the impeller blade angle on the PS and SS along the blade-aligned streamwise location is, respectively, modified and therefore the flow field can be reordered. The energy transfer in impeller is also redistributed after the modification of blade thickness distribution. What is more, experimental research upon impeller PRT and impeller OPT is also complemented to support the computational fluid dynamics (CFD) results. The experimental results show that the hydraulic performance of the impellers basically agree with the CFD results and the vibration data also proves a better vibration performance of the OPT.


Processes ◽  
2019 ◽  
Vol 7 (7) ◽  
pp. 431
Author(s):  
Si Huang ◽  
Jiaxing Huang ◽  
Jiawei Guo ◽  
Yushi Mo

By using EDEM–Fluent codes and coupling the continuous fluid medium with a solid particle discrete element, the solid–liquid two-phase flow field in a centrifugal pump was simulated under the same inlet conditions of the particle volume fraction and three flow conditions of 0.7Qd, 1.0Qd and 1.3Qd. By introducing the Archard wear model, the wear was calculated, and the wear law was obtained for the pump flow parts such as the leading edge of the impeller blade, blade tip, blade pressure side, blade suction side, impeller shroud, hub and volute. The results demonstrate that the wear of volute is about 70% of the total wear of pump. The wear in the impeller mainly occurs in the blade leading edge, the junction of the hub and the trailing part of the blade pressure side, and the junction of the shroud and the rear part of the blade suction side. Under lower flow conditions, the wear in the impeller shroud is relatively considerable. As the flow rate increases, the wear in the blade pressure side and the hub increases significantly.


Author(s):  
Dimitri P. Tselepidakis ◽  
Sung-Eun Kim

This paper presents the computation of the flow around a controlled diffusion compressor cascade. Features associated with by-pass transition close to the leading edge — including laminar leading-edge separation — contribute significantly to the evolution of the boundary layer on the blade surface. Previous studies have demonstrated that conventional k-ε models, based on linear or non-linear Boussinesq stress-strain relations, are able to capture by-pass transition in simple shear, but are unable to resolve transitional features in complex strain, like the leading-edge separation bubble, which is of considerable influence to the suction-side flow at high inlet angle. Here, the k-ω turbulence model has been implemented in a nonstaggered, finite-volume based segregated Reynolds-Averaged Navier-Stokes solver. We demonstrate that this model, if properly sensitized to the generation of turbulence by irrotational strains, is capable of capturing the laminar leading-edge separation bubble. The real flow around the leading edge is laminar and the transition is only provoked on the reattachment region. Additional investigation of transition in a flat-plate boundary layer development has also produced reasonably promising results.


Author(s):  
N. Arndt ◽  
A. J. Acosta ◽  
C. E. Brennen ◽  
T. K. Caughey

This paper describes an experimental investigation of rotor–stator interaction in a centrifugal pump with several vaned diffusers. Steady and unsteady diffuser vane pressure measurements were made for a two–dimensional test impeller. Unsteady impeller blade pressure measurements were made for a second two–dimensional impeller with blade number and blade geometry identical to the two–dimensional impeller used for the diffuser vane pressure measurements. The experiments were conducted for different flow coefficients and different radial gaps between the impeller blade trailing edge and the diffuser vane leading edge (5% and 8% of the impeller discharge radius). The largest pressure fluctuations on the diffuser vanes and the impeller blades were found to be of the same order of magnitude as the total pressure rise across the pump. The largest pressure fluctuations on the diffuser vanes were observed to occur on the suction side of the vane near the vane leading edge, whereas on the impeller blades the largest fluctuations were observed to occur at the blade trailing edge. However, the dependence of the fluctuations on the flow coefficient was found to be different for the diffuser vanes and the impeller blades; on the vane suction side, the fluctuations were largest for the maximum flow coefficient and decreased with decreasing flow coefficient, whereas at the blade trailing edge, the fluctuations were smallest for the maximum flow coefficient and increased with decreasing flow coefficient. Increasing the number of the diffuser vanes resulted in a significant decrease of the impeller blade pressure fluctuations. The resulting lift on the diffuser vanes was computed from the vane pressure measurements; the magnitude of the fluctuating lift was found to be larger than the steady lift.


1990 ◽  
Vol 112 (1) ◽  
pp. 98-108 ◽  
Author(s):  
N. Arndt ◽  
A. J. Acosta ◽  
C. E. Brennen ◽  
T. K. Caughey

This paper describes an experimental investigation of rotor-stator interaction in a centrifugal pump with several vaned diffusers. Steady and unsteady diffuser vane pressure measurements were made for a two-dimensional test impeller. Unsteady impeller blade pressure measurements were made for a second two-dimensional impeller with blade number and blade geometry identical to the two-dimensional impeller used for the diffuser vane pressure measurements. The experiments were conducted for different flow coefficients and different radial gaps between the impeller blade trailing edge and the diffuser vane leading edge (5 and 8 percent of the impeller discharge radius). The largest pressure fluctuations on the diffuser vanes and the impeller blades were found to be of the same order of magnitude as the total pressure rise across the pump. The largest pressure fluctuations on the diffuser vanes were observed to occur on the suction side of the vane near the vane leading edge, whereas on the impeller blades the largest fluctuations were observed to occur at the blade trailing edge. However, the dependence of the fluctuations on the flow coefficient was found to be different for the diffuser vanes and the impeller blades; on the vane suction side, the fluctuations were largest for the maximum flow coefficient and decreased with decreasing flow coefficient, whereas at the blade trailing edge, the fluctuations were smallest for the maximum flow coefficient and increased with decreasing flow coefficient. Increasing the number of the diffuser vanes resulted in a significant decrease of the impeller blade pressure fluctuations. The resulting lift on the diffuser vanes was computed from the vane pressure measurements; the magnitude of the fluctuating lift was found to be larger than the steady lift.


1989 ◽  
Vol 111 (3) ◽  
pp. 213-221 ◽  
Author(s):  
N. Arndt ◽  
A. J. Acosta ◽  
C. E. Brennen ◽  
T. K. Caughey

The interaction between impeller blades and diffuser vanes in a diffuser pump was investigated. Steady and unsteady pressure measurements were taken on the diffuser vanes, and the shroud wall of a vaned and a vaneless diffuser. Steady, unsteady, and ensemble-averaged unsteady data, as well as frequency spectra, are presented. The measurements were made for different flow coefficients, shaft speeds, and radial gaps between impeller blade trailing and diffuser vane leading edge (1.5 and 4.5 percent based on impeller discharge radius). The resulting lift on the vane, both steady and unsteady, was computed from the pressure measurements at midvane height. The magnitude of the fluctuating lift was found to be greater than the steady lift. The pressure fluctuations were larger on the suction side than on the pressure side attaining their maximum value, of the same order of magnitude as the total pressure rise across the pump, near the leading edge. Pressure fluctuations were also measured across the span of the vane, and those near the shroud were significantly smaller than those near the hub. The pressure fluctuations on the shroud wall itself were larger for the vaned diffuser than a vaneless diffuser. Lift, vane pressure, and shroud wall pressure fluctuations decreased strongly with increasing radial gap.


2019 ◽  
Vol 142 (1) ◽  
Author(s):  
B. Mischo ◽  
P. Jenny ◽  
Y. Bidaut ◽  
N. Fonzi ◽  
D. Hermann ◽  
...  

Abstract Unshrouded industrial centrifugal compressor impellers operate at high rotational speeds and volume flow rates. Under such conditions, impeller blade excitation is dominated by high frequency interaction with stationary parts, i.e., vaned diffusers or inlet guide vanes. In a previous study conducted on two full compression units of the original equipment manufacturer (OEM), the authors identified, characterized, and quantified resonant blade vibration caused by the interaction of the impeller blades with rotating stall cells during severe off-design conditions. This caused significant dynamic stress in the blades. In a follow-up study, this phenomenon was reproduced successfully experimentally under representative off-design conditions in a downscaled test rig and numerically with unsteady computational fluid dynamics (CFD) and structural mechanical finite element method (FEM) analysis. The gained knowledge was translated into a new diffuser design philosophy, based on sectorwise circumferential variation of the leading edge angle. This paper presents the patented philosophy, which is experimentally verified on the same test rig configuration in terms of flow path geometry and measurement equipment that was used in the mentioned prior study to assess resonant blade interaction. The results confirm the design aims: rotating stall onset was delayed without affecting the aerodynamic performance of the stage. Resonant blade interaction with rotating stall observed in the baseline diffuser could not be avoided with the two new diffuser designs. However, with the two new diffusers, the induced mechanical stresses in the impeller and the excitability were reduced by up to 12%.


Author(s):  
R. Nanthini ◽  
B. V. S. S. S. Prasad ◽  
Y. V. S. S. Sanyasiraju

Abstract In an iterative inverse design of a turbine blade, choice of initial guess profile is crucial. As the pressure distribution is very sensitive to the leading and trailing edge shapes and the profile slope and curvature, a good initial guess profile will help in faster convergence. In this paper, the sensitivity of the pressure distribution is determined by carrying out numerical simulations with ANSYS Fluent 17.2 for the inviscid flow. The flow domain comprises of a two dimensional transonic turbine cascade. It consists of a turbine blade enclosed by inlet, outlet and periodic boundaries. Inlet total pressure, total temperature and inlet angle are given as the boundary conditions at the inlet and static pressure is imposed at the outlet boundary. The flow is solved for continuity, momentum and energy equations. Sensitivity of different parameters — leading edge thickness, trailing edge thickness, leading edge shape, inlet and outlet wedge angle on the pressure distribution is demonstrated for VKI blade cascade. It is found that the pressure side of the profile is less sensitive and that even a small variation in suction side of the profile geometry can affect the performance of the blade significantly. It is shown that, with the proposed methodology and sequence of steps, the final guess blade is quite close to the original blade.


Sign in / Sign up

Export Citation Format

Share Document