You are on page 1of 7

Full Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No.

3, July 2012

Performance Analysis of Centrifugal Pumps Impeller of Low Specific Speed for Various Designs and Validation of Result by CFD
S N Sapali, Prof in Mech Engg Dept. College of Engg Pune-05
Rajiv Kaul , College of Military Engg, Pune-3(India) Email :- rajivkaulcme@yahoo.com
Abstract: The design of the pump is primarily dictated by the discharge rate and head to be developed. Various parameters affect the pump performance and energy consumption. The design of centrifugal pump has reached the stage where improvements can only be achieved through a detailed understanding of the internal flow. The prediction of the flow in such equipment is very complicated due to the rotation and the curved three-dimensional shape of the impellers. Furthermore, the flows in pump show unsteady behavior, especially at off- design conditions, as a result of interaction between impeller and pump casing. CFD analysis is very useful for predicting pump performance at various mass-flow rates. For designers, prediction of operating characteristics curve is most important to evaluate the performance of pump. Key words: - Centrifugal pump, Impeller, CFD, blade angle, turbulence, triangles, solid modeling, Meshing

in the machine for various designs (blades angle) and thus help the engineer to obtain a thorough performance evaluation of a particular design. The Centrifugal Pumps impeller under consideration was designed as a radial type, in which the flow noise or pulsation generally remains low. The inlet passage is entirely perpendicular to the rotating axis to maximize the centrifugal force and a semi-open construction (having only the back shroud) to facilitate fabrication and minimize disk friction. II. DESIGN OF IMPELLERS BLADES The blade design serves to define the shape of the blades along the outer, inner and mean streamlines so that the inlet and outlet angles could be obtained. The blade describes a three-dimensional curve along each surface of revolution. The length of the streamlines in circumferential direction is still unknown; within the certain limits it can be freely selected. It has been observed in impeller type, the most serious problem for adopting a semi-open impeller is the control of the impeller-axial tip clearance. A. Semi-open and open impellers Impellers without a front shroud are termed as semiopen. The axial thrust is higher in semi-open than with closed impellers, because the pressure in the vane-less impeller, sidewall gap always exceeds the pressure within the bladed impeller. A large clearance would invite leakage loss, while small clearance would run the risk of seizure by contact. The angles between two consecutive blades are different (720 in 5-blade pump, 600 in 6-blade pump and 51.430 in 7blade pump), therefore the distances between the nearest blade before the tongue and tongue denoted by parameter dare different in three pumps. Also it has been observed that this distance has a great influence on the flow field. 6-blade pump has a larger distance of blade with the tongue; therefore flow separation has begun later than the other two cases. The unique behavior of the 6-blade pump is related to its large distance of blade with the tongue. Larger distance has stronger backflow effects.

I. INTRODUCTION Centrifugal pumps are probably among the most often used machinery in industrial facilities as well as in common life. After being invented they passed long evolutionary way until they became accessible for various applications. Their physical principle is described centuries ago by Euler through a well known equation named after him called Eulers equation for turbo-machinery. Many technical applications of centrifugal pumps cannot be fulfilled without proper analysis, especially regarding pumps output parameters i.e. its head and efficiency. Pumps with specific speed lower than 500 are of no practical use, and pumps with specific speed lower than this value suffer reduced efficiency and are no longer practical. Specific speed, Ns = (discharge Q is in lit/s) (1)

Nspeed of pump (rpm), HHead developed by pump (m). The trend toward higher speed, high power density liquid turbo-machinery has inevitably increased the potential for fluid/structure interaction problems, and the severity of those problems. The current investigation is aimed to simulate the complex internal flow in a centrifugal pump impeller (semi-open) for low specific speed (900-1250 rpm) with six twisted blades by using a three-dimensional Navier-Stokes equation with a standard k- (SST) turbulence model. Numerical simulations can provide quite accurate information on the fluid behavior 2012 AMAE DOI: 01.IJRTET.7.3.507 1

Full Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 3, July 2012

Fig. 1: Graphical view of Impeller with blades (dimensions are in mm) TABLE I: DENOMINATION OF I MPELLER (TYPE A)

T ABLE II: DENOMINATION OF N EW IMPELLER (TYPE B)

2012 AMAE DOI: 01.IJRTET.7.3.507

Full Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 3, July 2012

Fig. 2: Impellers of various materials (Brass, CI, Copper)

III. CFD AND NAVIER-STOKES (N-S) EQUATIONS Computational Fluid Dynamics (CFD) complements experimental and theoretical fluid dynamics by providing an efficient means of simulating fluid flow of practical interest. Many tasks can numerically be solved much faster and cheaper than by means of experiments. N-S equation for Turbulent Flow:- Navier-Stokes (N-S) equations describe the flow of viscous fluid most accurately. However, at higher Reynolds Number the flow grows turbulent as in the case of water pump. Study of turbulent flow is rather a difficult area, analytically as well as computationally. For more complex situation like pump, turbine etc, solution of the N-S equation must be found with the help of computers. This is a field of science by its own, called Computational Fluid Dynamics (CFD). Turbulent Flow is characterized by random and rapid fluctuations swirling regions of fluid, called eddies, through out the flow. These fluctuations provide an additional mechanism for momentum and energy transfer. The swirling eddies transport mass, momentum and energy to other regions of flow much more rapidly than molecular diffusion greatly enhancing mass, momentum and heat transfer. As a result, turbulent flow is associated with much higher values of friction, heat transfer and mass transfer coefficients. Even when the average flow is steady, the eddy motion in turbulent flow causes significant fluctuations in the values of velocity, temperature, pressure and even density. Some scientists conceived turbulent motion as a simple superposition of a fluctuating motion on a mean motion. Velocity vector component V = (u,v,w) at a point in a turbulent motion could be written as

Where u , v, w, p ; denote mean and u , v , w, p denote fluctuating components. The magnitude of is usually just a few percent of , but the high frequencies of eddies make them very effective fo the transport of momentum, thermal energy and mass. The chaotic fluctuations of fluid particles play dominant role in pressure drop on these random motions must be considered in analyses with the average velocity. For viscous term, time averaged N-S equation for turbulent flow

(u

2 u v p u u v ) ( u ) ( uv) x t x x x y y

(2)
2 v w p v v w ) ( v ) ( vw) y t y y x z z

(v

(3) 2 w u p w w (w u ) ( w ) ( uw) z t z z x x x (4) Where p = pressure of flowing fluid, = density of fluid, = viscosity of fluid. These equations are same as the one for laminar flow except that the shear force terms have been modified due to the presence of turbulence of the flow. These additional stresses are called apparent stresses. Since stress is a tensor of order 2, it has nine components

u u u , v v v , w w w ,
p p p .
2012 AMAE DOI: 01.IJRTET.7.3.507 3

Full Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 3, July 2012 The determination of these stresses posses a big problem. Since this not only depend on the properties of the fluid but also on the nature of the flow. However different models were developed for determining these stresses. No turbulence model is currently available that is valid for all types of flows and so it is necessary to choose and fine-tune a model for particular classes of flows. A. Shear Stress Transport Model (SST Model) It is the variation of the standard k- model. It is so named because the definition of turbulent viscosity is modified to account for the transport of the principal turbulent shear stress. It takes into account the transport of turbulent shear stress, and makes a gradual change of solution variables from the standard k- model in the inner region of the boundary layer to a high Reynolds number version in the outer part of the boundary layer. The numerical solution of the discretized threedimensional, incompressible Navier-Stokes equations over an unstructured grid is accomplished with a commercial CFD finite-volume code. For each impeller, the flow pattern and the pressure distribution in the blade passages are calculated and finally the head-discharge and efficiency-discharge curves are compared and discussed. IV. SIMULATION USING ANSYS SOFTWARE (12.2 VERSION) A CFD code has three basic components:- pre-processor, solver, post-processor. The solver is the heart of the code, carrying out the major computations and providing the numerical solutions. The pre-processor and post-processor are the front and end of the code, providing the user /machine interface that allows a CFD operator to communicate with the solver: inputting data to define the problem to be simulated and commanding the solver to use certain models to carry out the simulation. Solid modelling is at the core of this integrated design system. The solid model also provides the grid for finite volume (FV) simulation. The pump is divided into three regions; inlet, rotating impeller and outlet. First 3D modelling has been done using dimension of Pumps impeller. Then inlet, impeller, casing and outlet region was meshed using tetra-hedral elements; but a mixture of structured and unstructured grids was used for the inlet, rotating regions and outlet region. Number of nodes formed in meshing is 152346. A reduction of the grid spacing and time step could lead to an improved solution. Then using turbulence model k- (SST) and giving inputs (inlet velocity, pressure, rotating speed of impeller, discharge, and density of material), simulation was done. Since the fluid surrounding the impeller rotates around the axis of the pump, stationary and rotating frames were considered for flow equations. To accomplish this Multiple Reference Frame (MRF) model has been used. In the present case, volume flow rate and pressure, at inlet were used for the boundary conditions. Outer wall was stationary but impeller was rotating with certain speed (rpm). At solid walls defined by the user, the boundary conditions are set automatically by the CFD-program: (i) No-slip conditions at the impeller blades and walls. (ii) Velocity components perpendicular to the wall are zero. Roughness of all walls was considered 100mm. Validation normally has two parts: - first, a mathematical/ numerical verification, including convergence, grid independence and stability test; and second, physical validation through comparing predicted results with experimental data;

Fig. 3: Impeller (type A)

Fig.4: Sliced view of Impeller with casing (typeA)

2012 AMAE DOI: 01.IJRTET.7.3.507

Full Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 3, July 2012

Fig. 5: Meshing of Sliced Impeller with casing

Fig. 6: Contours of Velocity Magnitude (m/s)

Fig. 7:

Contours of static pressure (pascal)

Fig. 8:

Sliced view of Impeller with casing (Type B)

2012 AMAE DOI: 01.IJRTET.7.3.507

Full Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 3, July 2012

Fig. 9: Meshing of sliced Impeller and casing (Type B)

Fig. 10:

Contours of Velocity Magnitude (m/s) (Type B)

Fig.11: Contours of static pressure (pascal) (Type B)

V. EXPERIMENTAL RESULTS WITH VARIOUS TYPES OF IMPELLERS Experiments were conducted in the lab with available singe suction centrifugal pump with the various types of impellers

of different materials (CI, brass, aluminium) and design (blades angles 150,280; 180,300) at various speed. Results were plotted as shown below.

Fig. 12: Graph between flow Vs head and efficiency at 2600 rpm (Type A)

2012 AMAE DOI: 01.IJRTET.7.3.507

Full Paper Int. J. on Recent Trends in Engineering and Technology, Vol. 7, No. 3, July 2012

Fig. 13: Graph between flow Vs head and efficiency at 2600 rpm (Type B)

VI. RESULT ANALYSIS AND CONCLUSION In the present study, we intend to investigate the effect of use light materials of impeller and blade design on the pump efficiency. This paper introduces an integrated system with solid modeling, flow analysis, mechanical design, and all interconnected in a comprehensive suite of products. Considering the various turbulence models, turbulence model of standard k- (SST) with blade number of 6 for the pump, was analyzed. From experimental results (for Type A) it is observed that at 2600 rpm, efficiency of Aluminium impeller is the maximum and remains in the region of 37.5% with head, 3.67m. Whereas in case of CI impeller, maximum efficiency is 34% and head is about 4.25m, but at part loads, efficiency is quite low. Brass impeller is having maximum efficiency 32.72% with head 3.82m. In each case velocity increases from inlet of impeller, becomes maximum in the middle of impeller and then decreases up to outlet of the impeller. From experimental results (for Type B) it is observed that at 2600 rpm, efficiency of new impeller (Aluminium) is maximum (32.36%) with head, 3.27m. It is concluded that with the increase in blade angles, efficiency decreases due the separation and detachment of flow. Results were also compared with CFD simulation result (using software AnsysVersion12.1) and it was found that pressure varies in the outer region of pump from 2.03x104 to 5.07x104 Pascal, which resembles with experimental data at 2600 rpm. In simulation it was also observed that maximum pressure (1.01x105 Pascal) gets developed on rotating impeller zone but due to recirculation and separation just at the outlet, pressure of the fluid drops drastically. Fluid velocity variation in the CFD simulation was found to be 1.952.19 m/s while actually during experiment it was in the range of 1.581.93 m/s. Maximum losses in pressure occurs just at the exit of impeller (casing) due to the formation of eddies and separation of flow. Minimum clearance must be provided at the exit of 2012 AMAE DOI: 01.IJRTET.7.3.507 7

impeller and casing. Degree of finishing of impeller is also important in minimizing the mechanical losses which in turn increases the efficiency of pump and life of impeller blades. The impeller blades with the higher outlet blades angles do not reduce the flow separation at the blade exit, which accompanied by an increase of hydraulic losses that causes a decrease in pressure head. It is therefore recommended to use low blade angles and light material (like Aluminium, Brass) to fabricate pumps impeller. Apart from higher efficiency, transportation of pump will be an additional advantage. REFERENCES
[1] Takamatu, Young-Do Choi, Junichi Korokawa Internal Flow Cheracteristics of a very low Specific Speed Centrifugal Pump Journal of Fluids Engineering March 2006 - Volume 128, Issue 2, pp.341-349 [2] Soyama, Tatebayashi and Kazuhiro, Characteristics of screw centrifugal pump Journal of Turbomachinery October 2005 Volume 127, Issued 4, pp.755-762 [3] Majidi K and Seikman, Numerical simulation of a radial multi stage centrifugal pump, AIAA 2006, 44 th AIAA Aerospace Science Meeting. [4] Shi A. and Tsukmann Analysis of centrifugal pump BEP and recirculating flows, 11 th International Pump Users Symp, Houston 2006 [5] Shum Y, Comparison of steady state and transient RotorStator Interaction of an Industrial centrifugal Pump CFX Users Conference. Tokyo 2007 [6] H.Y.Li, M.M. Athwale, and Y.Jiang. Mathematical Basis and Validation of the Full Cavitation Model ASME FEDSM01, New Orleans, Louisiana, 2004. [7] J Plutecki and J Skrzypackz Application of CFD in incompressible flow visualization Proceeding of tenth IHSC, Poland, 2004. [8] Boaling Cui, Zuchao Zhu, Jianci Zhang, Ying Chen, The flow simulation and experimental study of low-specific-speed highspeed complex centrifugal impellers, Chin. J. Chem. Eng. 14 (4) (2006) 435441.

You might also like