INSTITUTE OF TECHNOLOGY, NIRMA UNIVERSITY, AHMEDABAD 382 481, 08-10 DECEMBER, 2011

Design and Development of Split Case Pump Using Computational Fluid Dynamics
A. V.S. Kadam, B. S.S. Gawade, C. H.H. Mohite, D. N.K.Chapkhane

Abstract -- Todays requirement regarding Pump is to satisfy specified duty point means required head and discharge, with less power consumption. For the pump used for water service applications, periodic maintenance is required to check the wear out parts like wear ring, shaft sleeves. In such case horizontal split case pumps are more suitable. If the given duty parameter is unable to satisfy with existing split case pump range, it is necessity to develop new pump and developed pump should have more efficiency with energy crises taking into consideration. With the aid of computational fluid dynamics, the complex internal flow in horizontal split case pump can be well predicted, thus facilitating the design of pump. The numerical analysis is carried out with multiple frame of reference to predict the flow field inside the entire pump casing and impeller. Key Words- centrifugal pump, computational fluid dynamics, pump performance
I.

INTRODUCTION

entrifugal pump is a hydrodynamic machine, in which rotating impeller continuously transmits mechanical work from the driving machine to fluid. The kinetic energy is converted into potential pressure energy. Computational Fluid Dynamics (CFD) is one of the Computer Aided Engineering (CAE) tool. CFD has recently come up as an alternative approach to investigate the complex fluid flow phenomenon in pumps. It is rapidly becoming an important tool for analysis and design in hydraulic engineering. The input to CAE software is meshed geometry of the model and this geometric modeling is done by using Computer Aided Design (CAD) software. II. NEED OF DESIGN OF PUMP FOR NEW SERIES Now a day in most of the water service applications duty point requirements are of high head and discharge is between 1000 m3/hr to 2500 m3/hr. Current duty point requirement is head (H) is 218 m and discharge (Q) is 2100 m3/hr. The existing split case pump model range is up to 175m head for 2100 m3/hr head. So it is necessary to develop the pump for new higher head series for given discharge range. With the same time as the energy crises taking into consideration the developed pump should consume less power with more efficiency. The various types of pumps like multistage pump, Vertical Turbine (VT) pump, mixed flow (MF) pump can be offered for present required duty of high head and discharge. The Horizontal Split case pumps are more suitable for water

supply scheme projects because of following reasons: 1. Simple in design as compare to multistage pump and Vertical turbine pump. 2. Less rotating components and easy to assemble. 3. Less cost of production. 4. Easy for maintenance. Over the last 25 years, CFD has been increasingly used for a wide variety of engineering applications. In the beginning, the use of these techniques was customary only in the aerospace and nuclear fields. Subsequently, the use has spread to a variety of products, physical situations, and manufacturing processes. In the case of pump industry applications, CFD tools are important in view of analysis of the hydraulic passage of the pumps. For impeller inlet, CFD can help to improve the inlet flow distribution by proper designing/checking the quality of flow in pump [7]. The experimental and theoretical study requires enormous time. Also in experiment actual physical model of prototype is required to manufacture which is costly. As the CFD analysis can be carried out on prototype also. CFD tools avoid physical modeling and testing every time. Better and faster design of pump and its analysis leads to shorter design cycles. III. FLOW SIMULATION OF PUMP In order to obtain better design in CFD, following procedure is applied so that fluid flow can easily be modeled in the centrifugal pump impeller.

Fig.1. Flow chart of CFD procedure

It is now feasible to use CFD codes for a realistic prediction of the complex three dimensional (3D) turbulent flow in the entire pump. The 3D views of hydraulic flow passage of the suction volute, impeller and delivery volute as

2 shown in Fig.2.

INTERNATIONAL CONFERENCE ON CURRENT TRENDS IN TECHNOLOGY, NUiCONE 2011

Conditions: The first task to accomplish on numerical flow simulation is the definition of the geometry, followed by the grid generation. A structured grid is created using the HyperMesh software as shown in Fig.3. The grid of total flow domain passages has 55860 nodes. The blades, hub and shroud surfaces and casings surfaces are modeled as solid walls. All periodic grid surfaces are defined as periodic boundary conditions. Finally mesh file is imported in fluent software.

Fig.2. Flow domains of pump for CFD

A 3D flow simulation is carried out for hydraulic flow passage. Several numerical calculations are carried out to predict the flow pattern inside the entire impeller and casing, whereby the impeller flow field is solved in a rotating frame and the casing in a fixed one [2]. The two frames of reference are connected in such a way that they each have a fixed relative position throughout the calculation. The appropriate transformation occurs across a sliding interface without any interface averaging. The impeller with specific speed 51.8 metric unit is used. The pump main dimensional parameters are presented in Table.1. Impeller Value Description 820 mm Impeller outside diameter 37 Blade inlet angle 70 Blade outlet angle 26 mm Impeller Outlet Width 300 mm Impeller Eye Diameter 7 mm Blade Thickness 5 Number of Blades Volute Value Description 52 mm Volute inlet width 861 mm Cut water diameter 300 mm Volute outlet diameter

Fig.3. Meshed model for CFD analysis

The general parameters and boundary conditions used for the 3D flow simulation of the pump are summarized in Table.2. Parameters Description Extended Inlet and outlet, Suction volute, Delivery volute and impeller Structured Water at room temperature Mass flow= 2100 m3/hr=583.33 kg/s Pressure outlet Standard k-epsilon model Second order upwind 10-4 (RSM)

Parameter D2 1 2 b2 D1 Su z Parameter b3 D3 outlet

Flow simulation domain Grid Fluid Inlet Outlet Turbulence Model Discretization Maximum residual conversions criteria

Table.2. Simulation Parameters used in Fluent.

Table.1. Geometrical Parameters of the pump

It has found that a good volute design could improve the pump performance and operating range [4]. Volute studies have been performed on existing geometries and can be categorized into two groups: volute flow analysis and flow interactions between volute and impeller. There are five key geometrical parameters in volute design: circumferential variation of the cross sectional area; shape of the crosssection; radial position of the cross section; position of the volute inlet; and tongue geometry [1], [3]. A design requirement is to make the circumferential pressure distribution uniform.

Computational Domain, Grid Generation and boundary

The simulation domain at the inlet and outlet sections was sufficiently extended to allow inlet recirculation and the elliptic influence of the flow [5]. The flow solver of the CFD code employs for incompressible turbulent flow the continuity equation. Three criteria influence the choice of a turbulence model: (1) the physical nature of the problem, (2) the quality of attended results, and (3) computing power. Depending on the flow complexity, this choice is always crucial when using CFD codes. Nevertheless, the traditional Reynolds Averaged Navier Stoke (RANS) model, standard k- viscous model is used [6]. The walls are modeled using a log-law wall function. The transport equations are discretized using the conservative finite volume method. For the present study the advection terms in the integral equations are modeled employing a second order accurate skewed upwind

INSTITUTE OF TECHNOLOGY, NIRMA UNIVERSITY, AHMEDABAD 382 481, 08-10 DECEMBER, 2011

differencing scheme with physical advection correction. At the inlet of the computational domain, the mass flow rate in the absolute frame of reference [13], turbulence intensity, length scale and the reference pressure at one grid point are specified. The inlet velocity vector is in the axial direction. The blades, hub and shroud surfaces and casings surfaces are modeled as solid walls [5]. All periodic grid surfaces are defined as periodic boundary conditions. The fluid variables like outlet pressure and velocity are extrapolated at the outlet of the discharge nozzle sectional plane. IV. RESULTS The entire three-dimensional flow field in the centrifugal impeller, volute and circular casings, is obtained by numerical analysis. The fringed contours of total pressure and velocity distributions at mid span of the impeller and volute casings at the design point are shown in Fig.4. According to Fig.5 the total pressure distribution shows some negligible circumferential non uniformity at the impeller outlet. The CFD simulation offers a virtual image of the internal flow in the machine allowing the analysis and comprehension of more complex phenomena [1], [3]. Even though the effect of the volute tongue is appreciable, a relatively uniform pressure distribution is obtained around the impeller. The flow pattern in casings also affects the performance of the pump [7], [8].

point, the flow in the volute encounters a dividing plane, below which the flow exits the pump, above which the flow enters the volute. The leading edge of the tongue is analogous to that of an airfoil, a sharp leading edge has low blockage but is sensitive to angle of attack; while a well-rounded leading edge has large blockage but is insensitive to incidence. This difference manifests itself as a design tradeoff. A sharp volute tongue provides the absolute peak efficiency, while a volute with a round tongue has better off-design performance [9].

Fig.6. Contours of Total Pressure

Fig.7. Path lines colored by velocity magnitude

Fig.4. Velocity Vectors Colored by Velocity Magnitude (m/s)

Comparison between CFD analysis and Experimental testing: Comparison between the CFD analysis and experimental results for the hydraulic performance of the Pump is as shown in Table.3 and Fig.8. Sr. 1 2 3 4 Parameters Discharge Total head Pump input Efficiency CFD Analysis 2100 m3/hr 222 m 1548.30 kW 82 % Experimental Results 2100 m3/hr 219.7 m 1544.12 kW 81.37 %

Table.3. Comparison between CFD analysis and Experimental testing Fig.5. Contours of Total Pressure field on middle plane

The tongue of a volute is located corresponding to a joint of the volute smallest area, largest area, and exit cone. At this

INTERNATIONAL CONFERENCE ON CURRENT TRENDS IN TECHNOLOGY, NUiCONE 2011

[5]

John S. Anagnostopoulos, Numerical method for flow analysis and blade design in centrifugal pump impeller. Journal of Computers and Fluids 38, 2009, pp. 284289. [6] Chris Reeves, Computer simulation helps design more efficient water pumps. World Pumps Magazine, June 2004, pp. 32-34. [7] Craig Hornsby, CFD-driving pump design forward. World Pumps Magazine, August 2002, pp. 18-22. [8] Val S. Lobanoff and Robert R. Ross, Centrifugal Pumps design and application, Jaico publishing house, 1995. [9] A. J. Stepanoff, Centrifugal and Axial flow pumps Theory, Design and Application, John Wiley and Sons, 1953. [10] Stephen Lazarkiewicz and Adam T. Troskolanski, Impeller pumps, Pergramon Publication, 1965. [11] Shigley J.E. and Mischeke C.R., Mechanical Engineering Design, Tata McGraw-Hill Book Company, 2003. [12] John D. Anderson, Computational Fluid Dynamics-The Basics with Applications Taylor and Francis Publications, 1993.

Fig.8. H-Q characteristics using CFD and Experimental testing

Performance found satisfactory as per the Acceptance standard IS 9137. V. CONCLUSION The impeller of the existing closer range pump has been modified by increasing the diameter to 820 mm from 770 mm to suit the higher efficiency, required head and discharge. Considering economic incentive for operating range and efficiency is gained by better understanding of the influence of the tongue. Pump with higher efficiency and greater stable operating region is designed. The CFD analysis of the pump with modified impeller diameter is carried out to check the performance and efficiency of the pump. Efficiency of the pump from CFD results is coming 82 % and by actual performance test efficiency is coming 81.37%, by which it is confirmed that CFD analysis is clearly validated. Traditional volute design is based on two-dimensional analysis, and the emphasis is on collection (Impeller) and less on the diffusion (Volute) function. However, with the use of advanced fluid modeling tools, it is possible to design a volute using three-dimensional analysis. The volute design and analysis presented is indeed part of a new product in development at Kirloskar Brothers Limited. This shows that CAD and CAE tools are very useful in hydraulic design. VI. REFERENCES
[1] Ahmad Nourbakhsh Experimental study of characteristic curves of centrifugal pumps in different specific speeds. Journal of Experimental Thermal and Fluid Science 32, 2008, pp. 800807. W.K. Chan, L.P. ChuaThe flow patterns within the impeller passages of a centrifugal pump model. Journal of Medical Engineering and Physics 22, 2000, pp. 381393. S. Ahmad Nourbakhsh, Experimental and Finite Element Method (FEM) failure analysis and optimization of centrifugal pump volute casing. Journal of Engineering Failure Analysis 16, 2009. Dr. Ing Friedrich, Economic development of efficient centrifugal pump impellers by numerical methods. World Pumps Magazine May 2001, pp. 48-53. R. Spence, CFD for parametric study of geometrical variations on the pressure pulsations and performance characteristics of a centrifugal pump. Journal of Computers and Fluids 38, 2009, pp. 12431257.

[2]

[3]

[4]