0
Research Papers: Flows in Complex Systems

Shape Optimization of the Pick-Up Tube in a Pitot-Tube Jet Pump

[+] Author and Article Information
J. Meyer

Laboratory of Fluid Dynamics,
University of Magdeburg OVGU,
Magdeburg 39106, Germany
e-mail: jan.meyer@ovgu.de

L. Daróczy

Laboratory of Fluid Dynamics,
University of Magdeburg OVGU,
Magdeburg 39106, Germany
e-mail: laszlo.daroczy@ovgu.de

D. Thévenin

Professor
Laboratory of Fluid Dynamics,
University of Magdeburg OVGU,
Magdeburg 39106, Germany
e-mail: thevenin@ovgu.de

Contributed by the Fluids Engineering Division of ASME for publication in the JOURNAL OF FLUIDS ENGINEERING. Manuscript received February 9, 2016; final manuscript received August 10, 2016; published online November 3, 2016. Assoc. Editor: Bart van Esch.

J. Fluids Eng 139(2), 021103 (Nov 03, 2016) (11 pages) Paper No: FE-16-1089; doi: 10.1115/1.4034455 History: Received February 09, 2016; Revised August 10, 2016

At a very low specific speed (VLSS), pumps normally suffer from high disk friction losses. In order to solve this issue, it can be helpful to use a different centrifugal pump design, which is not often found in the pump industry: the Pitot-tube jet pump (PTJ pump). It shows superior performance at low specific speed due to a rather unconventional working principle, described in detail in this paper. The key design feature of the PTJ pump is the (fixed) pick-up tube. Its geometry has not varied over the last decades; it is referred to in this study as “initial” or “standard” design configuration. However, optimizing the pick-up tube might lead to a considerably higher performance. Therefore, a parameterized three-dimensional (3D) computer-aided design (CAD) model is used in this study to investigate the impact of shape deformation on pump performance with the help of computational fluid dynamics (CFD). Two CFD approaches are presented and compared for this purpose: a computationally efficient approach with limited accuracy (low-fidelity method) and a more detailed, but computationally more expensive, high-fidelity approach. Using both approaches, it is possible to obtain highly efficient PTJ pumps. As a consequence, first design rules can be derived. Finally, the optimized design has been tested for various operation points, showing that the performance is favorably impacted along the complete characteristic curve.

Copyright © 2017 by ASME
Your Session has timed out. Please sign back in to continue.

References

Figures

Grahic Jump Location
Fig. 1

Cutoff representation of a standard Pitot-tube jet pump. This CAD model shows rotating (denoted r) and static (denoted s) parts.

Grahic Jump Location
Fig. 2

Optimization workflow in the present study

Grahic Jump Location
Fig. 3

Left: 3D diffusor geometry with extended (fixed) inlets and outlets. In this configuration, only static pressure recovery can be simulated and optimized. Measuring planes used for velocity and static pressure are shown as boundary between green and purple regions. Right: complete PTJ pump with a parameterized diffusor. The interior wetted surface S within the pick-up tube is colored green.

Grahic Jump Location
Fig. 4

Definition of the design parameters used to describe the diffusor channel within the pick-up tube, both for low- and high-fidelity simulations. The projected area Ap is the area in the x–z plane.

Grahic Jump Location
Fig. 5

Grid convergence study for diffusor channel regarding diffusor efficiency ηD when varying cell number and turbulence model

Grahic Jump Location
Fig. 6

Parity plot showing diffusor efficiency obtained using either k–ε Realizable or k–ω SST turbulence model (filled square symbols). The design for the grid convergence study presented in Fig. 5 is shown as Δ in this plot.

Grahic Jump Location
Fig. 7

Grid convergence study for complete pump, showing pressure head H (left scale) and pump efficiency η (right scale)

Grahic Jump Location
Fig. 8

Typical mesh for CFD of a Pitot-tube jet pump near the pick-up tube, showing an arbitrarily selected design

Grahic Jump Location
Fig. 9

Test rig with a pipe system (L-1 to L-3), a storage tank with water as fluid (B-1), pressure gauges (P-1 and P-2), a motor (M), a ball valve (V-1), a PTJ pump (PTJ), a flow meter (F-1), and a temperature sensor (T-1)

Grahic Jump Location
Fig. 10

Characteristic curves for head and efficiency overflow rate

Grahic Jump Location
Fig. 11

Low-fidelity DOE for diffusor optimization showing diffusor efficiency ηD, projected area ratio Ap/Ap,stand, and inner wetted surface ratio S/Sstand. The index stand represents the respective value for the standard diffusor design. All results are colored by efficiency isosurfaces for an easier 3D visualization. The resulting Pareto front is shown with filled circles. It contains all optimal solutions when considering simultaneously as objective functions efficiency (to be maximized) and projected area and inner wetted surface (to be minimized). The initial, standard pick-up tube is shown as a filled triangle.

Grahic Jump Location
Fig. 12

Comparing diffusor efficiency for Pareto-optimal individuals computed with low- or high-fidelity approach

Grahic Jump Location
Fig. 13

High-fidelity simulations: comparing diffusor efficiency ηD with pump efficiency of a complete PTJ pump η. The filled symbol corresponds to the initial, standard pick-up tube.

Grahic Jump Location
Fig. 14

High-fidelity simulations: for the same individuals (i.e., Pareto members), the pump efficiency η and diffusor efficiency ηD are plotted versus the projected area ratio Ap/Ap,stand. Arrows indicate designs with increasing efficiency (top left) or diffusor efficiency (top right).

Grahic Jump Location
Fig. 15

High-fidelity simulations: for the same individuals (i.e., Pareto members), the pump efficiency η and diffusor efficiency ηD are plotted versus the inner wetted surface ratio S/Sstand. The arrow indicates designs with increasing efficiency and diffusor efficiency.

Grahic Jump Location
Fig. 16

Tangential velocity profile in rotor cavity at β = 90 deg and β = 270 deg downstream of the pick-up tube for Q = 15.88 m3/h and n = 3000 rpm. Theoretical velocity distribution for a forced vortex is shown as solid line. Optimal design corresponds to velocity plotted as filled square symbols (ηmax). The index “max” represents the respective value at the rotating casing.

Grahic Jump Location
Fig. 17

Characteristic curves for best design (ηmax), worst design (ηmin), and initial, standard configuration. Pressure head H is nondimensionalized by head at the BEP for the standard configuration. The same is done for the volume flow rate Q. The pressure head is shown with filled symbols and pump efficiency η with hollow symbols.

Grahic Jump Location
Fig. 18

Power curve for best design (ηmax), worst design (ηmin), and initial, standard configuration. Power and flow rate are nondimensionalized by respective value at BEP for standard configuration.

Tables

Errata

Discussions

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In