8. CFD Seminar “Power Generation”
Dimensioning, Design and Simulation of high-
performance centrifugal pumps with CFturbo and
Ansys CFX
Dr. Oliver Velde CFturbo® Software & Engineering GmbH
3.9.2009 CFturbo Software & Engineering GmbH Page 2/16
Content
Introduction 03
Example 04
Design process and meshing 06
Optimization 14
Summary 16
3.9.2009 CFturbo Software & Engineering GmbH Page 3/16
Design process overview
Dimensioning,
Design
CFturbo®
Grid generation
ICEM-CFD, TurboGrid,
Pointwise, Gambit, …
CAD
Catia, SolidWorks, UG-NX,
ProE, BladeGen, …
ProductionOptimization:
interactiv or automatic
CFD/FEM Simulation
ANSYS-CFX, CCM+,
CFdesign, FineTurbo, …
Measurement
Rapid Prototyping,
Validation
3.9.2009 CFturbo Software & Engineering GmbH Page 4/16
Example high-performance centrifugal pump
of a cooling cycle in a power station
Cooling pump
This picture is based on Kohlekraftwerk.svg from the free media
database Wikimedia Commons. It has been published under the terms
of the GNU Free Documentation License. The copyright holder of the
picture is Kolossos.
3.9.2009 CFturbo Software & Engineering GmbH Page 5/16
Volume flow = 1500 m3/h
Head = 140 m
Speed = 3.000 min-1
• Main dimensions
• Meridional contour
• Blade design
• Volute design
Stage Design
Example high-performance centrifugal pump
of a cooling cycle in a power station
Pump has to fit optimally to the
system
Design goal
Operating point
Start with
System characteristic
Pump characteristic
(n = const.)
Operating point
3.9.2009 CFturbo Software & Engineering GmbH Page 6/16
Material data,
operating point,
parameter
Main dimensions
Design process with CFturbo: Impeller design
3.9.2009 CFturbo Software & Engineering GmbH Page 7/16
Hub & Shroud
Leading & Trailing edge
Design process with CFturbo: Impeller design
Meridional contour
Blade shape
Velocity triangles
Leading & Trailing edge
angle
Blade properties
b2
Blade angle distribution
Wrap angle
Mean lines
Blade thickness
Shape of Leading &
Trailing edge
Blade profiles
b1
Mean lines
3.9.2009 CFturbo Software & Engineering GmbH Page 8/16
Complete design:Impeller and Volute
Stage
Design process with CFturbo: Volute → Stage
3.9.2009 CFturbo Software & Engineering GmbH Page 9/16
Preparation of Model and
Geometry
Direct Export to ICEM CFD
Design process with CFturbo: Pre-Processing
3.9.2009 CFturbo Software & Engineering GmbH Page 10/16
Meshing Parameter Dialog
Meshing with ICEM CFD
Automatic, script-based meshing
complete parameter setup in CFturbo
Design process with CFturbo: Pre-Processing
Tetra Mesh with Prism Layers
Meshing
Design and meshing of an entire pump
stage within less than ½ hour
3.9.2009 CFturbo Software & Engineering GmbH Page 11/16
Inlet pressure
Volume flow
Steady Simulation
Frozen Rotor
Turbulence Model: SST
Simulation Setup Goals
• Fast determination of characteristic
• As many simulations as necessary,
but as few as possible
• Comparing two or more designs
• head
• efficiency
• Characteristic’s range
Goal, Model, Boundary conditions
Simulation Setup
Simulation
3.9.2009 CFturbo Software & Engineering GmbH Page 12/16
PInlet = 0
Determination of boundary conditions
Simulation strategy
Boundary conditions for the
complete characteristicPerformance estimation
Simulation
Volume flow(outlet B.C.)
He
ad
3.9.2009 CFturbo Software & Engineering GmbH Page 13/16
Characteristic, pressure
distribution, velocities
Results
Post-Processing
0,00
40,00
80,00
120,00
160,00
200,00
800 1000 1200 1400 1600 1800
H [
m]
Volume flow [m3/h]
Operating point
failed by 10%
3.9.2009 CFturbo Software & Engineering GmbH Page 14/16
Optimization
Design Geometry & Mesh
SimulationPost-Processing
Design goal
First design:
Examination:
Determination of para-
meter and target
function for optimization
Performance prediction:
Estimation of the influence o the
impeller size based on empirical
correlations
Target function:
Head at operating point
Parameter: Impeller diameter
d = 363 mm
d = 380 mm
3.9.2009 CFturbo Software & Engineering GmbH Page 15/16
0,00
40,00
80,00
120,00
160,00
200,00
800 1000 1200 1400 1600 1800
H [
m]
Volume flow [m3/h]
Optimization
Parameter change
First design,
Target function & Parameter
Simulation & Post-
Processing
Evaluation:
| Actual — Target| < ε ?
Geometry & mesh with
ICEM-Scripts
usually 3..5 loops
d = 363 mm
d = 380 mm
3.9.2009 CFturbo Software & Engineering GmbH Page 16/16
► Parametric/semi-automatic design of radial and mixed-flow turbomachines
► Stabile process of performance prediction
► Complete process can be run in one batch
Rapid-Design-Process with CFturbo® and ANSYS ®
Summary