13th Annual Weimar Optimization and Stochastic Days 2016

Conference for CAE-based parametric optimization, stochastic analysis and Robust Design Optimization

Optimization of an Axial Pump using

CFturbo, PumpLinx & optiSLang

Dr. Gero Kreuzfeld, Sebastian Jakisch

CFturbo Software & Engineering GmbH
 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

Content
1. Motivation
2. Software tools
a) CFturbo
b) PumpLinx
3. Parameters
4. Optimization workflow
5. Results
6. Conclusion, prospects

WOST 2016 CFturbo Software & Engineering GmbH page 2/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

1. Motivation
Turbomachinery design is
Complex
Time-consuming
Expensive

PumpLinx
optiSLang
Meshing +
CFD/FEA Simulation

Conceptual
Iterative loops required Product
Design

CFturbo CAD + Prototyping

+ Testing

WOST 2016 CFturbo Software & Engineering GmbH page 3/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

2. Software tools a) CFturbo

Modern turbomachinery
design software for
Pumps
Ventilators
Compressors
Turbines
Covers all major types:
Radial
Mixed-flow
Axial
Available components:
Impeller
Volute
Vaned/vaneless stators
Analytical design theory &
empirical correlations Fully parametric geometry model of machines

WOST 2016 CFturbo Software & Engineering GmbH page 4/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

2. Software tools a) CFturbo

Several design steps from design point to 3D
geometry
All parameters can be accessed via batch
mode per XML file
optiSLang integration available

WOST 2016 CFturbo Software & Engineering GmbH page 5/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

2. Software tools b) PumpLinx

CFD system with high solver speed, especially for
fluid systems with rotating/ sliding components

Geometry model with

pt Q

inflow region impeller outflow region

Mesh density after analysis:

292 000 nodes
200 000 cells
Solver speed accuracy
Binary tree mesh
WOST 2016 CFturbo Software & Engineering GmbH page 6/ 19
 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

2. Software tools b) PumpLinx

Additional convergence checks Relative velocity vectors
Total pressure pt
oscillation Checked iterations
Allowed

Final value

Iterations

Pressure colored
streamlines
WOST 2016 CFturbo Software & Engineering GmbH page 7/ 19
 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

3. Parameters
zTE,S
# Parameter Reference Minimum Maximum zLE,S
1 dH1 = dH2 176 mm 140 mm 210 mm Shroud dS2
Main
2 dS1 = dS2 584 mm 467 mm 700 mm dS1
dimensions
= dH1/ dS1 0.30 0.20 0.45

3 z 204 mm 160 mm 320 mm

4 zLE,H* 0.1 0.2 0.1

Meridional
contour 5 zLE,S* 0.2 0.02 0.4 Hub
6 zTE,H* 0.9 0.8 0.9
zLE,H dH2
dH1 zTE,H
7 zTE,S* 0.9 0.6 0.98
z

Simplifications: r
Hub and Shroud (Tip) axis-parallel z Meridional view
Straight meridional leading and trailing edge

WOST 2016 CFturbo Software & Engineering GmbH page 8/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

3. Parameters B2,H
# Parameter Reference Minimum Maximum

Blade
8 nBl 3 2 6
properties

9 tLE,S 0 -25 25 mB2,H

10 tTE,S 80.3 64.25 96.37

11 tTE,H 80.3 64.25 96.37

Mean lines 12 mB1,H* 0.333 0.1 0.4 mB1,H mB2,S

13 mB1,S* 0.166 0.1 0.4
B1,H mB1,S
14 mB2,H* 0.718 0.6 0.9 Leading edge
15 mB2,S* 0.773 0.6 0.9 tLE,S
tTE,H
tTE,S
Simplifications:
Free vortex velocity distribution m
Automatic calculation of blade angles t Mean lines
B1 (shock-less inflow), B2 (Euler equation)
WOST 2016 CFturbo Software & Engineering GmbH page 9/ 19
 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

4. Optimization workflow
Design point
Flow rate Q = 1.476 m/s
Total pressure difference pt = 0.466 bar (H = 4.755 m)
Rotational speed = 780 /min
Water, no pre-swirl

Objective
Max. hydraulic efficiency

Constraints
B2 < 90
Total pressure difference pt 10%

WOST 2016 CFturbo Software & Engineering GmbH page 10/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

4. Optimization workflow

Meshing, simulation, Check for meshing

PumpLinx convergence check, and simulations
interpretation problems

Variation of Geometry update Check log for B2 < 90?

CFturbo geometry & export warnings and
parameters (*.spro, *.stl) errors

WOST 2016 CFturbo Software & Engineering GmbH page 11/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

5. Results Some geometry examples

WOST 2016 CFturbo Software & Engineering GmbH page 12/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

5. Results Sensitivity analysis/ Number of samples

Advanced Latin Hypercube Sampling

Samples
100 200 300
CFturbo 15 32 46
Failed
PumpLinx 1 3 9
Valid 84 165 245
samples
Reduced 74 139 233
Q 83.6 % 86.6 % 89.1%
pt 85.8 % 90.6 % 88.4 %
CoP
P 91.5 % 92.9 % 94.0 %
57.4 % 66.5 % 73.6 %

WOST 2016 CFturbo Software & Engineering GmbH page 13/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

5. Results Sensitivity analysis/ : CoP=74%

Coefficient of Prognosis CoP

hydraulic efficiency (objective)

Q Flow rate mB2,S
Q pt
= pt Total pressure difference dS1
P P Power consumption zTE,S
tTE,H

Q: CoP=89% pt: CoP=88% P: CoP=94%

zTE,S zTE,S nBl

tTE,S nBl tLE,S
zLE,S tTE,S tTE,S
tLE,S zLE,S zTE,S
nBl tLE,S zLE,S
z z z
dS1 dS1 dS1

WOST 2016 CFturbo Software & Engineering GmbH page 14/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

5. Results Sensitivity analysis/ Metamodel of Optimal Prognosis MOP

Response surface for

pt = f(dS1, z)

pt: CoP=88%

z
dS1

WOST 2016 CFturbo Software & Engineering GmbH page 15/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

5. Results Sensitivity analysis/ Parallel coordinate plot

parameters objectives constraints

WOST 2016 CFturbo Software & Engineering GmbH page 16/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

5. Results Optimization

Simulation Simulation
Algorithm Samples
time time/ sample
EA 72.2 h
Evolutionary 330 13.1 min 69,9 % + 5.0 %
Algorithm (3.0 d)
ARSM 126.3 h
Adaptive Response 540 14.0 min 69,3 % + 4.4%
Surface Method (5.3 d)

Desktop PC
2 x Intel Xeon 3.07 GHz, 6 cores
64 GB RAM
Max. 2 parallel simulation jobs

WOST 2016 CFturbo Software & Engineering GmbH page 17/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

5. Results Optimization

Reference Reference ARSM

ARSM EA EA

WOST 2016 CFturbo Software & Engineering GmbH page 18/ 19

 Optimization of an Axial Pump using CFturbo, PumpLinx & optiSLang

6. Conclusion, prospects
Optimization process CFturbo + PumpLinx + optiSLang successfully tested
PumpLinx solver speed is beneficial for optimization and enables optimization on
desktop PCs
CFturbo initial design can be used as very reasonable starting point (pre-optimized) to
save optimization time
Number of parameters could be reduced to ~ 50% by sensitivity analysis
automatic parameter reduction in CFturbo possible in near future
Metamodel for efficiency seems to be not possible
direct optimization with reduced parameter set seems to be the best solution
Removing geometry simplifications is necessary
Empirical based performance prediction in CFturbo could be used to further reduce
simulation effort CFturbo included metamodel

Thanks to Dynardo (optiSLang) and Simerics (PumpLinx) for licenses and support!

WOST 2016 CFturbo Software & Engineering GmbH page 19/ 19