swirl velocity measurements for igt swirler concepts · • pitot probes are used in a variety of...

Andrew SislerUTSR Fellow 2016

Florida Turbine Technologies, Inc.561-427-6400

[email protected] © 2016 by Florida Turbine Technologies, Inc.

Swirl Velocity Measurements for IGT Swirler Concepts

Copyright © 2016 by Florida Turbine Technologies, Inc.

My background

2

BSME

MSME


OutlineIntroduction

Motivation

Literature Review

Approach & Setup

Results

Conclusions

3


Introduction

• Pitot probes are used in a variety of industries

– Aerospace– Weather stations– Maritime usage

• Used for measuring air speed• From Bernoulli’s principle, velocity is found if

static & stagnation pressure are known

𝑃𝑃𝑜𝑜 = 𝑃𝑃𝑠𝑠 +ρ𝑉𝑉2

2

4

Flytime.ca


Introduction• Sensitivity to environmental and geometric

applications

• On aircraft, far away from lifting surfaces or plane body—avoid boundary layer/local changes in flow due to fluid/structure interaction

• Icing, clogging (from dirt, debris, sand, etc.) can have devastating impact on effectiveness/accuracy

• Potential cause for deadly accidents5

Aviationweather.ws

Simhq.com


Introduction• Standard 1-hole probes only measure mean

velocity (1 direction)• Multi-hole probes capable of measuring

multiple velocity components– 1 hole: 1 component (Vx)– 3 hole: 2 components (Vx,Vy)– 5 hole: 3 components of flow (Vx,Vy,Vz)– >5 holes: highly 3D flow field with high angle variation (60° and

above)

6

1. Shaw-Ward, et al.Aeroprobe


Motivation

• Swirl-stabilized combustors– Common method for stabilizing combustors

• Recirculation zone is created, anchors the flame• Majority of gas turbine combustors implement high

swirl (HS) number combustors (S > 0.6)• Low swirl (LS) number criteria of

0.4 < S < 0.55• Research has shown that

low swirl number designs can reduce emissions/operate stably over wide range of conditions

7

2. Meadows et al.


Motivation• Low swirl number designs can burn very lean • Lean blowout limit (LBO) capability improved over high

swirl designs– Stresses reduced, decreasing non-uniform heat release

• Reduction in emissions for LS vs. HS• Although counter-intuitive, creates more stable flame• Create divergence in flow field, where mean air speed

matches turbulent flame speed

83. Cheng et al.


Literature Review• Some patents and licensing exists

– Maxon Corporation for ultra-low NOx industrial market– Patent from Lawrence Berkeley Lab

• Low swirl injectors tested in a Solar Turbines T70 gas turbine –converted their high swirl number injector to low swirl number

• Difference in residence times

2.5X NOx reduction

9

4. Cheng et al.


Literature Review• Attractive for industrial burners due to lower

emissions, high turndown ratio• Lifted flame keeps burner cool to touch• Normalized shear stresses

10

Divergence ZoneRecirculation Zone

5. Cheng et al.


Approach

11

• Measure swirl components with 5-hole conical Aeroprobe™ pressure probe

• Map swirler’s flow field as precursor to hot fired combustion testing

• Confirm flow exhibits expected behavior as CFD

• Realize optimal design for combustor performance


Setup

• Utilize FTT’s flowbench for testing

12

~ 5” ∅

Chimney

End of diffusing section

Straight 5-hole probe0.063” ∅

Traverse system

Ps

• Mount and position traverse system for taking measurements1/32” intervals

• Create mounting system for probe

FTT Air Flow Bench Schematic

50 HP Screw Compressors

37 Deg. Dewpt. Air Dryer

1320 Gallon Air Storage (120 psig)

0.5 Micron Coalescing

Filter

Manual Valves

To Sys. “B”

ER 3000 Controller

Sonic Nozzle or Venturi

Burst Disc (typ 20 psig)

MANIFOLD

System “A”

T1 and P1 P2 P3Test Item

P4/ptap

Specifications120 PSIG, 0.5 micron filtered air @ 37 deg. F dewpoint dryness.

Max Flow Rate is 0.48 lbm / sec (pps) or 400 SCFM.

ER3000 controls P2 to within ~ 0.002 psig.

Can Measure down to 0.0003 pps.

Flow bench inaccuracy is <1%

SS Plenum or “Cherry-Bomb”

Mechanical Room

Inlet air from compressor

Flow from compressors


Setup – Exploded View

13

Swirl velocity experiments conducted at ambient conditions only


Setup

14


Setup• Use the switching board designed at FTT to switch between pressure readings (large

cost savings)• Using handheld barometer to measure pressures in each port• 3 DOF traverse system for mapping flow field• Experimental coefficients can then be interpolated to find flow angles• Probe measurements valid in flow field that is within 5 m/s to Mach 2.0 range• Probe is accurate within ± 1°• Handheld barometer accurate to 0.025%• Probe’s measured velocity magnitudes within <1% or 1 m/s (whichever is larger value)

15

6. Lee et al.


Setup• Calibrated in a uniform flow field with known velocity• Create a map with even spread of flow angles

16

𝐶𝐶𝑝𝑝,𝑦𝑦𝑦𝑦𝑦𝑦 =𝑃𝑃2 − 𝑃𝑃3𝑃𝑃1 − �𝑃𝑃

Swirler Vane Exit

Static Pressure Tap

Diffuser Section


Setup• Structural analysis of probe mounting fixture (fluid-structure interaction)• Vortex shedding frequency vs. natural frequency of mounting structure• Ensure resonance does not occur and damage the probe

• St = f∗LU

• f ~ 77 Hz• Probe mounting estimated as simplified beam structure

• ω = 1.8752 E�I�gρ�A�L4

• f = 542 𝐻𝐻𝐻𝐻• Mass flux ratio (between swirler and perf plate flow) is critical

• m = 𝑚𝑚𝑐𝑐𝑚𝑚𝑠𝑠

17


Results – Velocities• 3-Component velocity results for each configuration• Help determine what downstream plane the flame will anchor at

(based on axial velocity)

18

W3 = 0.12854 lbm/sP3 = 1.4836 psigP4 = 0.0519 psig(PR = 10%)

Axial

RadialTangential

PN94IR-010-002


Results• Measurements taken on 1/32” intervals• Probe diameters and separating wall marked on data

19

W3 = 0.12854 lbm/sP3 = 1.4836 psigP4 = 0.0519 psig(PR = 10%)

PN94IR-010-002


• Increase vane height so that measurements can be taken within swirl passages, but away from wall

• Literature suggests anywhere from 2 – 5 probe diameters away from wall is sufficient

• Measurements taken every 1/16”

Results

20

PN94IR-010-010


Data Reduction

• Probe calibration valid in ±60° window• Starting at ±40°, standard calibration technique breaks

down• Need for various normalization schemes to reduce

data• Sectoring method is successful for flow angles in

fringes

21


Data Reduction – Baseline Design

22

Traditional Method PN94IR-395-012



23

Sectoring MethodPN94IR-395-012



24

Traditional MethodPN94IR-395-012



25

Sectoring MethodPN94IR-395-012


Results

26

• Importance of vane-to-vane-positioning for the probe


Results

27

• Flow variation across vane profile• Flow angles can vary significantly depending on

position relative to vane trailing edge


Conclusions

• For low swirl swirler - mass flow through core (perf plate) is too low, so that velocities are below specified range for probe (5 m/s to ~ 700 m/s)– Increase PR so that velocities are within range

• Vane height is too small for ample traversing area• Wall effects are/may be present throughout much of

vane section• Swirl angle is most accurately measured close to vane

exit• Swirled flow diffuses significantly as downstream

distance increases (addition of α flow component)

28


• Streamlines diverge, flow spreads at diffuser section

Conclusions

Reduction in radial swirl angle near wall, flow spreading outboard (αcomponent)

29

PN94IR-395-001


• Low velocities shown to be leading cause for unsteady & unreliable readings

• High Swirl perf plate flow within bounds (V ~ 17 ft/s)• Low Swirl perf plate flow out of bounds (V ~ 7 ft/s)

Conclusions

30


Conclusions

31

• Effect of core flow on swirl angle is significant – effect of aerodynamic wall• Careful consideration towards design target accounting for core flow effect in design

tool• Influence of vane height, combined with lack of perf plate flow, increases overall swirl

angleNO Core Flow Core Flow + Swirler Flow

PN94IR-010-010


Conclusions

32

Cost savings by replacing expensive commercialized software

PN94IR-010-009


• Reasonably good error limits, within a couple of degrees• Error includes propagation of pressure measuring device, and

intrinsic error of probe

Conclusions

33


• Permitting velocities and swirl angles are within range, swirl angles and velocity components can be obtained

• Effective diagnostic tool for pre-hot fired testing• Flow field can be mapped throughout different regions of the swirler• Surface finish differences between wax printed & SLA

• SLA Metal parts• Wax printed Casting or DMLS

• Hone in on MATLAB’s interpolation techniques• Use of non-invasive diagnostics (LDV, PIV) to measure velocities• Validate flow behavior with CFD

Overall Takeaways – Future

34


1. Shaw-Ward, S., Titchmarsh, A., Birch, D.M., “The Calibration and Use of n-hole Velocity Probes.” AIAA.

2. Meadows, J., Agrawal, A.K., “Time-Resolved Particle Image Velocimetry Measurements of Nonreacting Flow Field in a Swirl-Stabilized Combustor Without and With Porous Inserts for Acoustic Control.” ASME Journal of Engineering for Gas Turbines and Power.

3. Marc Day, Shigeru Tachibana, John Bell, Michael Lijewski, Vince Beckner, and Robert Cheng, “A combined computational and experimental characterization of lean premixed turbulent low swirl laboratory flames. i. methane flames”. Technical report, Lawrence Berkeley National Laboratory, Berkeley, CA, USA, 2012.

4. Cheng, R. K., (2006), “Low Swirl Combustion. DOE Gas Turbine Handbook”.

5. Cheng, R. K., "Ultra-Clearn Low-Swirl Burner for Heating and Power Systems." ARPA-e Workshop.

6. Lee, S.W., Yoon, T.J. “An investigation of wall-proximity effect using a typical large-scale five-hole probe” KSME International Journal.

References

35

swirl velocity measurements for igt swirler concepts · • pitot probes are used in a variety of...

Documents