Measurement of Temperature, Soot Diameter and Soot Volume Fraction in a Gulder Burner

Transcription

1 Department of Engineering Science University of Oxford Measurement of Temperature, Soot Diameter and Soot Volume Fraction in a Gulder Burner Huayong Zhao, Ben William, Richard Stone Project Meeting in Cambridge, 17 st May 1

2 Outline Motivation Introduction to CBT-TCS Experimental Setup Accuracy Improvement Techniques Sample Data and Assumptions Conclusions

3 Motivation Sub-micron particulate matter is of concern to human health Legislation places limits on the particulate emissions For model validation, a simple flame is needed Spatially resolved non-intrusive measurement is needed of: Flame temperature Soot volume fraction Soot particle diameter 3

4 y (cm) z (cm) z (cm) z (cm) Introductions to CBT-TCS (Cone Beam Tomographic Three Colour Spectrometry) D Tomography + Three Colour Pyrometry Temperature (K) y (cm) Diameter (micrometer) y (cm) Soot Volume Fraction (ppm) 5.5 Optical diagnostic techniques used to measure the 3D temperature, soot diameter and soot volume fraction distributions Combination of three colour pyrometry and 3D tomography technique 1.5

5 y (cm) z (cm) z (cm) z (cm) Introductions to CBT-TCS (Cone Beam Tomographic Three Colour Spectrometry) D Tomography + Three Colour Pyrometry Temperature (K) y (cm) Diameter (micrometer) y (cm) Soot Volume Fraction (ppm) 5.5 Optical diagnostic techniques used to measure the 3D temperature, soot diameter and soot volume fraction distributions Combination of three colour pyrometry and 3D tomography technique 1.5

6 Three Colour Pyrometry - Strategy Calibration of the spectral response of the Camera Sensor 6

7 Emissivity & Scattering Models (1/) Energy balance (Kirchhoff s Law) Emissivity (ε) = Absorption efficiency factor (Qabs) Different scattering models Rayleigh-Gans Theory (X [x = πd/λ] <.3; T and fv); Rayleigh-Gans Theory with Penndorf extension (x <.8; T, D and fv) Mie Scattering Theory (analytical solution; T, D and fv) Hottel and Broughton Correlation (Empirical correlation; T, KL) 7

8 Error of the Absorption Efficiency Particle Diameter (nm) Absorption Efficiency Fraction of Scattering Emissivity & Scattering Models (/) RG Penndorf Mie RG Penndorf Mie Size Parameter RG Penndorf Size Parameter Size Parameter Size Parameter Wavelength (nm) 8

9 CBT-TCS - Strategy Thermal Radiation Law and calibrated camera spectral response curve Measure the projection and do the colour interpolation Using cone-beam tomography to reconstruct the 3D colour ratio distributions Construct the temperature and diameter look-up tables (entries: Red/Blue, Red/Green) 3D Temperature and diameter distributions Soot volume fraction Camera sensitivity factor and pixel value 9

10 Experimental Setup - Diagram Gulder Burner Fused Silica Window Camera Computer Labview Interface Control Box Optical Rail Supporting Column Fuel tank FUEL AIR MFC MFC Central Air Compressor 1

11 Experimental Setup - Photos 1

12 Computer Absolute Calibration sensitivity factor Burner Opaque Plate Calibrated Spectrometer Tracing paper Fused Silica Window Lens and filter Bulb Camera Optical Rail 11

13 Accuracy Improvement Techniques Downsampling (ratio: ) Trade-off between accuracy and spatial resolution: smaller pixel width smaller difference between two adjacent projections, if the difference is comparable to the background noise poor accuracy but good spatial resolution 3D Median Filter apply to the T, D, or KL data Blurs the image apply to the reconstructed RGB colour map Enhances the smoothness Increase the zero-padding length when using the fast Fourier transform so as to increase the resolution in the frequency domain Circumferential averaging for axi-symmetric flame Using optical filters to make more use of the dynamic range of different colour channels (especially blue) 1

14 Sample data ethylene co-flow laminar diffusion flame Test Conditions Ethylene flow rate: 19 ml/min; Air flow rate: L/min Camera setup: Frame rate: 6 frames/sec Aperture size: f/3 (with * converter) Focus length: 1 mm Object to lens distance: 3.6 cm Lens to detector distance: 17. cm Spatial resolution: 6 μm Optical components Lens: Nikon 5 mm lens with * converter Filter: LEE E81 filter (around. transmission efficiency for red and green light and.8 for blue light) 13

15 Transmission Sample data ethylene co-flow laminar diffusion flame Test Conditions Ethylene flow rate: 19 ml/min; Air flow rate: L/min Camera setup: Frame rate: 6 frames/sec Aperture size: f/3 (with * converter) Focus length: 1 mm.8 Object to lens distance: 3.6 cm Lens to detector distance: 17. cm Spatial resolution: 6 μm Optical components. Lens: Nikon 5 mm lens with * converter Filter: LEE E81 filter (around. transmission efficiency for red and green light and.8 for blue light) 1.6. LEE81.DX Wavelength (nm) 13

16 z (cm) z (cm) z (cm) 6 Results Mie scattering theory Temperature (K) 6 Diameter (nm) Soot Volume Fraction (ppm) T (K) at 3 mm Soot volume fraction (ppm) at 3 mm 8 RG 6 Penn Mie x (cm) x (cm) x (cm) Radial distance (cm) Radial distance (cm) T (K) at mm Soot volume fraction (ppm) at mm Radial distance (cm) Radial distance (cm) T (K) at 1 mm HB RG Penn Mie Radial distance (cm) Soot volume fraction (ppm) at 1 mm Radial distance (cm) 1

17 Temperature (K) Soot volume fraction (ppm) Temperature (K) Soot volume fraction (ppm) Comparison of Results with Snelling et al. [1] 3 mm 3 mm 1 1 CARS LOSA SST 8 SST 19 CBT-TCS CBT-TCS Radial distance (mm) mm 1 CARS SST 19 CBT-TCS Radial distance (mm) mm LOSA SST CBT-TCS Radial distance (mm) Radial distance (mm) [1] Snelling DR, Thomson KA, Smallwood GJ, Gulder OL. AIAA J. ; ;

18 Assumptions (1/) Particulate temperature is the same as local flame temperature Thermal radiation from other species is negligible compared to soot particles CO:.,.7,.3, 9., 1. and 15 µm HO: 1.38, 1.87,.7, and 6.3 µm Chemilluminescence from radicals is negligible except from the circumferential base of the flame 16

19 Assumptions (1/) Particulate temperature is the same as local flame temperature Thermal radiation from other species is negligible compared to soot particles CO:.,.7,.3, 9., 1. and 15 µm HO: 1.38, 1.87,.7, and 6.3 µm Chemilluminescence from radicals is negligible except from the circumferential base of the flame 16

20 Assumptions (/) The radiation attenuation along the optical path is negligible (the optical-thin approximation) Error is expected to be small because of the low soot volume fraction (~1 ppm) Can be partially corrected by using an iterative method suggested by Lu et al. (9) 17

21 Concluding Remarks The CBT-TCS technique has been successfully applied to the Gulder burner to measure the spatially distributed temperature, soot diameter and soot volume fraction for an ethylene laminar diffusion flame The comparison with the data reported by Snelling et al. shows consistent temperature and soot volume fraction profiles CBT-TCS has also been applied to the Santoro burner to investigate the soot formation in helium-diluted and hydrogen-added ethylene flames 18

22

23 COV of soot volume fraction Temperature (K) COV of the measured light intensity Standard deviation of temperature (K) Precision of the experiment.8 1 Red Green 1.6 Blue mm mm 3 mm 5 mm Height above the fuel tube exit (cm) 1 mm mm 3 mm 5 mm Temperature (K) RG RG-fitting RB RB-fitting Temperature (K) Colour ratio