Optical diagnostics of soot formation in low pressure laminar premixed flames

Abstract

Soot particles generated during combustion processes enhance heat transfer by thermal radiation. These particles are harmful to human health; therefore, control of soot emissions from combustion systems and mitigation of their negative effects is highly desirable. Soot particle formation is a complex process involving many processes. These are nucleation, surface growth, coagulation, aggregation, and finally oxidation. These processes are not completely understood. The aim of this research was to apply optical diagnostics as a tool to better understand the inception phase and surface growth of soot particles in low pressure premixed laminar flames. The work described in this thesis is based on quasi-one-dimensional, premixed C2H4–air (plus other additives) laminar flames, stabilised on a McKenna burner. Three different flame settings were used to study the dependence of soot particle formation on pressure variation in the range of 48–27 kPa. Two flames were at stoichiometric ratios, at phi (Φ) of 2.1 and 2.3. The third flame, at Φ of 2.1 and pressure of 40 kPa, was chosen to evaluate the effect of gas additives on the soot formation process. Three gas additives to ethylene base flame (C2H4-air) were used. These gas additives are argon (Ar), nitrogen (N2) and carbon dioxide (CO2). Laser-induced incandescence (LII) was used to carefully measure the spatial profile of the soot volume fraction (fv). Spatially resolved emission spectroscopy was then utilised to measure two key radicals (CH* and C2*) and to verify the location of the flame front (yff) and soot particle temperature (Ts). Probe thermocouple was employed to measure gas temperature (Tg), while Laser Induced Fluorescence (LIF) was used to record the Polycyclic Aromatic Hydrocarbons (PAHs) with 2 – 3 rings (2-3R), 3 – 4 rings (3-4 R) and >5 rings (>5 R). The gas velocity (v) was modelling by using the Ansys-Fluent software package. The time (t), at each axial location was calculated in a stepwise fashion, based on the modelled velocity profile. This helps to compute the soot surface growth rate and the phenomenological removing rates of PAH (2-3R) and PAH (3-4R). From Φ of 2.1 and 2.3 flames at different pressure settings, it was found that the thickness zone for CH*, used as an indicator of the flame front, was larger than for C2*. Furthermore, it was observed that the distance between the maximum recorded intensity of CH* and C2* decreased linearly with increasing pressure - with a slope of 25 × 10–9 ± 0.062 × 10–9 (mPa–1) and 28 × 10–9 ± 0.048 × 10–9 (mPa–1) for Φ of 2.1 and 2.3, respectively. It was found that the lowest value of fv was 0.0003 ppm, observed at a spatial location of 6 mm away from the burner surface. It was also observed that fv scales with pressure following a simple power function of the form fv = kPrn, where k is a scaling factor and n was measured at a value of 2.15 ± 0.7 and 1.5 ± 0.4 for Φ = 2.1 and Φ = 2.3, respectively. The analysis of soot particle surface growth pointed to a soot growth rate constant, kSG, of 20 s-1 for Φ of 2.1, whereas at Φ 2.3 the values of kSG was found to be 32 s-1, 25.13 s-1 and 12.11 s-1 for pressures of 27 kPa, 32 kPa and 35 kPa, respectively. This indicates that kSG has a weak dependence on the pressure and equivalence ratio. The measured values of Tg and Ts aligned well, with less than70 degrees difference between the two. The values for Tg that were measured from the first recorded soot particles were ~1465 ± 66 K. This was termed the ‘soot inception temperature’. The spatially phenomenological removing rate of PAHs with 2 – 3 rings (𝑘𝑝ℎ𝑒𝑛2−3𝑅) and 3 – 4 rings (𝑘𝑝ℎ𝑒𝑛3−4𝑅) were measured as 24.61 s–1 and 21.64 s–1, respectively, at Φ of 2.1 and pressure of 40 kPa. At a pressure of 27 kPa, the spatially phenomenological removing rate constants were measured as 15.29 s–1 and 18.26 s–1 for PAHs with 2 – 3 rings and 3 – 4 rings, respectively. This indicates that at a pressure of 40 kPa, (𝑘𝑝ℎ𝑒𝑛2−3𝑅) is faster than (𝑘𝑝ℎ𝑒𝑛3−4𝑅)by a factor of 1.14, whereas at a pressure of 27 kPa, (𝑘𝑝ℎ𝑒𝑛2−3𝑅) is faster than (𝑘𝑝ℎ𝑒𝑛2−3𝑅)by a factor of 1.19. At Φ = 2.3 and pressures of 40 kPa, (𝑘𝑝ℎ𝑒𝑛2−3𝑅) and (𝑘𝑝ℎ𝑒𝑛3−4𝑅)were found to be 23.33 s–1 and 16.9 s–1, respectively. This indicates that (𝑘𝑝ℎ𝑒𝑛2−3𝑅) is faster than (𝑘𝑝ℎ𝑒𝑛3−4𝑅)by a factor of 1.13 under these flame conditions. At Φ of 2.1 and pressure of 40 kPa and with respect to C2H4-air, it was found that fv decreased after addition of N2, CO2 and Ar. Ar was found to be the most effective additive for reducing fv, and increasing the soot surface growth rate constant. The soot surface growth rate constant (kSG) was calculated to be 8.3 s–1, 15.35 s–1, 35.65 s–1 and 60.35 s–1 for C2H4-air, C2H4-air:N2, C2H4-air:CO2 and C2H4-air:Ar, respectively. However, it was found that fv was reduced significantly in the presence of additives. The values of (𝑘𝑝ℎ𝑒𝑛2−3𝑅)were measured as 14.1 s–1, 20.58 s–1, 7.8 s–1 and 11.2 s–1 for C2H4-air, C2H4-air:N2, C2H4-air:CO2 and C2H4-air:Ar, respectively; whereas the values of (𝑘𝑝ℎ𝑒𝑛3−4𝑅) were measured as 10.3 s–1, 14.53 s–1, 4.7 s–1 and 3.9 s–1 for C2H4-air, C2H4-air:N2, C2H4-air:CO2 and C2H4-air:Ar, respectively. It was also observed that in these flames, the initial detection of the soot particles took place at a temperature of 1458.52 K, 1414.51 K, 1406.21 K, and 1377.16 K for C2H4-air, C2H4-air:N2, C2H4-air:CO2 and C2H4-air:Ar, respectively.