Applying laser-induced incandescence (LII) technique to soot volume fraction measurements for pyrolyzing pulverized coal

Staged combustion and gasification of coal are conducive to soot formation. Presence of soot in combustion systems can affect heat transfer, pollutant formation, and unburned carbon levels.Development of an accurate soot production model requires high-quality experimental data. For this purpose, the laser-induced incandescence (LII) technique was applied for in-situ measurements of soot volume fraction in the vicinity of pyrolyzing pulverized Pittsburgh #8 coal particles passing through the center of a fuel-rich flat flame burner. A pulsed ND:YAG laser was used as the light source, while an intensified CCD camera captured the spatially-resolved two-dimensional distributions of soot volume fraction. Spatially-resolved measurements with good signal-to-noise ratios were made downstream of the gaseous flame sheet.

16th International IFRF Members' Conference, Boston, USA APPLYING LASER-INDUCED INCANDESCENCE (LII) TECHNIQUE TO SOOT VOLUME FRACTION MEASUREMENTS FOR PYROLYZING PULVERIZED COAL Shengteng Hu1, Dong Zeng2, and Hamid Sarv3 The Babcock & Wilcox Company Abstract Staged combustion and gasification of coal are conducive to soot formation. Presence of soot in combustion systems can affect heat transfer, pollutant formation, and unburned carbon levels. Development of an accurate soot production model requires high-quality experimental data. For this purpose, the laser-induced incandescence (LII) technique was applied for in-situ measurements of soot volume fraction in the vicinity of pyrolyzing pulverized Pittsburgh #8 coal particles passing through the center of a fuel-rich flat flame burner. A pulsed ND:YAG laser was used as the light source, while an intensified CCD camera captured the spatially-resolved two-dimensional distributions of soot volume fraction. Spatially-resolved measurements with good signal-to-noise ratios were made downstream of the gaseous flame sheet. 1. Introduction Soot is an important byproduct of fossil fuel combustion. A heavily sooting flame typically has a lower flame temperature due to the increased radiative heat flux. In addition, soot contributes to the unburned carbon in the fly ash. Soot formation could also alter the formation pathways of certain emissions (e.g., NOx). Thus, there is a great need to develop a model that can simulate the processes of soot formation, growth, and destruction. Laser-induced incandescence (LII) detection and measurement is an emerging technology that is capable of spatial and temporal determination of soot volume fraction and the primary soot particle size [1]. In this technique, a laser beam with a high power density is used to heat soot particles rapidly. The soot almost reaches its vaporization temperature (~4000 K) and incandesces. With appropriate calibration and analysis, the measured incandescence from soot particles can yield information about the soot volume fraction and the primary soot particle size. Theoretical investigations have shown that the LII signal is approximately proportional to the soot volume fraction [1]. Several research teams [210] have used this technique to determine the soot volume fraction and measure soot particle size in gaseous sooting flames. LII has also been applied in diesel engines to monitor particulate emissions [11-13]. However, there has been no reported application of the technique to particle-laden pulverized coal (PC) flames. This paper presents the developmental and application of the LII technique to soot measurements in PC flames. 2. Model development In order to distinguish the temperature histories of soot and char particles in a PC flame, their relative contributions to the LII signal must be determined via heat transfer modeling. Spatial effects can be neglected for small soot particles from a time scale analysis (<100 nm). Char particles on the other hand are 3-4 orders of magnitude bigger and as such, they can be simulated by a two-dimensional, axisymmetric, and time dependent model. The following sections provide more details. 1 Corresponding author, Research Engineer, The Babcock & Wilcox Research Center, [phone], [e-mail], 180 S. Van Buren Avenue, Barberton, OH 44203 2 Research Engineer, The Babcock & Wilcox Research Center 3 Director, The Babcock & Wilcox Research Center Page 2 of 15 2.1 Soot model This model is to capture the temperature history of soot particles under intensive laser heating. The theory and physical processes of LII was started by Eckbreth [14], summarized by Melton in 1984 [1] and developed further by several groups around the globe. Lots of theoretical bases for model developed herein are from the paper by the second LII workshop [15]. The model describes energy balance of the soot particle considering rapid heating by the laser fluence, sublimation, heat conduction to surrounding gas, and radiation heat loss as shown in Figure 1. The energy balance equation can be expressed as dU int ernal = q absorption − q conduction − q sub lim ation − q radiation dt (1) Details of the mathematical formula for each term in Equation 1 can be found elsewhere [15]. Under the Rayleigh approximation ( a << λ ), the equation takes the following form. ~) 2πa 2αT pref 4 3 dT 8π 2a 3E (m RT0 ⎛ γ * + 1 ⎞ ⎟(T − T0 ) ⎜ πa ρ s C s q(t ) − = 3 dt λ T0 2πWa ⎜⎝ γ * − 1 ⎟⎠ − ΔH v 4πa 2α M pv RT ⎛ RT ⎞ ⎜⎜ ⎟⎟ ⎝ 2πWv ⎠ 0.5 − ~) 192π 3D 3 (kT )5 E (m (2) h (hc )3 where the temperature dependent physical properties are are summarized in Table 1. Radiation Sublimation Internal Energy Change Heat Conduction Absorption Figure 1. Energy balance of laser heated soot particle C s [16] 2.90041× 103 − 36.4073× 103 1 T , J / kg ⋅K ρ s [17] ~ [18] m 2.3031 × 10 3 − 7.3106 × 10 −2 T , kg m 3 1.90 − 0.55 i αT 0.3 p v , atm exp − 122.96 + 9.0558 × 10 −2 T − 2.7637 × 10 −5 T 2 + 4.1754 × 10 −9 T 3 − 2.4875 × 10 −13 T 4 ( ) Wv , g mol 17.179 + 6.8654 × 10 −4 T + 2.9962 × 10 −6 T 2 − 8.5954 × 10 −10 T 3 + 1.0486 × 10 −13 T 4 ΔH v , J mol 2.05398 × 10 5 + 7.3660 × 10 2 T − 0.40713 T 2 + 1.1992×10−4 T 3 − 1.7946×10−8 T 4 + 1.0717 ×10−12 T 5 Table 1. Parameters for the LII model Debates exist on the choice of the complex refractive index of soot. Here we adopted a value reported by Lee and Tien [18]. Under the experimental conditions considered in this work (pressure = 1 atm and temperature > 1200 K), the mean free path of the surrounding gases greatly exceeds the particle size, and therefore the conduction cooling is in the free molecular flow regime. In Equation 2, γ * satisfy the following equation that takes into account the steep temperature gradient in the boundary layer surrounding the soot particles [19], 1 γ * −1 = 1 T − T0 T 1 ∫ γ − 1dT ′ T0 babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y (3) Page 3 of 15 The choice of the thermal accommodation factor is far from conclusive in the LII society. A value of 0.3 is selected based in this study. The molecular weight of the surrounding gases is assumed to be that of air. This assumption is known to introduce errors in the conductive heat loss term[19], but given the experimental condition where the major heat loss channel is sublimation, the error introduced to the overall LII signal prediction is minimal. The size of the soot particle is reduced by surface sublimation if temperature is high. The equation governs the size evolution is given by Wα p da =− v M v dt RTρ s ⎛ RT ⎜ ⎜ 2πW v ⎝ 0.5 ⎞ ⎟ ⎟ ⎠ (4) where the temperature dependent values of related parameters are given in Table 1. A forth order Runge-Kutta algorithm is utilized to solve the coupled Equations 1 and 4. The calculation is performed for 400 ns with a step size of 20 ps. The incandescence signal is calculated according to the Planck blackbody formula integrated over all solid angels as 2πc 2 h S LII = 5 ε λS 4π a 2 (5) λ S [exp(hc λ S kT ) − 1] Implicit in Equation 5 is that within the Rayleigh limit, the signal from single soot particles is proportional to the volume of the particle multiplied by a temperature term that only weakly depends on size. ( ) 2.2 Char model A 2-D axisymmetric model is also developed to simulate the incandescence signal from the much larger char particle under intensive laser heating. The specific heat of char particles including the 3 effects of ash can be calculated by weighted additivity as C = ∑w C j j , j=1, … 3 represents char, j =1 ash and moisture respectively. For char particles, the moisture content is zero, and specific heat of ash can be calculated by C 2 = 754 + 0.586 T . Specific heat of char can be the simulated based on Einstein formula g1 (z ) = exp(z ) C char = (R μ )[g1 (380 T ) + 2 g1 (1800 T )] [21], where ⎛ 5 ⎞ μ =⎜ yi μi ⎟ ⎜ ⎟ ⎝ i =1 ⎠ ∑ −1 and {[exp(z ) − 1]/ z}2 . The particle swelling ratio is assumed unity in this model. The true density of coal/char on a daf basis can be calculated based on the model proposed by Merrick [22]. In this model, the contributions from each elements are assumed to be additive, and therefore the density can be expressed as ρ t ,daf ⎛ =⎜ ⎜ ⎝ 5 ∑yγ i i i =1 ⎞ μi ⎟ ⎟ ⎠ −1 (6) where γ i , i = 1,…,5 is the coefficient for the corresponding elements, and their values are shown in Table 2. Including the effects of ash and moisture by assuming that the specific volumes are additive, the true density of coal/char is calculated by ⎛ ρt = ⎜ ⎜ ⎝ ⎞ wj ρ j ⎟ ⎟ j =1 ⎠ 3 −1 ∑ babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y (7) Page 4 of 15 where the densities of ash and moisture were taken as 3000 and 1000 kg/m3, respectively. Element C H O N S 3 γ , m /mol 5.3E-6 5.77E-6 3.46E-6 6.69E-5 3.84E-5 Table 2. Density coefficients A correlation was developed for calculating the thermal conductivity of coal/char [23]. It takes the following form k a = (ρ t 4511)3.5 T 0.5 (8) This correlation is capable of predicting conductivity over a wide temperature range for amorphous carbon, but only validated against experimental data for coal at room temperatures. The effect of ash is also neglected. The thermal conductivity is a relatively weak function of temperature. Heat transfer simulations Particle surface temperature was calculated using the Fluent software by considering laser heating, internal heat conduction towards the core, heat conduction to the surrounding air, radiation heat loss, and sublimation/vaporization heat losses of both carbon and ash in the char. Basic assumptions included the following: 1. Spherical particle and axisymmetric geometry as shown in Figure 2. 2. Uniform particle surface absorption efficiency 3. Uniform particle thermal conductivity, specific heat, and density 4. Complete coal devolatization to char q(t ) r z A θ a Figure 2. Schematic of the char particle laser heating geometry The above axisymmetric geometry was discretized with sufficient structured quadrilateral mesh cells including finer grids close to the surface where higher temperature gradients were expected. In this setup, a = 50 μm , and q(t ) , the laser power density, was presented in a lookup table as a function of time. The integrated laser fluence was 0.1 J/cm2. At the boundary ( r = a ), only half of the particle was exposed to laser radiation. Heat conduction to the surroundings was modeled by qcond = − k a (T − T∞ ) a . In the case of Knudsen conduction regime, an additional term 1 + G Kn would be added to the denominator. Initial particle and gas temperatures were T p = 1600 K and T g = 1600 K . Char particle heat absorption efficiency and emissivity were calculated based on the Mie theory [24], assuming a complex index of refraction m~ = 2.050 − 1.127 i . Given the particle size and laser wavelength (532 nm) used in this model, the calculated absorption efficiency was 0.74. Particle radiation emissivity is a function of temperature and it was evaluated by integrating over the entire wavelength at given temperatures [25]. However, for the temperature range in this study, the emissivity was nearly constant, and therefore a value of 0.76 was used in the model. babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y Page 5 of 15 The most important mineral contents in coal are kaolinite, quartz, pyrite, siderite, and gypsum. Together they amount to more than 70 wt% of the total mineral matter [26]. Upon heating, these minerals undergo decomposition reactions at different temperatures. However, at typical flame temperatures (> 1000°C), the end constituents are always SiO2 (highly crystalline cristobalite), Al2O3, and Fe2O3. Upon heating by the laser pulse, these constituents approach their boiling points and begin to vaporize. In this model, we assumed that SiO2 was the only constituent in the coal ash. Its heat of vaporization was obtained from the NASA equilibrium code. With these considerations, the char particle surface temperature during a 4.04 µs pulse at a laser fluence of F = 0.1 J cm 2 was calculated using the Fluent computer model. The calculation for char surface temperature using Fluent was performed for 2000 time steps with a step size of 20 ps, i.e. total of 40 ns, followed by another 2000 time steps with a step size of 2 ns, total of 4.04 µs. The laser fluence is F = 0.1 J cm 2 . 2.3 Comparison of signals The temperature of the node at the leading position (point A in Figure 2) on the char particle surface is plotted as a function of time shown in Figure 3. Figure 3. Computed particle sufrace temperature history at point A (Figure 2; The laser pulse is also shown on an arbitrary scale to indicate the time sequence). F = 0.1 J cm 2 ; d = 100μm 3 With an assumed density of 2000 kg/m , the equivalent layers of carbon and SiO2 that would be removed from a 100 µm char particle were 0.0003 and 81 nm, respectively. From the temperature profiles in Figure 3, the relative contributions from the char and soot particles to the overall LII signal can be evaluated. Correct interpretation of the LII signal can only be made when the contribution from char particles is negligible compared to that of soot particles. Equation 5 gives the temperature dependent emission for individual particles. The total emission per unit volume can be calculated by the following equation t 0 + t exp m& ⋅ y ρ (9) S LII ,total = n S LII dt , where the number density n = 1 t0 πd 3 ⋅ V& 6 ∫ The number densities of soot and char can be calculated with their respective properties. Combining Equations 5 and 9, the soot to char signal ratio is given by R LII = ~) ρ 4πd char E (m char λ S ε char ρ soot ∫ ∫ t0 + texp t0 t0 + texp t0 1 [exp(hc λ S kTsoot ) − 1]dt 1 [exp(hc λ S kTchar ) − 1]dt y soot y char (10) In Equation 10, the char and soot particle diameters are assumed constant, which is always valid for char particles and also justified for soot particles under low to moderate laser fluence. For particles with known temperature histories, the signal ratio becomes only a function of the mass fraction ratio of soot and char. Figure 4 illustrates this point by plotting the effects of soot-to-char mass ratio on babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y Page 6 of 15 the corresponding LII signal ratio for different soot particle sizes. When generating this plot, it was assumed that the char surface temperature was uniform. As one can see, the contribution from char particles only becomes important when the soot yield is three orders of magnitude lower than that of char in the sampling volume. Furthermore, this conclusion is not affected by the size of the soot particles as shown in Figure 4. Figure 4. The soot-char LII signal ratio as a function of their mass fraction ratio 3. Experimental setup A schematic of the LII setup is shown in Figure 5. A frequency doubled Nd:YAG laser is used as the light source (Continuum Powerlite 9010, 532 nm). A zero order waveplate on a rotatable mount followed by a thin film plate polarizer mounted at its Brewster angle is used to adjust the beam energy continuously. After passing through an aperture for spatial filtering, the laser beam is shaped into a 20.1 × 1.2 mm rectangular sheet with its long edge aligned in the vertical direction using a set of beam shaping optics detailed in Figure 5. The beam is first expanded in the vertical direction using a negative cylindrical lens (-100 mm f.l.), and then a 12 × 1.24 mm fixed slit passes only the portion of the beam with relatively uniform intensity. After being collimated by a positive cylindrical lens (+500 mm f.l.), the slit is imaged to the test section the light by passing through another +200 mm f.l. cylindrical lens. The effect of diffraction is minimized using this setup. 2X Nd:YAG Beam Dump Mirror +200 mm cyl. Polarizer Image the slit Beam in horizontal Slit Dump direction Aperture 12mm×1 Burner .24mm Waveplate -100 mm cyl. Computer +500 mm cyl. Camera Controller Filter ICCD DG535 Figure 5. LII setup for soot volume fraction measurement using LII (2D) The two-dimensional incandescence signal from the soot particles is collected at 90 degrees to the path of the laser beam using an intensified CCD camera (PI-MAX2, 1024 × 1024 pixels). The incandescence images are relayed into the ICCD camera via a 105 mm achromatic macro lens (Coastal Optical). Each pixel in the camera corresponds to a 29 × 29 µm (H × W) area in the test section. The camera readout is binned in the vertical direction by 4 pixels to improve the signal to noise ratio, which corresponds to a resolution of 115 × 29 µm. A narrowband filter (440 ±10 nm) is installed in front of the camera lens to select light signal that is relatively interference free. The camera intensifier is triggered 50 ns after the trailing edge of the laser pulse to avoid or reduce interferences from laser babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y Page 7 of 15 induced fluorescence (if any). The exposure duration is set to 100 ns for optimum signal intensity. A 3000-pulse average is used to minimize signal fluctuation due to unsteady coal feeding. At each condition, the tests were repeated at least 4 times. As demonstrated by other researchers (e.g. [4]), there exists a region where the LII signal is relatively insensitive to the variations of the laser fluence. An optimum fluence value of 0.087 J/cm2 (~21 mJ/pulse) was used in this work. 3.1 Flat Flame Burner (FFB) The Flat Flame Burner (Technologies for Research, Inc.) incorporates a hastalloy 50.8 × 50.8 mm, honeycomb burner surface supporting stainless steel hypodermic fuel tubes. Fuel supply is directed to the burner surface through the hypodermic tubes which are individually sealed from the oxidizing gas that flows around the fuel tubes and through the honeycomb to mix with the fuel at the burner surface. A central 1.33 mm ID coal particle feeding tube is installed along the flat flame burner centerline. A removable quartz viewing tower shields the flame from air entrainment. The feeder is similar to the one used by Ma [27]. It consists of a stepper-motor-driven syringe loaded with pulverized coal particles. The computer-controlled motor translates the plunger at a predefined speed which pushes the coal particles into a vertically mounted funnel from a side port. A carrier gas enters from the top of the funnel to transports coal particles into the central feeding tube. A CO-H2-N2 mixture is used as fuel and an O2-N2 mixture is used as oxidizer to generate fuel rich flames. Nitrogen dilution was varied to achieve flame temperatures in the 1000 to 1500 C (with 100 C intervals). The equivalence ratios for the 1100 to 1500 C flames were maintained around 1.70 but it had to be raised to 2.26 in order to produce a 1000 C flame. For each case, temperature profiles along the burner centerline axis were measured using a coated type-B thermocouple and the values were corrected for radiation heat loss. 3.2 Calibration Factor The proportionality factor for calculating the soot volume fraction from the LII signal can only be determined through calibration. We applied the conventional light extinction technique on an axisymmetric ethylene diffusion flame for this purpose. Ethylene was fed through the central tube of the FFB with air flowing through the honeycomb mesh. The light extinction technique was carried out using a pulsed Nd:YAG laser to obtain absolute volume fraction. The laser beam was focused to a ~135 µm diameter with pulse energy of ~10 µJ. This low energy ensures that there was no soot sublimation due to laser heating, which could alter the measurement results. A pyroelectric joule meter (Ophir PE9) was used to measure the laser energy after passing through the sooting flame. The FFB is mounted on a two-dimensional (x and z) translation stage with 0.01 mm resolution controlled by a computer. The laser beam passed through the calibration flame at a height above burner (HAB) of 50 mm. During the calibration measurement, the burner was moved 0.1 mm at a time. At each location, the energy of 300 laser pulses was registered on the joule meter and results were averaged. The calibration factors for LII can be determined by C = S LII f v (11) where, S LII is the signal detected when applying the LII technique on the same sooting flame. fv = λk ex ~ ) is obtained from the calibration procedure. A single calibration factor is applied to 6πE (m convert the LII signal to volume fraction. babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y Page 8 of 15 3.3 Signal Interference The light signal detected by the ICCD camera is subject to several possible sources of interference, including 1) LII from char and ash particles; 2) LIF (Laser Induced Fluorescence) from tar and light gas; 3) stray light leaked through the filter and the ICCD photocathode. We carefully investigated these possibilities to show that the LII signal is interference-free. The numerical model results have shown that as long as the soot-to-char yield ratio is higher than 10-3, the signal contribution for char particle is minimal. We also conducted some experiments to confirm this. Signals from a "non-sooting" flame (fuel lean CO/H2/O2/N2 flame) with PC feeding were compared to the signals from a rich, sooting flame. Data from the "non-sooting" flame showed a few orders in magnitude drop of the LII signal intensity within 10 to 20 mm above the burner. We believe the signal at 10 mm is from soot formed due to imperfect mixing between coal volatiles and oxidizer. Test results also showed that the signal intensity at 20 mm above the burner in a fuel-lean "nonsooting" flame was less than one percent of the intensity of a strongly sooting fuel-rich flame at the same location and temperature. These observations show that the signal contributions from the char/ash particles are insignificant and can be neglected. Most species that can interfere potentially with the soot signal detection are short-lived in their excited states (on the order of a few ns [14]). A detection delay of 50 ns was used to avoid interferences from LIF under low to medium laser fluence. Tests with time delays of 50, 200, 500 and 2000 ns between the arrival of the laser beam and camera exposure were conducted. Increasing the detection delay decreased the scattered laser light which shows that the stray light contribution was minimal. 4. Results and discussions Coal pyrolysis tests involved injecting dilute pulverized Pittsburgh #8 coal through the centerline of the flat flame burner operating at different flame temperatures. In all cases, the flame temperature rose rapidly and reached a maximum at ~12.7 mm above the burner. Beyond this point, the temperature decreased gradually due to heat loss to the surroundings through the quartz tower. The calibration factor was determined to be 8.95 × 10-8 ppm/(camera count). Light extinction across the quartz tower enclosure was about 10% loss per window (two surfaces). A correction was applied to the data to account for this. Two-D soot volume fraction distributions were also converted to soot production rates in terms of weight percentage of the parent coal on a daf basis. Particle residence times were calculated following the method of Ma [27]. To calculate the soot production rate from the measured LII data, we used the method illustrated in Figure 6. This is a view from the top of the burner, where the rings refer to different radii. At a given elevation and within the center square region (1.24 × 1.24 mm), the total volumetric rate of soot production is calculated as ∑fvi × Δr × W × v, where fvi represents soot volume fraction at a given radius ri measured via the LII technique, Δr is the radial spatial resolution (29 µm), W refers to the thickness of the beam, and v is the velocity in the vertical direction determined by method discussed in the previous section. Outside of this square region, it is calculated as ∑fvi / W × π 2ri × π × (ri+12 - ri2) × v. The summation of these two is multiplied by a constant soot density, 1.8 g/cc, to give the mass formation rate of soot in g/s. The ratio between the mass formation rate of soot and coal feed rate (0.8 g/s) on a dry ash-free basis gives us the weight percentage of soot formation of daf coal. To improve the signal-to-noise ratio (SNR), the data were averaged by 20 points in the vertical direction which results in a resolution of 2.3 mm (in height) × 29 µm (in radius). babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y Page 9 of 15 ∑fvi × Δr × W × v ∑fvi / W × π 2ri × π × (ri+12 - ri2) × v W = 1.24mm v is velocity Δr = 29 µm Figure 6. Illustration of the method to calculate soot production rate 4.1 Discussion Figure 7 shows a representative contour plot of the radial and axial distributions of soot volume fraction from pyrolyzing Pittsburgh #8 coal particles injected along the centerline of the flat flame burner. In this and other five cases involving different temperatures, soot formation started rapidly but gradually dropped as the height above the burner was increased. Possible reasons for the signal fall-off in the post-flame zone are being investigated. Soot oxidation (by radical species despite the fuel-rich environment), soot dispersion, and soot agglomeration, are among plausible mechanisms. Since the concentrations of oxidizing radicals generally increase with flame temperature, soot destruction should be accelerated in high-temperature flames. Indeed, this trend was observed in our experimental data, where soot volume fraction decreased at elevated flame temperatures. fv, ppm 140 0.0 0.4 0.8 1.2 1.6 2.0 2.4 Height, mm 120 100 80 60 40 20 -2 -1 0 1 2 r, mm Figure 7. 2-D soot volume fraction distribution (1200 °C) Soot production began closer to the coal particle injection point as the temperature was increased. Peak soot production location also moved closer to the burner as the temperature was raised. This is better shown in Figure 8, where the residence time corresponding to peak soot production is seen to decrease with increasing pyrolysis temperature. Figure 8 provides quantitative data on the rate of soot formation and can be used to calibrate the numerical model. babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y Residence time of max soot formation, ms Page 10 of 15 80 70 60 50 40 30 20 10 0 800 1000 1200 1400 1600 Temperature, C Figure 8. Residence time corresponding to peak soot production as a function of the coal pyrolysis temperature The maximum soot formation is also seen to increase with increased pyrolysis temperatures as shown in Figure 9. We also varied the coal feeding rate at 1200°C to investigate the effect of particle loading on the rate of soot production. A near-linear relationship between the maximum rate of soot production and the feeding rate is observed (Figure 10). This indicates that within our test range, particle-to-particle interactions have a minimal effect on soot production, which is desirable for kinetic studies. Max soot formation 3.5 Volume fraction, ppm 3 2.5 2 1.5 1 0.5 0 800 1000 1200 1400 1600 Temperature, C Figure 9. Maximum soot volume fraction as a function of the coal pyrolysis temperature babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y Page 11 of 15 4 Volume fraction, ppm 3.5 3 2.5 2 1.5 1 0.5 0 0 0.5 1 1.5 2 Coal feeding rate, g/hr Figure 10. Maximum soot volume fraction as a function of the coal feed rate 4.2 Error analysis Quantifying the error associated with the LII measurement is severely complicated by the intrinsic heterogeneity of coal. Here, the error analysis is conducted on a best effort basis. Furthermore, the analysis is carried out only on the volume fraction data which are direct results from Equation 11. The error associated with the LII signal involves several factors outlined below: 1. The heterogeneity of coal and unsteady feeding - These two are coupled together and there is no easy way to distinguish them. The best method to characterize it is to repeat the tests under the same conditions and calculate the standard deviation of the signal. The uncertainty estimated via this method is approximately ± 10%. 2. The reflectance of the window - The quartz windows become increasingly coated with tar and soot from coal pyrolysis as the tests proceed. Although the windows are cleaned from time to time, the change in their reflectance is unavoidable. The uncertainty is estimated to be ± 2% based on the variation of the light extinction through different parts of the windows. 3. The camera's CCD dark current and intensifier noise - The counts readout from the camera is in the order of tens of millions for an image of 3000-pulse integration. This is much higher than the error induced by dark current over the same time period of the camera exposure (~0.5 counts/pixel/sec). The noise from the intensifier characterized by equivalent background illumination (EBI, Max. 0.01 counts/pixel/sec) is also negligible. The error associated with the calibration factor also has several sources, including flame unsteadiness, the Abel-inversion process, the laser energy detector, etc. The best method for its characterization, however, is to estimate the maximum difference between volume fraction determined by LII and that of the light extinction technique. Only data that are larger than 1.0 ppm are used for this evaluation. The maximum uncertainty is estimated to be ± 5%. The overall uncertainty of any measurement result can be calculated by [28] 1/ 2 2 2 2 ⎡⎛ 1 ∂R ⎛ 1 ∂R ⎞ ⎤ ⎞ ⎛ 1 ∂R ⎞ wR = ⎢⎜⎜ w1 x1 ⎟⎟ + ⎜⎜ w2 x2 ⎟⎟ + ⋅ ⋅ ⋅ + ⎜⎜ wn xn ⎟⎟ ⎥ ⎢⎝ R ∂x1 ⎠ ⎝ R ∂x2 ⎠ ⎝ R ∂xn ⎠ ⎥⎦ ⎣ babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y (12) Page 12 of 15 where R is the quantity of interest, w is the relative uncertainty in percentage, and xi are the independent variables. Since the volume fraction is directly proportional to the variables mentioned above, the measurement uncertainty simplifies to w S LII ⎛ =⎜ ⎜ ⎝ n ∑ i =1 wi2 ⎞ ⎟ ⎟ ⎠ 1/ 2 with a value of ± 11.4%. 5. Conclusions In this work, we applied the laser-induced incandescence (LII) technique to characterize soot formation around pyrolyzing PC particles injected through a gaseous flame sheet using a flat flame burner. To our knowledge, it is the first time that the LII technique has been applied to a particle-laden flame. Numerical simulation and experimental results proved the feasibility of the technique and yielded spatially-resolved quantitative soot volume fraction data for pyrolyzing Pittsburgh #8 coal particles at 1000 to 1500°C. Soot formation rates can be derived from these data for inclusion in a computer model. The 2-D data can also be used to validate the overall prediction of the soot model. Soot volume fractions were not affected by particle-to-particle interactions but varied linearly with changes in coal feed rate. Further development of the technique for soot particle size distribution measurements in coal particle-laden flames is underway. Acknowledgement The authors would like to thank Alan Sayre and Rick Wessel of the B&W Combustion Analysis group for the help on the development of the model. Nomenclature a, d Radius and diameter of the particle c Speed of light j = 1, 2, 3; specific heat of daf coal/char, ash and moisture Cj Cs Specific heat of carbon F Laser fluence h Planck constant Heat of vaporization ΔH v m 8 = 2.998 × 10 m / s J/g-K = 1.90 J / g K J cm 2 = 6.626 × 10 −34 J s = 7.78 × 10 5 for carbon in Melton model, = 5.45 × 10 5 for SiO2, J mol = 1.381 × 10 −23 J / K m& Boltzmann constant Thermal conductivity Complex index of refraction Feeding rate of coal n pv q Particle number density Equilibrium vapor pressure of sublimed carbon Rate of energy transfer q(t ) Time dependent laser power density radial coordinate Universal gas constant Soot to char LII signal ratio Time Time delay at the start of the exposure k ka ~ m r R R LII t t0 W /mK 1.90 − 0.55 i for soot, = 2.050 − 1.127 i for char kg / s # / m3 atm W W m2 m = 8.314 J mol K babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y sec = 50 ns Page 13 of 15 t exp Exposure time of the ICCD camera T Particle temperature Initial particle temperature Internal energy of the particle T0 U int ernal V& = 100 ns K K J 3 Volumetric flow rate j = 1, 2, 3; mass fraction of daf coal/char, ash and moisture wj m /s x yi αT = 36 g / mol in Melton model Molecular weight of carbon vapor = 28.47 g / mol for air Molecular weight of ambient gases = πd λ Size parameter i = 1,…,5; daf-based mass fractions of carbon, hydrogen, oxygen, nitrogen, sulfur Thermal accommodation coefficient γ Specific heat ratio γi i = 1,…,5; coefficient of C, H, O, N and S for char density calculation γ* Effective specific heat ratio ρs Density of soot (graphite) ρv λ λS ε λS Density of carbon vapor Laser wavelength Signal wavelength Wavelength-dependent emissivity Ф μ Equivalence ratio of the FFB flame Mean atomic weight Wv Wa = c p cv = 1.4 for air at room temperature = 2.26 g / cm 3 = ( pWv ) (RT ) , kg / m 3 = 532 nm nm kg/mol References: [1] [2] [3] [4] [5] [6] [7] [8] [9] Melton, L.A., Soot Diagnostics Based on Laser Heating. Applied Optics, 1984. 23(13): p. 22012208. Boiarciuc, A., F. Foucher, and C. Mounaim-Rousselle, Soot volume fractions and primary particle size estimate by means of the simultaneous two-color-time-resolved and 2D laserinduced incandescence. APPLIED PHYSICS B-LASERS AND OPTICS, 2006. 83(3): p. 413421. Mewes, B. and J.M. Seitzman, Soot Volume Fraction and Particle Size Measurements with Laser-Induced Incandescence. Applied Optics, 1997. 36(3): p. 709-717. Ni, T., J.A. Pinson, S. Gupta, and R. Santoro, Two-Dimensional Imaging of Soot Volume Fraction by the Use of Laser-Induced Incandescence. Applied Optics, 1995. 34(30): p. 70837091. Shaddix, C.R. and K.C. Smyth, Laser-Induced Incandescence Measurements of Soot Production in Steady and Flickering Methane, Propane, and Ethylene Diffusion Flames. Combustion and Flame, 1996. 107: p. 418-452. Vander Wal, R.L. and T.M. Ticich, Cavity ringdown and laser-induced incandescence measurements of soot Applied Optics, 1999. 38(9): p. 1444-1451. Vander Wal, R.L., Z. Zhou, and M.Y. Choi, Laser-induced incandescence calibration via gravimetric sampling. Combustion and Flame, 1996. 105(4): p. 462-470 Geitlinger, H., T.H. Streibel, R. Suntz, and H. Bockhorn. Two-Dimensional Imaging of Soot Volume Fractions, Particle Number Densities, and Particle Radii in Laminar and Turbulent Diffusion Flames. in Twenty-Seventh Symposium (International) on Combustion. 1998: The Combustion Institute. Quay, B., T.W. Lee, T. Ni, and R.J. Santoro, Spatially resolved measurements of soot volume fraction using laser-induced incandescence Combustion and Flame, 1994. 97(3-4): p. 384392. babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y Page 14 of 15 [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] Will, S., S. Schraml, and A. Leipertz. Comprehensive Two-Dimensional Soot Diagnostics Based on Laser-Induced Incandescence (LII). in Twenty-Sixth Symposium (International) on Combustion. 1996: The Combustion Institute. Dec, J.E., A.O. Zur Loye, and D.L. Siebers, Soot distribution in a D. I. diesel engine using 2D laser-induced incandescence imaging, in SAE paper 910224 1991. Kock, B.F., B. Tribalet, C. Schulz, and P. Roth, Two-color time-resolved LII applied to soot particle sizing in the cylinder of a Diesel engine. COMBUSTION AND FLAME 2006. 147(12): p. 79-92. Bougie, B., L.C. Ganippa, N.J. Dam, and J.J. Ter Meulen, On particulate characterization in a heavy-duty diesel engine by time-resolved laser-induced incandescence. APPLIED PHYSICS B-LASERS AND OPTICS, 2006. 83(3): p. 477-485. Eckbreth, A.C., Laser Diagnostics for Combustion Temperature and Speices. Combustion Science & Technology Book Series. Vol. 3. 1996: Gordon and Breach Publishers. Michelsen, H.A., F. Liu, B.F. Kock, H. Bladh, A. Boiarciuc, M. Charwath, T. Dreier, R. Hadef, M. Hofmann, J. Reimann, S. Will, P.-E. Bengtsson, H. Bockhorn, F. Foucher, K.-P. Geigle, C. Mounaïm-Rousselle, C. Schulz, R. Stirn, B. Tribalet, and R. Suntz, Modeling laser-induced incandescence of soot: a summary and comparison of LII models Applied Physics B: Lasers and Optics, 2007. 87(3): p. 503-521. Chase, J., M. W., NIST-JANAF Thermochemical Tables, J. Phys. Chem. Ref. Data. Fourth ed. Vol. 14. 1998. Michelsen, H.A., Understanding and predicting the temporal response of laser-induced incandescence from carbonaceous particles JOURNAL OF CHEMICAL PHYSICS, 2003. 118: p. 7012-7045. Lee, S.C. and C.L. Tien. Optical Constants of Soot in Hydrocarbon Flames. in Eighteenth Symposium (International) on Combustion. 1981. Liu, F., K.J. Daun, D.R. Snelling, and G.J. Smallwood, Heat Conduction from a Spherical Nano-Particle: Status of Modeling Heat Conduction in Laser-Induced Incandescence. Applied Physics B: Lasers and Optics, 2006. 83: p. 355-382. Leider, H.R., O.H. Krikorian, and D.A. Young, Thermodynamic properties of carbon up to the critical point. Carbon 1973. 11: p. 555-563 Merrick, D., Mathematical Models of the Thermal Decomposition of Coal: 2. Specific Heats and Heats of Reaction. Fuel, 1982. 62: p. 540-546. Merrick, D., Mathematical Models of the Thermal Decomposition of Coal: 3. Density, Prosity and Contraction Behaviour. Fuel, 1983. 62: p. 547-552. Atkinson, B. and D. Merrick, Mathematical Models of the Thermal Decomposition of Coal: 4. Heat Transfer and Temperature Profiles in a Coke-oven Charge. Fuel, 1983. 62: p. 553-561. Wessel, R.A., Radiative Properties of Particle-Gas Mixtures - A Mie theory based program. 1985, Babcock & Wilcox. Gansman, Particle Radiation Properties Subroutine. 1992, Babcock & Wilcox. Glick, D.C. and A. Davis, Operation and Composition of the Penn State Coal Sample Bank and Data Base. Organic Geochemistry, 1991. 17(4): p. 421-430. Ma, J., Soot Formation and Secondary Reactions During Coal Pyrolysis, in Department of Chemical Engineering. 1996, Brigham Young University. Kline, S.J. and F.A. McClintock, Describing Uncertainties in Single-Sample Experiments. Mechnical Engineering, 1953. 75: p. 3-7. Copyright© 2009 by Babcock & Wilcox Power Generation Group, Inc. a Babcock & Wilcox company All rights reserved. No part of this work may be published, translated or reproduced in any form or by any means, or incorporated into any information retrieval system, without the written permission of the copyright holder. Permission requests should be addressed to: Marketing Communications, The Babcock & Wilcox Company, P.O. Box 351, Barberton, Ohio, U.S.A. 44203-0351. Disclaimer babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y Page 15 of 15 Although the information presented in this work is believed to be reliable, this work is published with the understanding that The Babcock & Wilcox Company and the authors are supplying general information and are not attempting to render or provide engineering or professional services. Neither The Babcock & Wilcox Company nor any of its employees make any warranty, guarantee, or representation, whether expressed or implied, with respect to the accuracy, completeness or usefulness of any information, product, process or apparatus discussed in this work; and neither The Babcock & Wilcox Company nor any of its employees shall be liable for any losses or damages with respect to or resulting from the use of, or the inability to use, any information, product, process or apparatus discussed in this work. babcock & wilcox power gener ation gr oup, inc., a B a b c o c k & W i l c ox c o m pa n y