Atomization and Co-Combustion of Crude Glycerin with Natural Gas and Hydrogen

Atomization and Co-Combustion of Crude Glycerin with Natural Gas and Hydrogen Pedro Queirós Abstract The present study focuses on the atomization and co-combustion of glycerin. Initially, glycerin sprays produced by two air assisted atomizers have been characterized under non-reacting conditions. For both atomizers the results revealed a decrease in the Sauter mean diameter with increasing values of the atomizing air to fuel mass ratio up to 1, beyond which no significant improvements were obtained, and the presence of larger droplets in the periphery of the sprays and smaller droplets in its central region. Subsequently, a selection of the characterized glycerin sprays has been burn in cocombustion with natural gas and hydrogen in a laboratory furnace fired by a swirl burner. Flue-gas data were obtained for various flames, which quantify the effects of the type of atomizer, the atomization quality, the excess air and the incorporation degree of glycerin in the fuel mixture on pollutant emissions. In addition, the deposits formed in the near burner region and in the furnace walls were collected and analyzed. The results showed: 1) improvements in combustion efficiency with increasing atomization quality; 2) reduction in CO and hydrocarbons (HC) emissions with increasing excess air; 3) decrease in CO and HC emissions as the thermal percentage of glycerin in the fuel mixture increases up to a value close to 45%, for higher values CO, HC and NO x emissions increase; 4) no acrolein emissions detected for all flames studied; and 5) deposits collected at the burner exit and at the furnace walls in the near burner region contain high concentrations of Na, K and Cl, which could endanger long term furnace utilization. Keywords: Crude glycerin, laboratory furnace, atomization, combustion, pollutant emissions, ash deposits Introduction With the growing concern over environmental sustainability and the need to develop renewable sources of energy, biodiesel production is due to increase substantially over the next years. In 11 the world biodiesel production is expected to be around 22 1 9 l, value that should exceed 4 1 9 l by 19 [1]. Biodiesel production is mostly done by the reutilization of animal and vegetable fats in a process called transesterification. During this process, the waste animal and the vegetable fats react with an alcohol to produce biodiesel and glycerin as a by-product. On a molar basis, one mole of glycerin is produced for every three mole of biodiesel, and volumetrically 1% of the initial reactants are transformed to glycerin. In the past decade refined glycerin was commercialized at 1-14 /ton but with the increasing biodiesel production and the resulting excess of the glycerin by-product the prices went down significantly. Nowadays refined glycerin has a price of 4-5 /ton and crude glycerin resulting from biodiesel production reaches 4-1 /ton [2]. The prices are so low that many producers are having difficulty to support the cost of storage and transportation of this product. In order to solve the problem of excess of glycerin several solutions are being investigated including combustion, conversion in biogas, chemical conversion, etc Combustion of glycerin would be an attractive solution but it presents many challenges. Glycerin has a low calorific value ( 13 MJ/Kg), high ignition temperature (37 ºC) and high viscosity (1 cp at 22 ºC). In addition, combustion of glycerin presents the risk of acrolein emissions. This unstable and toxic carbonyl compound is the result of the thermal decomposition of glycerin and can be formed at temperatures around 28 ºC with very high toxicity even at low concentrations (2 ppm). For all these reasons it is essential to understand better the processes occurring in the combustion of glycerin, including the way the atomization quality affects the overall process. Two years ago, at the 33 rd Combustion Symposium, Bohon et al. [3] presented an interesting study on the combustion characteristics of glycerin in a 7 kw prototype high-swirl burner and in a 82 kw laboratory-scale moderate-swirl furnace. Their results showed NO x emissions between 15 ppm and 24 ppm and low total hydrocarbon (HC) emissions along with ash formation with high concentrations of sodium, potassium, phosphorous and chlorine. Pratzer et al. [4] investigated the cocombustion of glycerin with yellow grease in a large-scale boiler. They results showed a significant increase in HC and SO 2 emissions with the incorporation of glycerin in the fuel mixture and deposit formation in the near burner region with high content of sodium, potassium and chlorine, representing possible problems for long term utilization. In regard to the acrolein emissions, Pratzer et al. [4] reported no measurable emissions, while Bohon et al. [3] reported emissions around 18 ppb well below the health hazards limits. The use of supporting fuels to enhance the combustion efficiency of the glycerin could be a good solution to deal with the problems related with pure glycerin combustion, namely deposit 1

formation and toxic pollutant emissions, as reported by Bohon et al. [3]. In this context, the present study focuses on the atomization and co-combustion of crude glycerin with natural gas and hydrogen in a laboratory furnace. Initially, glycerin sprays produced by two air assisted atomizers have been characterized under non-reacting conditions. Subsequently, a selection of the characterized glycerin sprays has been burn in co-combustion with natural gas and hydrogen in the laboratory furnace fired by a swirl burner. The data quantify the effects of atomizer type, atomization quality, excess air and incorporation degree of glycerin in the fuel mixture on combustion performance. In addition, the deposits formed in the near burner region and in the furnace walls have been collected and analyzed. Experimental Facilities and Measuring Techniques Figure 1 shows a schematic of the spray test rig used for the spray characterization experiments under non-reacting conditions. The spray chamber has a quadrangular shape with 1 m side and is equipped with an exhaust system to extract vapors that result from droplet vaporization. The Malvern Particle Size Analyzer, used for droplet size distribution measurements, is mounted in the top of chamber. The burner gun is located axially above the chamber with fixtures permitting axial and radial movement. Figure 2 shows a schematic of the laboratory furnace and auxiliary equipment. The furnace is a cylinder.6 m in inside diameter and 2.4 m in length. The cylindrical combustion chamber has a vertical axis to minimize asymmetry due to natural convection and biased ash particle deposition and it is down-fired to facilitate particulate removal. The furnace roof and the initial 1.2 m length of the cylindrical walls are refractory lined. The outer surfaces of the refractory walls are surrounded by cooling water jackets. The remaining 1.2 m length of the wall surfaces are water cooled only. Figure 3 shows a schematic of the furnace roof and burner arrangement. The burner consists of a central gun and a secondary air supply in a conventional double-concentric configuration, terminating in a refractory quarl. The secondary air stream is fitted with guide vanes of constant cord and angle of 6º for inducing swirl. The burner gun comprises a removable air assisted atomizer and a co-axial supply of natural gas and/or hydrogen. In this study two air-assisted atomizers (atomizers 1 and 2) were used, as shown in Figure 4. The glycerin is supplied to the burner using a mono-progressing positive displacement pump coupled to a variable-speed motor which is controlled electronically. A loss-in-weight technique incorporating a weighbridge and a timing device ensures the good maintenance of the desired fuel flow rate controlled by the pump speed. The observed variation in flow is within ±3% over a sampling period of 1 s, and negligible over longer periods. Natural gas is fed to the burner by a central line operating with a pressure of 22 mbar and a maximum flow rate of 4 m 3 /h. During the present study, hydrogen from pressurized bottles was also used. The gas flows were controlled with pressure regulators and valves and the flow rates measured using a digital rotameters. Flue-gas composition data were obtained using a stainless steel water-cooled probe. The probe was composed of a central 1.3 mm inner diameter tube through which quenched samples were evacuated. This central tube was surrounded by two concentric tubes for probe cooling. The gas sample was drawn through the probe and part of the system by an oil-free diaphragm pump. A condenser removed the main particulate burden and condensate. A filter and a drier removed any residual particles and moisture so that a constant supply of clean dry combustion gases was delivered to the analyzers through a manifold to give species concentration on a dry basis. The analytical instrumentation included a magnetic pressure analyzer for O 2 measurements, a non dispersive infrared gas analyzer for CO 2 and CO measurements, a flame ionization detector for hydrocarbons (HC) measurements and a chemiluminescent analyzer for NO x measurements. At the combustor exit, probe effects were negligible and errors arose mainly from quenching of chemical reactions, which was found to be adequate. Repeatability of the flue-gas data was, on average, within 5%. For acrolein emission measurements, 1 ml samples were taken from the exhaust with a proper needle and analyzed with gas chromatography. The system was calibrated with a pure acrolein sample and exhaust gas samples where then compared. The collected deposits were analyzed in a scanning electron microscope (SEM) JEOL, model JSM-71F. The microscope allows obtaining 3-D images from a selected area through the sample irradiation by an electron beam. The microscope is also equipped with an energy dispersive x-ray detector (EDS) that allows quantifying the ultimate composition of a deposit sample within areas of 5 5 μm Results and Discussion Tables 1 to 3 show the glycerin physical and chemical characteristics, the test conditions used in the spray characterization experiments and the furnace operating conditions, respectively. The spray 2

characterization experiments (Table 2) allowed quantifying the effects of the atomizer configuration, atomizing air to glycerin mass ratio (AFR) and glycerin mass flow rate in spray quality. The trials performed in the laboratory furnace (Table 3) allowed quantifying the effects of the AFR, excess air level and degree of glycerin incorporation in the fuel mixture on pollutant emissions. It will be noted that the reported pollutant emissions have been corrected for the dilution effect of the excess air. Figure 5 shows the influence of AFR on SMD for three glycerin mass flow rates for atomizers 1 and 2. The measurements were made along a diameter of the cross-section of the spray (r = ) at a distance of 5 mm downstream from the atomizer exits, where the break-up process is already completed. For atomizer 1, Figure 5a reveals a rapid decrease in SMD with increasing AFR up to values around 1, regardless of the glycerin flow rate, beyond which further increases yield only marginal improvements in spray quality. At low values of AFR, the kinetic energy of the atomizing air is insufficient to overcome the viscous and surface tension forces which oppose the disintegration of the liquid. As the AFR is increased the air mass flow and its kinetic energy increase and consequently more energy is available to shatter the glycerin jet into droplets. The leveling off of the curves beyond a certain AFR suggest that above this value the additional atomizing air does not effectively interact with the glycerin jet. Figure 5a also shows that, for a given AFR below 1, the SMD decreases as the glycerin flow rate increases. The decrease of SMD with the glycerin flow rate at a constant AFR is due to the increase in air mass flow and its kinetic energy which overcame the deleterious effect that an increase in glycerin mass flow rate alone would have on atomization. For atomizer 2, Figure 5b reveals that the effect of both the AFR and the glycerin mass flow rate on spray quality is much less pronounced, which is attributed to the atomizing air velocities in atomizer 2 being much higher than in atomizer 1 even at low AFR values. Figure 6 shows the influence of the AFR on CO, HC and NO x emissions for atomizers 1 and 2 maintaining constant the excess air level (O 2 in the flue gas = 5%). For both atomizers, it is seen that the CO and HC emissions increase substantially for AFR values below.8 as a result of the poorer atomization. Note that atomizer 2 yields systematically lower CO and HC emissions than atomizer 1, which is consistent with the SMD data in Figure 5, but the co-combustion of glycerin with the gaseous fuels yields always higher CO and HC emissions than the pure gaseous fuel firing, regardless of the atomizer used. Moreover, the top two panels of Figure 6 show an increase in the CO and HC emissions for AFR values above 1.5. Since the SMD does not change for values of AFR > 1.5 in any of the atomizers, the increase in CO and HC emissions suggests that some initial similar size, but higher momentum droplets were able to penetrate through the internal recirculation zone, leading to an overall reduction in the combustion efficiency. The last panel of Figure 6 shows that the NO x emissions from the co-combustion of glycerin with the gaseous fuels tend to increase with the value of AFR, being, however, lower than those from pure gaseous fuel firing. As the AFR increases, both the entrainment of secondary air by the central jet and the combustion intensity increases resulting, most likely, in higher flame temperatures. As a consequence, the NO formation through the thermal mechanism is promoted leading to higher NO x emissions. Figure 7 shows the influence of the excess air on CO, HC and NO x emissions for atomizers 1 and 2, both operating with an AFR =.5. Atomizer 1 yields high CO and HC emissions at low excess air levels possibly due to the formation of rich areas in the central region of the spray with low O 2 entrainment. As the excess air increases, more O 2 is available and the CO and HC emissions decrease. For values of O 2 in the flue gas higher than 5% the CO and HC emissions are insignificant. Atomizer 2 presents a different behaviour for low values of excess air, with the CO and HC emissions being much lower as compared to atomizer 1. As the excess air increases, the CO and HC emissions from both atomizers are similar. Atomizer 1 produces a solid cone spray that is very sensitive to excess air variations since the O 2 penetration in the central regions of the sprays is highly dependent on the excess air level. Atomizer 2 originates a flat spray that has less inner volume and higher exposure area, which enhances the mixing between air and fuel even for lower excess air levels. This avoids the formation of rich areas and reduces the CO and HC emissions. The last panel of Figure 7 shows that the NO x emissions tend to increase with increases in the excess air level for atomizer 1. Increasing the excess air level enhances mixing and droplet evaporation, which causes an increase in flame temperature promoting NO formation through the thermal mechanism. Atomizer 2 exhibits lower NO x emissions probably caused by lower flames temperatures in the near burner region, which suppresses NO formation via the thermal mechanism. Figure 8 shows the influence of the thermal percentage of glycerin in the fuel mixture on CO, HC and NO x emissions for atomizer 2 operating with an AFR =.5. The results show that NO x emissions do not vary and that CO and HC emissions decrease as the thermal percentage of glycerin in the fuel mixture increases up to a value close to 45%. For higher values CO, HC and NO x emissions increase. 3

In regard to pollutant emissions, it is important to point out that no acrolein emissions were detected for all flames studied. Sufficient residence times in regions of relatively high temperature ensured the oxidation of the acrolein. During the co-combustion of glycerin with the gaseous fuels deposits were formed at the burner exit and at the furnace walls in the near burner region. The deposits were collected and later analyzed in a SEM/EDS device. Figure 9 shows SEM images of typical deposits collected at the burner exit and Figure 1 shows the chemical analysis of a representative deposit sample, obtained from four areas of 5 5 μm of the sample. The deposits present a dense aspect, grainy consistency with some moisture and easy removal. The chemical composition shows the presence of C, Cl, Na, O, Fe and some small amounts of K, Si, Ca and Al, which is consistent with glycerin ash analysis included in Table 3. The presence of O and Fe suggest the existence of corrosion at the burner, enhanced by Cl, which is a highly corrosive element that could be prejudicial for long term furnace operation. The presence of Na in the deposits could promote K capture and form silicates with low melting point that improve deposit adherence. Figure 11 shows SEM images of typical deposits collected at the furnace walls in the near burner region and Figure 12 shows the chemical analysis of a representative deposit sample, obtained from four areas of 5 5 μm of the sample. The deposits formed at the furnace walls can be described as a white powder with no moisture and difficult to remove. The chemical composition shows shows higher concentrations of Cl and Na, which, combined with K, is very problematic for corrosion and can explain the higher adherence of these deposits. As a final remark, the present study shows the possibility of burning glycerin but the results seems to demonstrate that co-combustion of glycerin is a rather interesting solution compared with pure glycerin firing. Co-combustion could reduce the problems related with deposit formation and pollutant emissions and could easily be legally approved by the authorities for industrial purposes. In this context, further studies on the combustion and co-combustion of glycerin should be made namely the possibility of mixing glycerin with other liquids fuels in order to create emulsions that could be properly burned. Conclusions The main conclusions from this investigation are as follows. For both atomizers the spray fineness significantly increases with increasing AFR up to a value, regardless of the glycerin flow rate, beyond which any further increases in AFR causes only minor improvements in spray quality. Increasing AFR causes a general improvement in combustion process. For both atomizers, increasing AFR diminishes droplet diameter and increases entrainment of the secondary air leading to a reduction in CO and HC emissions. NO x emissions from the cocombustion of glycerin tend to increase with the value of AFR, as a result of the higher flame temperatures that promote thermal NO formation. Atomizer 1 yields high CO and HC emissions at low excess air while atomizer 2 presents low CO and HC emissions. As the excess air increases, the CO and HC emissions from both atomizers are similar and low. NO x emissions tend to increase with increases in the excess air level for atomizer 1; atomizer 2 exhibits lower NO x emissions probably caused by lower flames temperatures in the near burner region, which suppresses NO formation via the thermal mechanism. NO x emissions do not vary and CO and HC emissions decrease as the thermal percentage of glycerin in the fuel mixture increases up to a value close to 45%. For higher values CO, HC and NO x emissions increase. No acrolein emissions were detected for all flames studied. Deposits collected at the burner exit and at the furnace walls in the near burner region contain high concentrations of Na, K and Cl, which could endanger long term furnace utilization. References [1] OECD Conference Centre, Paris, 25-25 June (9). [2] Werpy, T., Petersen, G., Top value added chemicals from biomass, results of screening for potential candidates from sugars and synthesis gas. Available in http://www.osti.gov/, acessed in March 11. [3] Bohon, M. D., Metzger, B. A., Linak, W. P., King, C. J., Roberts, W. L. Glycerol combustion and emissions, Proceedings of the Combustion Institute, 33, 2717-2724 (11). [4] Pratzer, R., Norris, M., Doering, A., Jorgenson, R., Neece, C., Zimmerli, B., Stack emissions 4

evaluation: combustion of crude glycerin and yellow grease in an industrial fire tube boiler Agricultural Utilization Research Institute (7). Table 1. Glycerin physical and chemical characteristics. Property Value Density at 15 ºC (kg/m 3 ) 1268 Dynamic viscosity (Ns/m 2 ) at 23.3 ºC 271 at 5 ºC 5.7 at 8 ºC 22.2 Composition (wt %) Glicerol MONG Moisture Ash Methanol Elemental analysis (wt %, dry ash free) Carbon Hydrogen Nitrogen Sulfur Oxygen Ash composition (wt %) Chlorine Sodium Potassium Phosphorous Calcium Silicon Others Heating value (MJ/kg) Inferior Superior 84.5 2.13 1.31 3.5.1 32.5 8.8 58.7 63. 34.5 1..3.1.1 1. 12.9 14.8 Table 2. Test conditions used in the spray characterization experiments. Atomizer (kg/h) (kg/h) AFR (ºC) 1 6-12 3-24.5-2 6-93 2 6-12 3-12.5-1 8 Table 3. Furnace operating conditions. Operating parameter Atomizer 1 Atomizer 2 Trial series number I II III IV V Flue gas O 2 (%) 5 3.2-5.9 5. 2.1-6. 5. Hydrogen (kw) Natural gas (kw) 45-58 45-58 45-58 45-58 28-58 Glycerin thermal input (kw) mass flow rate (kg/h) temperature (ºC) Atomizing air mass flow rate (kg/h) temperature (ºC) 22-35 6-9.6 7 3-19.2 22-35 6-9.6 7 9-14.4 22-35 6-9.6 7 3-9.6 22-35 6-9.6 7 3-4.8 22-52 6-14.4 7 Spray AFR.5-2 1.5.5-1.5.5 3-7.2 5

Figure 1. Schematic of the spray test rig. Figure 2. Schematic of the laboratory furnace and auxiliary equipment. Figure 3. Schematic of the furnace roof and burner arrangement. 6

SMD (µm) SMD (µm) 7 6 5 4 3 1 7 6 5 4 3 1 a) b) Figure 4. Schematic of the air-assisted atomizers. Flow rate (kg/h) 6 9.6 12 Flow rate (kg/h) 6 9.6 12.2.4.6.8 1 1.2 1.4 1.6 1.8 2 AFR Figure 5. Influence of AFR on SMD for three glycerin mass flow rates. a) Atomizer 1; b) atomizer 2. 7

Figure 6. Influence of the AFR on CO, HC and NO x emissions for atomizers 1 and 2. The dotted lines represent the emissions values for the pure NG flame and for the NG + hydrogen flame. For all tests: O 2 in the flue gas = 5%. 8

NO x (volume ppm @ 6% O 2 ) HC (volume ppm @ 6% O 2 ) CO (volume ppm @ 6% O 2 ) 1 1 8 6 4 Symbol Glycerin NG H 2 Atm 1 8 22 58 1 35 45 1 22 58 2 22 58 2 8 6 4 4 3 1 2 3 4 5 6 O 2 in the flue gas (% vol.) Figure 7. Influence of the excess air on CO, HC and NO x emissions for atomizers 1 and 2, both operating with an AFR =.5. 9

Concentration (%) CO (volume ppm @ 6% O2) HC, NOx (volume ppm @ 6% O2) 1 4 8 6 4 CO NO x HC 3 1 1 3 4 5 6 7 Glycerin in the fuel mixture (% energy) Figure 8. Influence of the thermal percentage of glycerin in the fuel mixture on CO, HC and NO x emissions for atomizer 2 operating with an AFR =.5. For all tests: O 2 in the flue gas = 5%. Figure 9. SEM images of typical deposits collected at the burner exit. 7 6 Area 1 Area 2 Area 3 Area 4 5 4 3 1 C Cl Na O Fe K Si Ca Al Element Figure 1. Chemical analysis of representative deposits collected at the burner exit. 1

Concentration (%) Figure 11. SEM images of typical deposits collected at the furnace walls in the near burner region. 7 6 Area 1 Area 2 Area 3 Area 4 5 4 3 1 C Cl Na O Fe K Si Ca Al Element Figure 12. Chemical analysis of representative deposits collected at the furnace walls in the near burner region. 11