- Email: [email protected]
Study of water–oil emulsion combustion in large pilot power plants for fine particle matter emission reduction
Experimental Thermal and Fluid Science 31 (2007) 421–426 www.elsevier.com/locate/etfs
Study of water–oil emulsion combustion in large pilot power plants for fine particle matter emission reduction C. Allouis b
a,*
, A. L’Insalata b, L. Fortunato b, A. Saponaro b, F. Beretta
a
a Istituto di Ricerche sulla Combustione, C.N.R., P.le V. Tecchio 80, 80125 Napoli, Italy ANSALDO Caldaie, Centro Combustione e Ambiente, Via Milano km 1,6, 70023 Gioia del Colle, Italy
Received 10 October 2005; received in revised form 5 March 2006; accepted 30 April 2006
Abstract The combustion of heavy fuel oil for power generation is a great source of carbonaceous and inorganic particle emissions, even though the combustion technologies and their efficiency are improving. The information about the size distribution function of the particles originated by trace metals present into the fuels is not adequate. In this paper, we focused our attention the influence of emulsion oil-water on the larger distribution mode of both the carbonaceous and metallic particles. Isokinetic sampling was performed at the exhausts of flames of a low-sulphur content heavy oil and its emulsion with water produced in two large pilot plants. The samples were size-segregated by mean of an 8-stages Andersen impactor. Further investigation performed on the samples using electronic microscopy (SEM) coupled with X-ray analysis (EDX) evidenced the presence of solid spherical particles, plerosphere, with typical dimensions ranging between 200 nm and 2–3 lm, whose atomic composition contains a large amount of the trace metals present in the parent oils (Fe, V, Ni, etc.). EDX analyses revealed that the metal concentration increases as the plerosphere dimension decreases. We also observed that the use of emulsion slightly reduce the emission of fine particles (D50 < 8 lm) in the large scale plant. 2006 Elsevier Inc. All rights reserved. Keywords: Metal concentration; Heavy oil combustion; Particulate size distribution
1. Introduction It is now assessed, that small particles present in the atmosphere produce significant effects on human health even at concentration levels considered acceptable few years ago [1]. Furthermore toxicology points out the importance of the size and/or the chemical composition in the production of the damage. Particles produced by combustion processes exhibit typical sizes in the submicronic and even in the nanometric range and are consequently considered among the primary responsible of the harmful effects. Furthermore, combus-
*
Corresponding author. Tel.: +39 081 7682247; fax: +39 081 5936936. E-mail address: [email protected] (C. Allouis).
0894-1777/$ - see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.expthermflusci.2006.04.011
tion particles are composed by unburned carbon, sulphates and metals contained in the fuel itself. Heavy oils are crude liquid fractions containing high molecular mass organic structures within which inorganic functionalities are inserted. It implies that when they are introduced in the combustion chamber in form of sprays their oxidation and pyrolysis take place both in the liquid phase, inside the droplets, and in the gaseous phase around them. Consequently, the size, shape and composition of the particles are all determined by the relative roles of the different reactive channels, as it has been discussed, few years ago, by Linak and Wendt [2]. On the other extreme the pyrolysis in the liquid phase produces particles, carbonaceous and/or metallic, whose typical sizes are comparable with those of the parent droplets [3]. Many of such structures present themselves as hollow riddled spheres, called cenospheres, which are subject to fragmentation due to
422
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426
thermal stresses and attrition. However some authors have also reported the presence of denser compact particles, called plerospheres, with typical sizes of the order of few microns in the fly ashes produced during coal combustion [4,5] and heavy oil combustion [6]. The presence of these different particles have also a significant impact on the removal efficiency of the particles from the exhaust streams: larger particles, in the micron size range, are captured much more easily than those into the submicronic one. On the other side, nanometric particles escape almost without obstacles in the atmosphere. Thus, it becomes important to characterize the fly ash from the combustion of heavy oil and its emulsion, with particular emphases on particle morphology and the size distribution of potentially toxic elements. Therefore, it is a specific goal of this communication to analyze the sizes, shapes and metal contents of the larger and intermediate distribution modes of particles present at the exhaust of two pilot plants (6 MWt and 50 MWt) fired with heavy oil and its emulsion with water. 2. Experimental set-up Experiments were carried out in a 6 MWt (Fig. 1a) and in a 50 MWt (Fig. 1b) Ansaldo boilers. The 6 MWt boiler is 13 m high with a section of 13 · 7 m. The 50 MWt boiler is 5.5 m high with a section of 12.5 · 4.35 m. Two lines of ‘‘Main’’ burners are placed at the bottom of the boiler. At the top of them are placed the ‘‘Reburning’’ burner ‘‘RB’’ fuelled with the same oil. Finally, combustion can be completed by the ‘‘OFA’’,Over-Fire Air system. The combustion air is preheated at 300 C. The ‘‘OFA’’ quantity and ‘‘RB’’ burners were changed in order to obtain oxygen concentration at the exhaust ranging between 0.5% and 2.5% in order to study the effect of these systems on the particle formation and their influence on the metallic content of the particles.
OFA REBURNING
MAIN
Fig. 1a. 6 MWt boiler scheme.
Fig. 1b. 50 MWt boiler overview.
The 50 MWt plant is a test rig boiler that can be fuelled with different fuels (coal, heavy oils, etc.). The fuels used during the experiments are a low-sulphur content oil, ‘‘BTZ’’, and an emulsion BTZ-water (7.5% v). The heavy oil composition is presented in Table 1. A PentolTM additive stabilized the emulsion. In this plant the oil and its emulsion were burned under the same conditions changing only the oxygen concentration at the exhausts from 0.5% to 2.5%. Different combustion regimes were investigated to study the effects of emulsion on the emission reduction of the fine particle and on the metal content of the particles.
Table 1 Heavy fuel oil characteristics Parameter
Method
Value
Water Ash Carbon Hydrogen Oxygen Nitrogen Sulphur Asphaltens
ISO 3733 EN ISO 6245 ASTM 5291 ASTM 5291 ASTM 5291 ASTM 5291 EN ISO 8754 IP 143
<0.5% w 0.058% w 86.61% w 10.94% w 1.15% w 0.34% w 0.90% w 5.95 m/m
PCS Density@120 C Viscosity@120 C
ASTM D 240 ASTM D 1298 EN ISO 3104
10409 kcal/kg 980.3 kg/dm3 3 E
Vanadium Sodium Nickel Magnesium Lead Manganese Aluminium Cadmium Chromo Zinc Iron
Atomic Atomic Atomic Atomic Atomic Atomic Atomic Atomic Atomic Atomic Atomic
51 mg/kg 17.8 mg/kg 12.9 mg/kg 2.37 mg/kg 22 mg/kg 5.4 mg/kg 1.1 mg/kg 0.05 mg/kg 2.4 mg/kg 137 mg/kg 92 mg/kg
absorption absorption absorption absorption absorption absorption absorption absorption absorption absorption absorption
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426
Isokinetic samplings were performed at the exhausts of the boiler under the same atomization conditions. The samples were size-segregated by means of an 8-stages AndersenTM impactor according to the EPA method 5. This impactor is designed to collect samples on nine stages (including the back-up filter) ranging between 0.5 and 20 lm in diameter for subsequent gravimetric and Scanning Electronic Microscopy and EDX analysis. Because of the too small existing stack, the sampling nozzle was extended to allow external isokinetic sampling. The entire impactor and sampling nozzle were heated to avoid cooling down of the sampled gas. Samples were also analyzed by a SEM equipped with an energy dispersive X-ray (EDX) spectrometer that provided both morphological information of individual particles and their chemical composition. 3. Results and discussion The particulate distribution functions of the sampled material at the exhausts burning the heavy oil and its emulsion for the 50 MWt plant versus the oxygen concentration are presented in Fig. 2. We observe, as expected, that particulate concentrations are reduced increasing the combustion air for all conditions except the case of emulsion at 2.5% O2 that produces more fine particles than at lower oxygen concentration. An interesting results stands in the fact that the use of the emulsion reduces up to three times the emission of fine particle (D50 < 8 lm). This benefit is not appreciable for larger particles. This phenomenon is probably due to the processes involved during the particle formation. In fact, the formation of larger particulate is strongly related to the asphaltene content of the parent
423
oil [7,8], while the formation of the smallest particles is conditioned by flame temperature and in the case of emulsion micro-explosion phenomenon due to the presence of water. A further direct comparison between the size distributions obtained for the oil and its emulsion is presented in Fig. 3 that evidence more clearly that the emulsion produces less particulate than the parent oil, especially in the range of aerodynamic diameters 1–8 lm for 0.5% and 1.5% O2 concentrations. All the distributions apparently seem to be bimodal. Further capability of emulsion was tested in the other boiler equipped with the traditional technology one could find in industrial power plants, namely OFA and reburning systems. So, the performance of oil and its emulsion were also tested in the 6 MWt plant in normal condition (0.5% O2 without reburning and OFA), normal condition plus 35% of the total air fed to the OFA system and finally, including the previous OFA condition plus 10% of the fuel flow rate fuelled to the reburning burners. A direct comparison between the size distributions obtained for the oil and its emulsion is presented in Fig. 4. On this figure we first note that both fuels present similar distribution functions.
10 0.5% O2
Emulsion Oil
8 6 4 2 0
10 Oil
2.5% O 2
1.5% O2
1.5% O 2
8
3
mg/Nm3
0.5% O 2
6
2
4 1
mg/Nm3
2 0
0 2.5% O2
Emulsion
3
4 2
1
2
0
0
0
0
2
4
6
8
10
12
14
16
D50, μm Fig. 2. Size distribution functions obtained at the stack of the 50 MWt plant versus oxygen concentration for both emulsion and parent oil.
2
4
6
8
10
12
14
16
D50, μm Fig. 3. Comparison of the size distribution functions obtained at the stack of the 50 MWt plant versus oxygen concentration for both emulsion and parent oil.
424
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426 Table 2 Typical elemental composition of a 1 lm plerosphere
6 0.5% O2
4
2
0
Element
wt.%
C Na Mg S Ca V Fe Ni
23.8 22 6.2 2.8 4.8 7.6 25.0 2.8
0.5% O2 + 35% OFA
mg/Nm3
6 Oil Emulsion
4
2
0 0.5% O2 + 35% OFA + 10 % Reburning
8
4
0 0
2
4
6
8
10
12
14
16
D50, μm Fig. 4. Comparison of the size distribution functions obtained at the stack of the 6 MWt plant versus oxygen concentration for both emulsion and parent oil.
The introduction of the OFA system leads to an increase of the fine particle concentration for the emulsion, while for the oil the distribution is quite the same. This slight increase for the emulsion is probably due to lower temperature measured in the boiler. Finally, the further introduction of reburning drastically increased the emission levels
for the oil; while in the case of emulsion this phenomenon does not occur. In fact, the emulsion distribution is slightly influenced for fine particles (D50 < 8 lm) and not at all for larger ones. The increase of particle emission for the oil is probably due to a lack of optimization of the reburning burners. After the weighing operation and plotting the size distribution function of the sampled material, further investigation was performed on the samples form both plants using electronic microscopy (SEM) coupled with X-ray analysis (EDX). The X-ray investigations on the probe filters evidenced the presence of solid spherical particles, or plerosphere. The typical observed dimensions ranged between 200 nm and 2–3 lm, whose atomic composition contains a large amount of the trace metals present in the parent oils (Fe, V, Ni, etc.). A mean composition of a 1 lm plerosphere obtained by EDX analysis is presented in Table 2. Photographs of typical corresponding plerosphere are showed in Fig. 5. We can note that these particles are perfectly spherical and that they do not present any porous structures on the surface. The plerospheres were identified on filter with aerodynamic diameter lower than 4 lm. Moreover, we also observed on the same analysed sample the presence of cenospheres with larger dimensions than the expected aerodynamic diameter. It is also worthwhile to note that the number of the observed plerospheres is important. From a toxicological point of view this last information is the more relevant because it determines the real effect and/or influence on human health. Another
Fig. 5. Typical photographs of observed particles: cenospheres and plerospheres.
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426
In addition, we can say that the metal concentration distribution versus particle diameter is slightly different to the one found in a smaller ceramic fibre pilot plant fuelled with similar oils [6], where the D 1 law best fitted the data and where the oxygen concentration was constant at 3%. This difference could be attributed to the lower peak temperature reached in the boiler combustion chamber due to the different kind of combustion devices and to the different residence times.
60 0.5% O 2 1.5% O 2
Metal Concentration, %w
2.5% O 2 D-1 law
45
425
30
4. Conclusions 15
0 0
4
8
12
16
20
D, μm Fig. 6. Weight content of metals obtained after EDX analysis on the sampled particles at the stack for both oil and emulsion.
interesting factor could be also the estimation of what is the real number of these particles emitted in the atmosphere. This is an hard job and was not performed so far in this paper. A further analysis performed on the sampled material showed that the total metal concentration increases as the plerosphere dimension decreases. More generally, we noted that the concentration of metals inside the micronic particles strongly decreases as the particle size increases. This tendency is shown in Fig. 6. Moreover, this figure evidences that this trend is independent of the fuel we burned and consequently no effect of emulsion can be observed in this sense. Finally, we observe that the operative conditions have a significant influence on the metal concentration in the particles. Particles collected with low oxygen concentration present a low metallic content and a high carbon content. Considering the overall investigated point, a further fitting curve of the diameter dependence of the concentration trend evidenced a D 1 law. This relationship is similar to that found for trace metal present in the coal fly ash [9] and in smaller pilot plant fired with heavy oil [6]. This D 1 law is not true anymore if we consider the low oxygen concentration condition. The process led us to think that the depletion process of the particles is not well completed in such condition. The correlation between the particle size and its metal concentration could be connected to the atomization process assuming that parent droplet contains a high percentage of asphaltenic mesophase. Thus, the metals could concentrate during the carbon oxidation, the devolatilization and the depletion of the droplet itself. This hypothesis may be corroborated by the fact that the dimensions of the asphaltenic mesophases range between 200 nm and 2–3 lm [9,10].
Investigations were performed on the particulate emitted by flames of a low-sulphur content heavy oil and its emulsion with water by mean of aerodynamic impactor and successive electronic microscopy and EDX analysis. They revealed the presence of plerospheres, whose the composition is directly correlated to their dimensions. The smaller are the particles, the greater is their metallic content. We also observed that this metallic concentration is independent of the fuel (oil or emulsion). Combustion peak temperature and the initial droplet size distribution are strongly suspected to be responsible of the formation of this class of particles. Moreover, we observed that the use of water–oil emulsion reduce the fine particle (D50 < 8 lm) concentration. Finally, the combustion of this emulsion in traditional boiler equipped with OFA and reburning systems did not give improvement respect to the parent oil in term of fine particle emission in the atmosphere. Acknowledgement The authors gratefully acknowledge the assistance of Mrs. Clelia Zucchini, Istituto di Ricerche sulla Combustione—CNR, for the SEM imaging and the EDX analyses of the samples. References [1] J.M. Samet, F. Dominici, S.L. Zeger, J. Schwartz, D.W. Dockery, The national morbidity, mortality and air pollution study, Part I: methods and methodologic issues, Research Report 94, Part I, Health Effect Institute, Cambridge MA, 2000. [2] W.P. Linak, J.O.L. Wendt, Toxic metal emissions from incineration mechanisms and control, Progress in Energy and Combustion Sciences 19 (1993) 145–185. [3] A. Williams, Fundamental of oil combustion, Progress in Energy and Combustion Sciences 2 (1976) 167–179. [4] G.L. Fischer, D.P.Y. Chang, M. Brummer, Fly ash collected from electrostatic precipitators: microcrystalline structures and the mystery of the spheres, Science 192 (1976) 553–555. [5] R.D. Smith, J.A. Campbell, K.K. Nielson, Characterization and formation of submicron particles in coal-fired plants, Atmospheric Environment 13 (1979) 607–617. [6] C. Allouis, F. Beretta, A. D’Alessio, Structure of inorganic and carbonaceous particles emitted from heavy oil combustion, Chemosphere 51 (10) (2003) 1091–1096.
426
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426
[7] R. Villasenor, F. Garcia, Experimental study of the effect of asphaltene on heavy oil fuel droplet combustion, Fuel 78 (1999) 933–944. [8] D.L. Urban, S.P. Huey, F.L. Dryer, Evaluation of the coke formation potential of residual fuel oils, in: Proceeding of the Combustion Institute, vol. 24, 1992, pp. 1357–1364.
[9] B. Haynes, M. Neville, R.J. Quann, A.F. Sarofim, Factors governing the surface enrichment of fly ash in volatile trace species, Journal of Colloid Interface Science 87 (1982) 266–278. [10] T.F. Yen, G.V. Chilingarian, Asphaltenes and Asphalts, vol. 1, Elsevier Science, 1994.
Study of water–oil emulsion combustion in large pilot power plants for fine particle matter emission reduction C. Allouis b
a,*
, A. L’Insalata b, L. Fortunato b, A. Saponaro b, F. Beretta
a
a Istituto di Ricerche sulla Combustione, C.N.R., P.le V. Tecchio 80, 80125 Napoli, Italy ANSALDO Caldaie, Centro Combustione e Ambiente, Via Milano km 1,6, 70023 Gioia del Colle, Italy
Received 10 October 2005; received in revised form 5 March 2006; accepted 30 April 2006
Abstract The combustion of heavy fuel oil for power generation is a great source of carbonaceous and inorganic particle emissions, even though the combustion technologies and their efficiency are improving. The information about the size distribution function of the particles originated by trace metals present into the fuels is not adequate. In this paper, we focused our attention the influence of emulsion oil-water on the larger distribution mode of both the carbonaceous and metallic particles. Isokinetic sampling was performed at the exhausts of flames of a low-sulphur content heavy oil and its emulsion with water produced in two large pilot plants. The samples were size-segregated by mean of an 8-stages Andersen impactor. Further investigation performed on the samples using electronic microscopy (SEM) coupled with X-ray analysis (EDX) evidenced the presence of solid spherical particles, plerosphere, with typical dimensions ranging between 200 nm and 2–3 lm, whose atomic composition contains a large amount of the trace metals present in the parent oils (Fe, V, Ni, etc.). EDX analyses revealed that the metal concentration increases as the plerosphere dimension decreases. We also observed that the use of emulsion slightly reduce the emission of fine particles (D50 < 8 lm) in the large scale plant. 2006 Elsevier Inc. All rights reserved. Keywords: Metal concentration; Heavy oil combustion; Particulate size distribution
1. Introduction It is now assessed, that small particles present in the atmosphere produce significant effects on human health even at concentration levels considered acceptable few years ago [1]. Furthermore toxicology points out the importance of the size and/or the chemical composition in the production of the damage. Particles produced by combustion processes exhibit typical sizes in the submicronic and even in the nanometric range and are consequently considered among the primary responsible of the harmful effects. Furthermore, combus-
*
Corresponding author. Tel.: +39 081 7682247; fax: +39 081 5936936. E-mail address: [email protected] (C. Allouis).
0894-1777/$ - see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.expthermflusci.2006.04.011
tion particles are composed by unburned carbon, sulphates and metals contained in the fuel itself. Heavy oils are crude liquid fractions containing high molecular mass organic structures within which inorganic functionalities are inserted. It implies that when they are introduced in the combustion chamber in form of sprays their oxidation and pyrolysis take place both in the liquid phase, inside the droplets, and in the gaseous phase around them. Consequently, the size, shape and composition of the particles are all determined by the relative roles of the different reactive channels, as it has been discussed, few years ago, by Linak and Wendt [2]. On the other extreme the pyrolysis in the liquid phase produces particles, carbonaceous and/or metallic, whose typical sizes are comparable with those of the parent droplets [3]. Many of such structures present themselves as hollow riddled spheres, called cenospheres, which are subject to fragmentation due to
422
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426
thermal stresses and attrition. However some authors have also reported the presence of denser compact particles, called plerospheres, with typical sizes of the order of few microns in the fly ashes produced during coal combustion [4,5] and heavy oil combustion [6]. The presence of these different particles have also a significant impact on the removal efficiency of the particles from the exhaust streams: larger particles, in the micron size range, are captured much more easily than those into the submicronic one. On the other side, nanometric particles escape almost without obstacles in the atmosphere. Thus, it becomes important to characterize the fly ash from the combustion of heavy oil and its emulsion, with particular emphases on particle morphology and the size distribution of potentially toxic elements. Therefore, it is a specific goal of this communication to analyze the sizes, shapes and metal contents of the larger and intermediate distribution modes of particles present at the exhaust of two pilot plants (6 MWt and 50 MWt) fired with heavy oil and its emulsion with water. 2. Experimental set-up Experiments were carried out in a 6 MWt (Fig. 1a) and in a 50 MWt (Fig. 1b) Ansaldo boilers. The 6 MWt boiler is 13 m high with a section of 13 · 7 m. The 50 MWt boiler is 5.5 m high with a section of 12.5 · 4.35 m. Two lines of ‘‘Main’’ burners are placed at the bottom of the boiler. At the top of them are placed the ‘‘Reburning’’ burner ‘‘RB’’ fuelled with the same oil. Finally, combustion can be completed by the ‘‘OFA’’,Over-Fire Air system. The combustion air is preheated at 300 C. The ‘‘OFA’’ quantity and ‘‘RB’’ burners were changed in order to obtain oxygen concentration at the exhaust ranging between 0.5% and 2.5% in order to study the effect of these systems on the particle formation and their influence on the metallic content of the particles.
OFA REBURNING
MAIN
Fig. 1a. 6 MWt boiler scheme.
Fig. 1b. 50 MWt boiler overview.
The 50 MWt plant is a test rig boiler that can be fuelled with different fuels (coal, heavy oils, etc.). The fuels used during the experiments are a low-sulphur content oil, ‘‘BTZ’’, and an emulsion BTZ-water (7.5% v). The heavy oil composition is presented in Table 1. A PentolTM additive stabilized the emulsion. In this plant the oil and its emulsion were burned under the same conditions changing only the oxygen concentration at the exhausts from 0.5% to 2.5%. Different combustion regimes were investigated to study the effects of emulsion on the emission reduction of the fine particle and on the metal content of the particles.
Table 1 Heavy fuel oil characteristics Parameter
Method
Value
Water Ash Carbon Hydrogen Oxygen Nitrogen Sulphur Asphaltens
ISO 3733 EN ISO 6245 ASTM 5291 ASTM 5291 ASTM 5291 ASTM 5291 EN ISO 8754 IP 143
<0.5% w 0.058% w 86.61% w 10.94% w 1.15% w 0.34% w 0.90% w 5.95 m/m
PCS Density@120 C Viscosity@120 C
ASTM D 240 ASTM D 1298 EN ISO 3104
10409 kcal/kg 980.3 kg/dm3 3 E
Vanadium Sodium Nickel Magnesium Lead Manganese Aluminium Cadmium Chromo Zinc Iron
Atomic Atomic Atomic Atomic Atomic Atomic Atomic Atomic Atomic Atomic Atomic
51 mg/kg 17.8 mg/kg 12.9 mg/kg 2.37 mg/kg 22 mg/kg 5.4 mg/kg 1.1 mg/kg 0.05 mg/kg 2.4 mg/kg 137 mg/kg 92 mg/kg
absorption absorption absorption absorption absorption absorption absorption absorption absorption absorption absorption
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426
Isokinetic samplings were performed at the exhausts of the boiler under the same atomization conditions. The samples were size-segregated by means of an 8-stages AndersenTM impactor according to the EPA method 5. This impactor is designed to collect samples on nine stages (including the back-up filter) ranging between 0.5 and 20 lm in diameter for subsequent gravimetric and Scanning Electronic Microscopy and EDX analysis. Because of the too small existing stack, the sampling nozzle was extended to allow external isokinetic sampling. The entire impactor and sampling nozzle were heated to avoid cooling down of the sampled gas. Samples were also analyzed by a SEM equipped with an energy dispersive X-ray (EDX) spectrometer that provided both morphological information of individual particles and their chemical composition. 3. Results and discussion The particulate distribution functions of the sampled material at the exhausts burning the heavy oil and its emulsion for the 50 MWt plant versus the oxygen concentration are presented in Fig. 2. We observe, as expected, that particulate concentrations are reduced increasing the combustion air for all conditions except the case of emulsion at 2.5% O2 that produces more fine particles than at lower oxygen concentration. An interesting results stands in the fact that the use of the emulsion reduces up to three times the emission of fine particle (D50 < 8 lm). This benefit is not appreciable for larger particles. This phenomenon is probably due to the processes involved during the particle formation. In fact, the formation of larger particulate is strongly related to the asphaltene content of the parent
423
oil [7,8], while the formation of the smallest particles is conditioned by flame temperature and in the case of emulsion micro-explosion phenomenon due to the presence of water. A further direct comparison between the size distributions obtained for the oil and its emulsion is presented in Fig. 3 that evidence more clearly that the emulsion produces less particulate than the parent oil, especially in the range of aerodynamic diameters 1–8 lm for 0.5% and 1.5% O2 concentrations. All the distributions apparently seem to be bimodal. Further capability of emulsion was tested in the other boiler equipped with the traditional technology one could find in industrial power plants, namely OFA and reburning systems. So, the performance of oil and its emulsion were also tested in the 6 MWt plant in normal condition (0.5% O2 without reburning and OFA), normal condition plus 35% of the total air fed to the OFA system and finally, including the previous OFA condition plus 10% of the fuel flow rate fuelled to the reburning burners. A direct comparison between the size distributions obtained for the oil and its emulsion is presented in Fig. 4. On this figure we first note that both fuels present similar distribution functions.
10 0.5% O2
Emulsion Oil
8 6 4 2 0
10 Oil
2.5% O 2
1.5% O2
1.5% O 2
8
3
mg/Nm3
0.5% O 2
6
2
4 1
mg/Nm3
2 0
0 2.5% O2
Emulsion
3
4 2
1
2
0
0
0
0
2
4
6
8
10
12
14
16
D50, μm Fig. 2. Size distribution functions obtained at the stack of the 50 MWt plant versus oxygen concentration for both emulsion and parent oil.
2
4
6
8
10
12
14
16
D50, μm Fig. 3. Comparison of the size distribution functions obtained at the stack of the 50 MWt plant versus oxygen concentration for both emulsion and parent oil.
424
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426 Table 2 Typical elemental composition of a 1 lm plerosphere
6 0.5% O2
4
2
0
Element
wt.%
C Na Mg S Ca V Fe Ni
23.8 22 6.2 2.8 4.8 7.6 25.0 2.8
0.5% O2 + 35% OFA
mg/Nm3
6 Oil Emulsion
4
2
0 0.5% O2 + 35% OFA + 10 % Reburning
8
4
0 0
2
4
6
8
10
12
14
16
D50, μm Fig. 4. Comparison of the size distribution functions obtained at the stack of the 6 MWt plant versus oxygen concentration for both emulsion and parent oil.
The introduction of the OFA system leads to an increase of the fine particle concentration for the emulsion, while for the oil the distribution is quite the same. This slight increase for the emulsion is probably due to lower temperature measured in the boiler. Finally, the further introduction of reburning drastically increased the emission levels
for the oil; while in the case of emulsion this phenomenon does not occur. In fact, the emulsion distribution is slightly influenced for fine particles (D50 < 8 lm) and not at all for larger ones. The increase of particle emission for the oil is probably due to a lack of optimization of the reburning burners. After the weighing operation and plotting the size distribution function of the sampled material, further investigation was performed on the samples form both plants using electronic microscopy (SEM) coupled with X-ray analysis (EDX). The X-ray investigations on the probe filters evidenced the presence of solid spherical particles, or plerosphere. The typical observed dimensions ranged between 200 nm and 2–3 lm, whose atomic composition contains a large amount of the trace metals present in the parent oils (Fe, V, Ni, etc.). A mean composition of a 1 lm plerosphere obtained by EDX analysis is presented in Table 2. Photographs of typical corresponding plerosphere are showed in Fig. 5. We can note that these particles are perfectly spherical and that they do not present any porous structures on the surface. The plerospheres were identified on filter with aerodynamic diameter lower than 4 lm. Moreover, we also observed on the same analysed sample the presence of cenospheres with larger dimensions than the expected aerodynamic diameter. It is also worthwhile to note that the number of the observed plerospheres is important. From a toxicological point of view this last information is the more relevant because it determines the real effect and/or influence on human health. Another
Fig. 5. Typical photographs of observed particles: cenospheres and plerospheres.
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426
In addition, we can say that the metal concentration distribution versus particle diameter is slightly different to the one found in a smaller ceramic fibre pilot plant fuelled with similar oils [6], where the D 1 law best fitted the data and where the oxygen concentration was constant at 3%. This difference could be attributed to the lower peak temperature reached in the boiler combustion chamber due to the different kind of combustion devices and to the different residence times.
60 0.5% O 2 1.5% O 2
Metal Concentration, %w
2.5% O 2 D-1 law
45
425
30
4. Conclusions 15
0 0
4
8
12
16
20
D, μm Fig. 6. Weight content of metals obtained after EDX analysis on the sampled particles at the stack for both oil and emulsion.
interesting factor could be also the estimation of what is the real number of these particles emitted in the atmosphere. This is an hard job and was not performed so far in this paper. A further analysis performed on the sampled material showed that the total metal concentration increases as the plerosphere dimension decreases. More generally, we noted that the concentration of metals inside the micronic particles strongly decreases as the particle size increases. This tendency is shown in Fig. 6. Moreover, this figure evidences that this trend is independent of the fuel we burned and consequently no effect of emulsion can be observed in this sense. Finally, we observe that the operative conditions have a significant influence on the metal concentration in the particles. Particles collected with low oxygen concentration present a low metallic content and a high carbon content. Considering the overall investigated point, a further fitting curve of the diameter dependence of the concentration trend evidenced a D 1 law. This relationship is similar to that found for trace metal present in the coal fly ash [9] and in smaller pilot plant fired with heavy oil [6]. This D 1 law is not true anymore if we consider the low oxygen concentration condition. The process led us to think that the depletion process of the particles is not well completed in such condition. The correlation between the particle size and its metal concentration could be connected to the atomization process assuming that parent droplet contains a high percentage of asphaltenic mesophase. Thus, the metals could concentrate during the carbon oxidation, the devolatilization and the depletion of the droplet itself. This hypothesis may be corroborated by the fact that the dimensions of the asphaltenic mesophases range between 200 nm and 2–3 lm [9,10].
Investigations were performed on the particulate emitted by flames of a low-sulphur content heavy oil and its emulsion with water by mean of aerodynamic impactor and successive electronic microscopy and EDX analysis. They revealed the presence of plerospheres, whose the composition is directly correlated to their dimensions. The smaller are the particles, the greater is their metallic content. We also observed that this metallic concentration is independent of the fuel (oil or emulsion). Combustion peak temperature and the initial droplet size distribution are strongly suspected to be responsible of the formation of this class of particles. Moreover, we observed that the use of water–oil emulsion reduce the fine particle (D50 < 8 lm) concentration. Finally, the combustion of this emulsion in traditional boiler equipped with OFA and reburning systems did not give improvement respect to the parent oil in term of fine particle emission in the atmosphere. Acknowledgement The authors gratefully acknowledge the assistance of Mrs. Clelia Zucchini, Istituto di Ricerche sulla Combustione—CNR, for the SEM imaging and the EDX analyses of the samples. References [1] J.M. Samet, F. Dominici, S.L. Zeger, J. Schwartz, D.W. Dockery, The national morbidity, mortality and air pollution study, Part I: methods and methodologic issues, Research Report 94, Part I, Health Effect Institute, Cambridge MA, 2000. [2] W.P. Linak, J.O.L. Wendt, Toxic metal emissions from incineration mechanisms and control, Progress in Energy and Combustion Sciences 19 (1993) 145–185. [3] A. Williams, Fundamental of oil combustion, Progress in Energy and Combustion Sciences 2 (1976) 167–179. [4] G.L. Fischer, D.P.Y. Chang, M. Brummer, Fly ash collected from electrostatic precipitators: microcrystalline structures and the mystery of the spheres, Science 192 (1976) 553–555. [5] R.D. Smith, J.A. Campbell, K.K. Nielson, Characterization and formation of submicron particles in coal-fired plants, Atmospheric Environment 13 (1979) 607–617. [6] C. Allouis, F. Beretta, A. D’Alessio, Structure of inorganic and carbonaceous particles emitted from heavy oil combustion, Chemosphere 51 (10) (2003) 1091–1096.
426
C. Allouis et al. / Experimental Thermal and Fluid Science 31 (2007) 421–426
[7] R. Villasenor, F. Garcia, Experimental study of the effect of asphaltene on heavy oil fuel droplet combustion, Fuel 78 (1999) 933–944. [8] D.L. Urban, S.P. Huey, F.L. Dryer, Evaluation of the coke formation potential of residual fuel oils, in: Proceeding of the Combustion Institute, vol. 24, 1992, pp. 1357–1364.
[9] B. Haynes, M. Neville, R.J. Quann, A.F. Sarofim, Factors governing the surface enrichment of fly ash in volatile trace species, Journal of Colloid Interface Science 87 (1982) 266–278. [10] T.F. Yen, G.V. Chilingarian, Asphaltenes and Asphalts, vol. 1, Elsevier Science, 1994.