- Email: [email protected]
h scale pyrolysis-melting incinerator
Available online at www.sciencedirect.com
Waste Management 28 (2008) 2422–2427 www.elsevier.com/locate/wasman
Combustion characteristics of simulated gas fuel in a 30 kg/h scale pyrolysis-melting incinerator D. Shin a, T. Yu b, W. Yang b, B. Jeon c, S. Park c, J. Hwang c,* a b
School of Mechanical & Automotive Engineering, Kookmin University, Seoul 136-702, Republic of Korea Industrial Equipment Team, Korea Institute of Industrial Technology, Chonan 330-825, Republic of Korea c Department of Mechanical Engineering, Yonsei University, Seoul 120-749, Republic of Korea Accepted 25 November 2007 Available online 5 March 2008
Abstract Combustion characteristics of gas fuel in a pyrolysis-melting incinerator having a 30 kg/h capacity were investigated. Pyrolyzed gas from waste was simulated by propane that was injected in the combustion chamber, and burnt through multi-staged combustion by distributing the combustion air to primary, secondary, and tertiary air nozzles. Temperatures and the concentrations of gas components in the combustion chamber were measured. Combustion performance was evaluated with respect to the temperature distribution and combustion gas concentrations of O2, CO and NOx. Using secondary air and/or tertiary air, the combustion performance was improved, and, in particular, NOx concentration decreased significantly following the tertiary air injection. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction Pyrolysis-melting incineration is considered to be one of the alternatives to conventional waste incineration, as it removes dioxin and heavy metals, as well as other pollutants such as NOx and SOx. A typical pyrolysis-melting incinerator performs the following three processes: pyrolysis, gas combustion, and ash melting. The pyrolyzed gas of waste generated in the pyrolysis chamber, which consists of various hydrocarbons such as CH4, C3H6, C5H12, C7H8 etc., burns in the combustion chamber, and the char and ash, which are the byproducts of the pyrolysis, are introduced to the melting process where they are converted to slag in a high-temperature environment. Heat generated by the pyrolyzed gas combustion can be utilized for power generation or ash melting. Many experimental and numerical studies of conventional incinerators, including the stoker, rotary kiln and fluidized bed types, have been reported. Most results indicate
*
Corresponding author. Tel.: +82 2 2123 2821; fax: +82 2 2123 2159. E-mail address: [email protected] (J. Hwang).
0956-053X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.wasman.2007.11.016
that injection of secondary air jets significantly affects the flow pattern, mixing of gaseous species, and distribution of gas residence time in the combustion chamber (Han et al., 1996; Nasserzadeh and Swithenbank, 1991; Rogaume et al., 2002). In particular, multi-staged combustion using secondary and/or tertiary air has been applied to reduce NOx emission and to complete combustion (Ryu et al., 2002; Shin et al., 2005; Yang et al., 2003). The multi-staged combustion technique results in under-stoichiometric conditions in the first stage of combustion and supplies enough combustion air for the later stages to complete the combustion (Coghe et al., 2004; Turns, 1996). Several processes of pyrolysis-melting incineration technology have been introduced to Korea (Table 1) and are presently being verified (Fink, 1999). Recently, several commercial pyrolysis-melting systems have undergone construction in Korea and one blast furnace-type plant was operating as of September 2007. These processes were initially developed in Europe and commercialized in Japan with consideration of the characteristics of the waste produced in these regions. Operating parameters of these processes should be adjusted to the intrinsic characteristics of Korean waste.
D. Shin et al. / Waste Management 28 (2008) 2422–2427
2423
Table 1 Waste pyrolysis-melting technologies introduced to Korea Technology
Pyrolysis process
Melting process
Benefit
Weakness
Thermoselect
Compression and moving bed Rotary kiln Fluidized bed Blast furnace
Pure oxygen burber (1600 °C)
Generate high quality syngas (H2, CO) Low energy consumption
Pure oxygen consumption
R-21 Ebara Blast furnace The present study
Compression and moving bed
Combustion of pyrolyzed gas and char with air (1350 °C)
Pre-treatment of waste is necessary
Pyrolysis and melting in a chamber Oxygen enriched bubbling melting pot(1600 °C)
Relatively low energy consumption, uniform molten slag, smelting, and fully waste adaptable by oxygen concentration control
Table 2 Pyrolyzed gas composition (Park, 2006) Components Methane Ethylene Ethan Acetylene Propylene n-Pentane Toluene n-Heptane 2-Heptene Ethylbenzene Nitrogen Other
Cokes and oxygen consumption, and discrete discharge of molten slag
400 460
Volume% CH4 C2H4 C2H6 C2H2 C3H6 C5H12 C7H8 C7H16 C7H14 C8H10 N2
0.51 0.45 0.05 0.08 0.62 4.2 1.7 0.4 1.6 0.1 88.8 1.49
800
Cyclone Cyclone
360
2. Experimental setup Fig. 1 shows the schematic diagram of the pyrolysismelting incinerator designed especially for this study. A rectangular parallel-piped pyrolysis chamber is connected to the combustion chamber. In the pyrolysis chamber, waste is compressed by a feeder and pyrolyzed in a reduced atmosphere by 16 plates of an electric heater of 3 kW each.
1,000
800
Waste Pyrolysis chamber 120
This study aims to develop a novel pyrolysis-melting incineration process considering the intrinsic characteristics of waste generated in Korea. A 30 kg/h pyrolysis-melting incinerator was designed, constructed and pre-operated. The effects of secondary and tertiary air on flow pattern, mixing, and NOx emissions of the combustion chamber were investigated. The pyrolyzed gas consists of various hydrocarbons (Table 2), whose composition is dependent on pyrolyzing conditions such as waste composition, temperature, pyrolyzer size, etc. (Park, 2006). In the present study, the pyrolyzed gas was simulated by propane, a mid-sized molecule among all the components. The propane was injected in the combustion chamber, and burnt through multi-staged combustion by distributing the combustion air to primary, secondary, and tertiary air nozzles. Temperatures and gas components in the combustion chamber were measured, and combustion performance was determined by temperature distribution and O2, CO and NOx chemical species concentrations.
250
200 550
Combustion chamber
1,300
2,000
Feeder Melting furnace
1,000
Oil burner Fig. 1. Schematic diagram of 30 kg/h pyrolysis-melting incinerator.
The cylindrical combustion chamber has a height of 3850 mm and inner diameter of 460 mm. The particles present in the combustion gas are collected by a standard cyclone (Nevers, 1995). In the combustion chamber, two-staged secondary air and tertiary air nozzles having 10-mm inner diameters are installed. The nozzles for secondary air are positioned around the chamber and tilted 10° from the radial direction of the chamber to generate swirl for improvement of mixing and residence time, as shown in Fig. 2. Tertiary air nozzles are located 800 mm above the second stage of the secondary air nozzles. A light oil burner was used for initial heating and to simulate the gas from the melting pod. In this study, neither a pyrolysis nor melting furnace was operated. Therefore, propane was directly supplied instead of pyrolyzed gas at the combustor inlet, as shown in Fig. 3 (experimental setup). The effects of multi-staged combustion were examined by measurement of temperatures and gas compositions. Three ports were selected for both
2424
D. Shin et al. / Waste Management 28 (2008) 2422–2427
temperature and gas species concentration measurements in the combustion chamber, and four K-type thermocouples were installed along the radial direction for each port. For calibration of the measured temperature considering radiative heat loss, the following equation was used:
Combustion chamber o
Secondary air nozzle -2
120oo
o
60o
10o
Secondary air nozzle -1
Tg Tt ¼
o
120o
Fig. 2. Configuration of secondary air nozzles around the combustion chamber.
Port 33 Port
Thermo -stack
Tertiary air
Port 2 Port2
Secondary air 2 Computer DAQ
ð1Þ
where Tg is the actual gas temperature in the combustion chamber, Tt the thermocouple bead temperature, Ts the inner wall temperature, e is the emissivity set to 0.71, r is the Stefan-Boltzman constant, d is the diameter of thermocouple bead, and kg (W/m K) is the thermal conductivity of the gas. The unit for temperature is K. Flue gas samplings were performed in the measurement ports (10 cm of inner diameter) by using a water-cooled gas probe (3 cm outer diameter). In each measurement port, gas compositions were measured at five points along a linear direction at the cross section (Fig. 4a–e), where gas concentrations were adjusted by the surface area of each section (x, y, z). The adjusted concentration a of the gas species i was obtained by Eq. (2):
Port 1 Port1
Secondary air 1
erdðT 4t T 4s Þ 2k g
ai ¼
i i ci x þ ðbi þd Þy þ ðai þe Þz 2 2 xþyþz
ð2Þ
Propane
Gas analyzer
where ai, bi, ci, di, and ei are the concentrations of gas species i at each point, and x, y, and z are surface areas of the zone that contains the measurement point, as presented in Fig. 4. Values of x, y, z are 79 cm2, 537 cm2, and 1046 cm2, respectively. The concentrations of O2, CO, and NOx were measured by electro-chemical sensors on the Greenline gas analyzer (MKII). The measurement ranges were 0–25% for O2, 0–20,000 ppm for CO, and 0–2000 ppm for NOx. The temperatures and the gas compositions were continuously monitored by a data acquisition system (DA-100, National Instrument Inc.). The measurement time of each item was 10 min, and the average value was adopted for analysis.
Primary air Oil burner
Fig. 3. Experimental setup.
3. Experimental cases
Fig. 4. Measurement points at a cross-section of each measurement port level.
Four different operating conditions were selected to evaluate combustion air distribution. Table 3 summarizes the flow rates of the primary, secondary and tertiary air for each of the cases. The flow rates of the total combustion air, primary air and propane were fixed, maintaining an excess air ratio of 20% to ensure a theoretical oxygen concentration of 4% at the combustor exit. The secondary and
Table 3 Flow rates of air and fuel Case Total air (lpm)
Primary air (lpm)
Secondary air – 1 (lpm)
Secondary air – 2 (lpm)
Tertiary air (lpm)
Propane (lpm)
Light oil (g/min)
Product gas (lpm)
GF1 2740 GF2 GF3 GF4
1542
464 3573 271 237
464 357 271 237
0 213 385 454
30 (714*)
125 (1480*)
3058
*
Theoretical air rate (lpm) (Reed, 1978).
D. Shin et al. / Waste Management 28 (2008) 2422–2427
o
Temperature ( C)
1400
a
Case GF1 GF2 GF3 GF4 , , , Port 1 Port 2 , , , Port 3 , , ,
5
O2 concentration (%)
a
1200
1000
2425
4
3
Port 1 Port 2 Port 3
2
1
800 0
-20
-10
0
10
20
-20
Case GF1 Case GF2 Case GF3 Case GF4
o
Temperature ( C)
1400
b
0
10
20
1000
CO concentration (ppm)
b
-10
Measurement points (cm)
Measurement point (cm)
1200
1000
800
Port 2 Port 3
600
400
200
800
Port 2
0
Port 3
Fig. 5. Temperature distribution in the combustion chamber: (a) temperature distribution along the radial location and (b) area-averaged temperature along the axial distance.
tertiary airflow rates were varied for each case. The tertiary air was not supplied for Case GF1, and the tertiary airflow rate increased gradually from GF2 to GF4. A light oil burner was applied to maintain the primary chamber temperature over 1100 °C. 4. Results and discussion
-20
-10
0
10
20
Measurement points (cm)
c
100
NOx concentration (ppm)
Port 1
80
Port 1 Port 2 Port 3
60
40
20
Fig. 5a shows the radial distributions of temperatures measured at three ports in the combustion chamber for the cases described in Table 3. For any case and any port location, temperatures were uniform along the radial distance due to the high mixing performance from the secondary air injection. The temperatures at Ports 1 and 2 did not show significant differences for all cases, implying that the tertiary air did not directly affect the temperature distribution of Ports 1 and 2. The temperatures at Port 3 were lower than those of Ports 1 and 2 due to convective/radiative heat loss, as shown in Fig. 5b. Fig. 6 shows the O2, CO and NOx concentrations at all measurement points for Case GF4. O2 levels at Ports 1 and 2 have similar values of 1 ± 0.5%. Ideally, the stoichiometric air rate is approximately 2200 lpm to completely burn the propane (30 lpm) and light oil (125 g/min). The zone of Port 2 is located in the fuel rich condition as the total air
0 -20
-10
0
10
20
Measurement points (cm)
Fig. 6. Concentrations of flue gas species for Case GF4: (a) O2 concentrations, (b) CO concentrations and (c) NOx concentrations.
injected by the primary air and the secondary air 1 and 2 are 2016 lpm (Table 3). However, the actual mixing by secondary air jets cannot be ideally completed and the combustion continues such that the oxygen concentration is not 0% (Reed, 1978). At Port 3, the oxygen concentration increases to 4% as the tertiary air is injected, suggesting that the combustion is completed. CO levels measured at Port 1 exceeded the measurement limit of the gas analyzer (20,000 ppm). The levels of CO were sharply decreased at Port 2 due to secondary air injection. At Port
D. Shin et al. / Waste Management 28 (2008) 2422–2427
a
5
O2 concentration (%)
3, combustion was completed and the CO level was lower than 100 ppm. Fig. 6c shows that the NOx concentrations at Port 1 were in the range of 0–20 ppm. Even though the gas temperature is 11001200 °C at Port 1, starved oxygen conditions did not generate NOx. Meanwhile, the NOx concentrations at Port 2 increased to 40–50 ppm due to the sufficient air supply from the secondary air nozzle at the high-temperature range around 1100–1200 °C. However, this value is still low compared to an ordinary combustion system, which is around 100–200 ppm (Cho et al., 2000). The NOx concentration at Port 3 decreased due to the tertiary air. As shown in Fig. 7, the tertiary air is injected through the center tube with uniform distribution along the flow path, resulting in gradual mixing and reaction with air as the gas flows. This slow mixing characteristic reduces the hot zone by preventing rapid reaction. Fig. 8 shows the area-averaged concentrations of O2 and CO at the three ports for all cases. At Port 1, O2 concentrations were lower than 1%, and CO concentrations exceeded the measurement limit (20,000 ppm) for all cases because of the fuel-rich condition of the first combustion stage. Due to the injections of the secondary and the tertiary air, the CO concentrations decreased sharply and the O2 concentrations increased at Ports 2 and 3. However, O2 concentrations at Port 2 in Cases GF3 and GF4 were lower than those of Cases GF1 and GF2, suggesting that a fuel-rich condition was generated at Port 2 due to the distribution of the secondary air to the tertiary air. As the stoichiometric air rate was about 2200 lpm, sufficient air was injected by the primary and secondary airflows for case GF1 (2470 lpm) and case GF2 (2250 lpm). However, for cases GF3 and GF4, the air injected by the primary and the secondary airflows were 2085 lpm and 2016 lpm, respectively, which resulted in oxygen starvation and fuel rich conditions at the Port 2 zone.
Case GF1 Case GF2 Case GF3 Case GF4
4
3
2
1
0
Port 1
b
Port 2
Port 3
1000
All CO concentrations at Port 1 exceeds the measurement limit (20,000ppm)
800
CO concentration (ppm)
2426
600
400
Case GF1 Case GF2 Case GF3 Case GF4
200
0
Port 1
Port 2
Port 3
Fig. 8. O2 and CO concentrations for Case GF1–GF4: (a) O2 concentrations and (b) CO concentrations.
Area-averaged NOx concentrations at the three ports for all cases are shown in Fig. 9. As the tertiary air increased (from GF1 to GF4), the NOx concentration at Port 3 decreased. These results show that the NOx emission was controlled by multi-staged combustion, and the effect of tertiary air was significant. As shown in Table 3, the overall air–fuel ratio was the same for all of the cases, and the oxygen concentration at the furnace exit was about 4% volumetric. Hence, the difference of NOx concentration between the cases was not only the result of dilution with air but also the
NOx concentration (ppm)
100
80
Case GF1 Case GF2 Case GF3 Case GF4
60
40
20
0
Port 1
Fig. 7. DeNOx mechanism by tertiary air injection.
Port 2
Port 3
Fig. 9. NOx concentrations for Case GF1–GF4.
D. Shin et al. / Waste Management 28 (2008) 2422–2427
NOx reduction mechanism of air distributed combustion. The tertiary air rates for GF3 and GF4 were 12.5% and 15% of total gas, respectively, so that NOx dilution effects are expected to approximate these amounts. However, the NOx reduction rates of GF3 and GF4 were 40% and 38%, respectively, suggesting that the NOx reduction was the result of not only the dilution of tertiary air but also of the combustion air distribution effect. 5. Conclusion A 30 kg/h pyrolysis-melting incineration system was designed, and the combustion characteristics of the simulated gas fuel in the combustion chamber of the incinerator were investigated as well as NOx generation/reduction characteristics. Propane was employed as a pyrolyzed gas. CO and NOx emissions significantly varied by combustion air distributions for staged combustion. The temperatures at Ports 1 and 2 were almost uniform along the measurement line, implying that the swirling flow from the secondary air injection enhanced flow uniformity. Meanwhile, the concentrations of CO at Port 1 exceeded the measurement limit (20,000 ppm) due to incomplete combustion. However, at Port 2, they were drastically reduced due to active combustion by additional air injection. At Port 3 where tertiary air was supplied, CO concentrations were under 100 ppm through completed combustion. As the tertiary air rate increased, the NOx concentration decreased, while the CO concentration at Port 3 increased slightly in case GF4. Acknowledgement This research was supported by the Strategic R&D Cluster Development Program of Seoul City 2006 and
2427
the Faculty Research Program 2007 of Kookmin University in Korea. References Cho, G.C., Choi, J.W., Park, H.K., Yoo, B.T., 2000. The evaluation of NOx emission factor from large combustion facilities in Seoul. Korean Journal of Sanitation 15, 78–83. Coghe, A., Solero, G., Scribano, G., 2004. Recirculation phenomena in a natural gas swirl combustor. Experimental Thermal and Fluid Science 28, 709–714. Fink, J.K., 1999. Pyrolysis and combustion of polymer wastes in combination with metallurgical processes and the cement industry. Journal of Analytical and Applied Pyrolysis 51, 239–252. Han, J.H., Jeong, K.K., Choi, J.H., Choi, S., 1996. A hot flow model analysis of MSW incinerator. International Journal of Energy Research 21, 899–910. Nasserzadeh, V., Swithenbank, J., 1991. Design optimization of a large municipal solid waste incinerator. Waste Management 11, 249–261. Nevers, N.D., 1995. Air pollution control engineering. McGraw-Hill International Ed., 209. Park, S., 2006. Pyrolysis Characteristics of RDF in Pyrolysis Chamber of a Pyrolysis Melting Incinerator, M.Sc., Thesis, Yonsei University, Seoul, Korea. Reed, R.J., 1978. North American Combustion Handbook, second ed. North American Mfg. Co., Cleveland, USA (pp. 16–17). Rogaume, T., Auzanneau, M., Jabouille, F., Goudeau, J.C., Torero, J.L., 2002. The effects of different airflows on the formation of pollutants during waste incineration. Fuel 81, 2277–2288. Ryu, C., Shin, D., Choi, S., 2002. Combined bed combustion and gas flow simulation for a grate type incinerator. Journal of Air and Waste Management Association 52, 189–197. Shin, D., Jang, S.D., Hwang, J., 2005. Combustion characteristics of paper mill sludge in a combustor with internally cycloned circulating fluidized bed. Waste Management 25, 680–685. Turns, S.R., 1996. An Introduction to Combustion, first ed. McGraw-Hill, 499. Yang, Y., Reuter, M.A., Hartman, D.T.M., 2003. CFD Modelling for control of hazardous waste incinerator. Control Engineering Practice 11, 93–101.
Waste Management 28 (2008) 2422–2427 www.elsevier.com/locate/wasman
Combustion characteristics of simulated gas fuel in a 30 kg/h scale pyrolysis-melting incinerator D. Shin a, T. Yu b, W. Yang b, B. Jeon c, S. Park c, J. Hwang c,* a b
School of Mechanical & Automotive Engineering, Kookmin University, Seoul 136-702, Republic of Korea Industrial Equipment Team, Korea Institute of Industrial Technology, Chonan 330-825, Republic of Korea c Department of Mechanical Engineering, Yonsei University, Seoul 120-749, Republic of Korea Accepted 25 November 2007 Available online 5 March 2008
Abstract Combustion characteristics of gas fuel in a pyrolysis-melting incinerator having a 30 kg/h capacity were investigated. Pyrolyzed gas from waste was simulated by propane that was injected in the combustion chamber, and burnt through multi-staged combustion by distributing the combustion air to primary, secondary, and tertiary air nozzles. Temperatures and the concentrations of gas components in the combustion chamber were measured. Combustion performance was evaluated with respect to the temperature distribution and combustion gas concentrations of O2, CO and NOx. Using secondary air and/or tertiary air, the combustion performance was improved, and, in particular, NOx concentration decreased significantly following the tertiary air injection. Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction Pyrolysis-melting incineration is considered to be one of the alternatives to conventional waste incineration, as it removes dioxin and heavy metals, as well as other pollutants such as NOx and SOx. A typical pyrolysis-melting incinerator performs the following three processes: pyrolysis, gas combustion, and ash melting. The pyrolyzed gas of waste generated in the pyrolysis chamber, which consists of various hydrocarbons such as CH4, C3H6, C5H12, C7H8 etc., burns in the combustion chamber, and the char and ash, which are the byproducts of the pyrolysis, are introduced to the melting process where they are converted to slag in a high-temperature environment. Heat generated by the pyrolyzed gas combustion can be utilized for power generation or ash melting. Many experimental and numerical studies of conventional incinerators, including the stoker, rotary kiln and fluidized bed types, have been reported. Most results indicate
*
Corresponding author. Tel.: +82 2 2123 2821; fax: +82 2 2123 2159. E-mail address: [email protected] (J. Hwang).
0956-053X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.wasman.2007.11.016
that injection of secondary air jets significantly affects the flow pattern, mixing of gaseous species, and distribution of gas residence time in the combustion chamber (Han et al., 1996; Nasserzadeh and Swithenbank, 1991; Rogaume et al., 2002). In particular, multi-staged combustion using secondary and/or tertiary air has been applied to reduce NOx emission and to complete combustion (Ryu et al., 2002; Shin et al., 2005; Yang et al., 2003). The multi-staged combustion technique results in under-stoichiometric conditions in the first stage of combustion and supplies enough combustion air for the later stages to complete the combustion (Coghe et al., 2004; Turns, 1996). Several processes of pyrolysis-melting incineration technology have been introduced to Korea (Table 1) and are presently being verified (Fink, 1999). Recently, several commercial pyrolysis-melting systems have undergone construction in Korea and one blast furnace-type plant was operating as of September 2007. These processes were initially developed in Europe and commercialized in Japan with consideration of the characteristics of the waste produced in these regions. Operating parameters of these processes should be adjusted to the intrinsic characteristics of Korean waste.
D. Shin et al. / Waste Management 28 (2008) 2422–2427
2423
Table 1 Waste pyrolysis-melting technologies introduced to Korea Technology
Pyrolysis process
Melting process
Benefit
Weakness
Thermoselect
Compression and moving bed Rotary kiln Fluidized bed Blast furnace
Pure oxygen burber (1600 °C)
Generate high quality syngas (H2, CO) Low energy consumption
Pure oxygen consumption
R-21 Ebara Blast furnace The present study
Compression and moving bed
Combustion of pyrolyzed gas and char with air (1350 °C)
Pre-treatment of waste is necessary
Pyrolysis and melting in a chamber Oxygen enriched bubbling melting pot(1600 °C)
Relatively low energy consumption, uniform molten slag, smelting, and fully waste adaptable by oxygen concentration control
Table 2 Pyrolyzed gas composition (Park, 2006) Components Methane Ethylene Ethan Acetylene Propylene n-Pentane Toluene n-Heptane 2-Heptene Ethylbenzene Nitrogen Other
Cokes and oxygen consumption, and discrete discharge of molten slag
400 460
Volume% CH4 C2H4 C2H6 C2H2 C3H6 C5H12 C7H8 C7H16 C7H14 C8H10 N2
0.51 0.45 0.05 0.08 0.62 4.2 1.7 0.4 1.6 0.1 88.8 1.49
800
Cyclone Cyclone
360
2. Experimental setup Fig. 1 shows the schematic diagram of the pyrolysismelting incinerator designed especially for this study. A rectangular parallel-piped pyrolysis chamber is connected to the combustion chamber. In the pyrolysis chamber, waste is compressed by a feeder and pyrolyzed in a reduced atmosphere by 16 plates of an electric heater of 3 kW each.
1,000
800
Waste Pyrolysis chamber 120
This study aims to develop a novel pyrolysis-melting incineration process considering the intrinsic characteristics of waste generated in Korea. A 30 kg/h pyrolysis-melting incinerator was designed, constructed and pre-operated. The effects of secondary and tertiary air on flow pattern, mixing, and NOx emissions of the combustion chamber were investigated. The pyrolyzed gas consists of various hydrocarbons (Table 2), whose composition is dependent on pyrolyzing conditions such as waste composition, temperature, pyrolyzer size, etc. (Park, 2006). In the present study, the pyrolyzed gas was simulated by propane, a mid-sized molecule among all the components. The propane was injected in the combustion chamber, and burnt through multi-staged combustion by distributing the combustion air to primary, secondary, and tertiary air nozzles. Temperatures and gas components in the combustion chamber were measured, and combustion performance was determined by temperature distribution and O2, CO and NOx chemical species concentrations.
250
200 550
Combustion chamber
1,300
2,000
Feeder Melting furnace
1,000
Oil burner Fig. 1. Schematic diagram of 30 kg/h pyrolysis-melting incinerator.
The cylindrical combustion chamber has a height of 3850 mm and inner diameter of 460 mm. The particles present in the combustion gas are collected by a standard cyclone (Nevers, 1995). In the combustion chamber, two-staged secondary air and tertiary air nozzles having 10-mm inner diameters are installed. The nozzles for secondary air are positioned around the chamber and tilted 10° from the radial direction of the chamber to generate swirl for improvement of mixing and residence time, as shown in Fig. 2. Tertiary air nozzles are located 800 mm above the second stage of the secondary air nozzles. A light oil burner was used for initial heating and to simulate the gas from the melting pod. In this study, neither a pyrolysis nor melting furnace was operated. Therefore, propane was directly supplied instead of pyrolyzed gas at the combustor inlet, as shown in Fig. 3 (experimental setup). The effects of multi-staged combustion were examined by measurement of temperatures and gas compositions. Three ports were selected for both
2424
D. Shin et al. / Waste Management 28 (2008) 2422–2427
temperature and gas species concentration measurements in the combustion chamber, and four K-type thermocouples were installed along the radial direction for each port. For calibration of the measured temperature considering radiative heat loss, the following equation was used:
Combustion chamber o
Secondary air nozzle -2
120oo
o
60o
10o
Secondary air nozzle -1
Tg Tt ¼
o
120o
Fig. 2. Configuration of secondary air nozzles around the combustion chamber.
Port 33 Port
Thermo -stack
Tertiary air
Port 2 Port2
Secondary air 2 Computer DAQ
ð1Þ
where Tg is the actual gas temperature in the combustion chamber, Tt the thermocouple bead temperature, Ts the inner wall temperature, e is the emissivity set to 0.71, r is the Stefan-Boltzman constant, d is the diameter of thermocouple bead, and kg (W/m K) is the thermal conductivity of the gas. The unit for temperature is K. Flue gas samplings were performed in the measurement ports (10 cm of inner diameter) by using a water-cooled gas probe (3 cm outer diameter). In each measurement port, gas compositions were measured at five points along a linear direction at the cross section (Fig. 4a–e), where gas concentrations were adjusted by the surface area of each section (x, y, z). The adjusted concentration a of the gas species i was obtained by Eq. (2):
Port 1 Port1
Secondary air 1
erdðT 4t T 4s Þ 2k g
ai ¼
i i ci x þ ðbi þd Þy þ ðai þe Þz 2 2 xþyþz
ð2Þ
Propane
Gas analyzer
where ai, bi, ci, di, and ei are the concentrations of gas species i at each point, and x, y, and z are surface areas of the zone that contains the measurement point, as presented in Fig. 4. Values of x, y, z are 79 cm2, 537 cm2, and 1046 cm2, respectively. The concentrations of O2, CO, and NOx were measured by electro-chemical sensors on the Greenline gas analyzer (MKII). The measurement ranges were 0–25% for O2, 0–20,000 ppm for CO, and 0–2000 ppm for NOx. The temperatures and the gas compositions were continuously monitored by a data acquisition system (DA-100, National Instrument Inc.). The measurement time of each item was 10 min, and the average value was adopted for analysis.
Primary air Oil burner
Fig. 3. Experimental setup.
3. Experimental cases
Fig. 4. Measurement points at a cross-section of each measurement port level.
Four different operating conditions were selected to evaluate combustion air distribution. Table 3 summarizes the flow rates of the primary, secondary and tertiary air for each of the cases. The flow rates of the total combustion air, primary air and propane were fixed, maintaining an excess air ratio of 20% to ensure a theoretical oxygen concentration of 4% at the combustor exit. The secondary and
Table 3 Flow rates of air and fuel Case Total air (lpm)
Primary air (lpm)
Secondary air – 1 (lpm)
Secondary air – 2 (lpm)
Tertiary air (lpm)
Propane (lpm)
Light oil (g/min)
Product gas (lpm)
GF1 2740 GF2 GF3 GF4
1542
464 3573 271 237
464 357 271 237
0 213 385 454
30 (714*)
125 (1480*)
3058
*
Theoretical air rate (lpm) (Reed, 1978).
D. Shin et al. / Waste Management 28 (2008) 2422–2427
o
Temperature ( C)
1400
a
Case GF1 GF2 GF3 GF4 , , , Port 1 Port 2 , , , Port 3 , , ,
5
O2 concentration (%)
a
1200
1000
2425
4
3
Port 1 Port 2 Port 3
2
1
800 0
-20
-10
0
10
20
-20
Case GF1 Case GF2 Case GF3 Case GF4
o
Temperature ( C)
1400
b
0
10
20
1000
CO concentration (ppm)
b
-10
Measurement points (cm)
Measurement point (cm)
1200
1000
800
Port 2 Port 3
600
400
200
800
Port 2
0
Port 3
Fig. 5. Temperature distribution in the combustion chamber: (a) temperature distribution along the radial location and (b) area-averaged temperature along the axial distance.
tertiary airflow rates were varied for each case. The tertiary air was not supplied for Case GF1, and the tertiary airflow rate increased gradually from GF2 to GF4. A light oil burner was applied to maintain the primary chamber temperature over 1100 °C. 4. Results and discussion
-20
-10
0
10
20
Measurement points (cm)
c
100
NOx concentration (ppm)
Port 1
80
Port 1 Port 2 Port 3
60
40
20
Fig. 5a shows the radial distributions of temperatures measured at three ports in the combustion chamber for the cases described in Table 3. For any case and any port location, temperatures were uniform along the radial distance due to the high mixing performance from the secondary air injection. The temperatures at Ports 1 and 2 did not show significant differences for all cases, implying that the tertiary air did not directly affect the temperature distribution of Ports 1 and 2. The temperatures at Port 3 were lower than those of Ports 1 and 2 due to convective/radiative heat loss, as shown in Fig. 5b. Fig. 6 shows the O2, CO and NOx concentrations at all measurement points for Case GF4. O2 levels at Ports 1 and 2 have similar values of 1 ± 0.5%. Ideally, the stoichiometric air rate is approximately 2200 lpm to completely burn the propane (30 lpm) and light oil (125 g/min). The zone of Port 2 is located in the fuel rich condition as the total air
0 -20
-10
0
10
20
Measurement points (cm)
Fig. 6. Concentrations of flue gas species for Case GF4: (a) O2 concentrations, (b) CO concentrations and (c) NOx concentrations.
injected by the primary air and the secondary air 1 and 2 are 2016 lpm (Table 3). However, the actual mixing by secondary air jets cannot be ideally completed and the combustion continues such that the oxygen concentration is not 0% (Reed, 1978). At Port 3, the oxygen concentration increases to 4% as the tertiary air is injected, suggesting that the combustion is completed. CO levels measured at Port 1 exceeded the measurement limit of the gas analyzer (20,000 ppm). The levels of CO were sharply decreased at Port 2 due to secondary air injection. At Port
D. Shin et al. / Waste Management 28 (2008) 2422–2427
a
5
O2 concentration (%)
3, combustion was completed and the CO level was lower than 100 ppm. Fig. 6c shows that the NOx concentrations at Port 1 were in the range of 0–20 ppm. Even though the gas temperature is 11001200 °C at Port 1, starved oxygen conditions did not generate NOx. Meanwhile, the NOx concentrations at Port 2 increased to 40–50 ppm due to the sufficient air supply from the secondary air nozzle at the high-temperature range around 1100–1200 °C. However, this value is still low compared to an ordinary combustion system, which is around 100–200 ppm (Cho et al., 2000). The NOx concentration at Port 3 decreased due to the tertiary air. As shown in Fig. 7, the tertiary air is injected through the center tube with uniform distribution along the flow path, resulting in gradual mixing and reaction with air as the gas flows. This slow mixing characteristic reduces the hot zone by preventing rapid reaction. Fig. 8 shows the area-averaged concentrations of O2 and CO at the three ports for all cases. At Port 1, O2 concentrations were lower than 1%, and CO concentrations exceeded the measurement limit (20,000 ppm) for all cases because of the fuel-rich condition of the first combustion stage. Due to the injections of the secondary and the tertiary air, the CO concentrations decreased sharply and the O2 concentrations increased at Ports 2 and 3. However, O2 concentrations at Port 2 in Cases GF3 and GF4 were lower than those of Cases GF1 and GF2, suggesting that a fuel-rich condition was generated at Port 2 due to the distribution of the secondary air to the tertiary air. As the stoichiometric air rate was about 2200 lpm, sufficient air was injected by the primary and secondary airflows for case GF1 (2470 lpm) and case GF2 (2250 lpm). However, for cases GF3 and GF4, the air injected by the primary and the secondary airflows were 2085 lpm and 2016 lpm, respectively, which resulted in oxygen starvation and fuel rich conditions at the Port 2 zone.
Case GF1 Case GF2 Case GF3 Case GF4
4
3
2
1
0
Port 1
b
Port 2
Port 3
1000
All CO concentrations at Port 1 exceeds the measurement limit (20,000ppm)
800
CO concentration (ppm)
2426
600
400
Case GF1 Case GF2 Case GF3 Case GF4
200
0
Port 1
Port 2
Port 3
Fig. 8. O2 and CO concentrations for Case GF1–GF4: (a) O2 concentrations and (b) CO concentrations.
Area-averaged NOx concentrations at the three ports for all cases are shown in Fig. 9. As the tertiary air increased (from GF1 to GF4), the NOx concentration at Port 3 decreased. These results show that the NOx emission was controlled by multi-staged combustion, and the effect of tertiary air was significant. As shown in Table 3, the overall air–fuel ratio was the same for all of the cases, and the oxygen concentration at the furnace exit was about 4% volumetric. Hence, the difference of NOx concentration between the cases was not only the result of dilution with air but also the
NOx concentration (ppm)
100
80
Case GF1 Case GF2 Case GF3 Case GF4
60
40
20
0
Port 1
Fig. 7. DeNOx mechanism by tertiary air injection.
Port 2
Port 3
Fig. 9. NOx concentrations for Case GF1–GF4.
D. Shin et al. / Waste Management 28 (2008) 2422–2427
NOx reduction mechanism of air distributed combustion. The tertiary air rates for GF3 and GF4 were 12.5% and 15% of total gas, respectively, so that NOx dilution effects are expected to approximate these amounts. However, the NOx reduction rates of GF3 and GF4 were 40% and 38%, respectively, suggesting that the NOx reduction was the result of not only the dilution of tertiary air but also of the combustion air distribution effect. 5. Conclusion A 30 kg/h pyrolysis-melting incineration system was designed, and the combustion characteristics of the simulated gas fuel in the combustion chamber of the incinerator were investigated as well as NOx generation/reduction characteristics. Propane was employed as a pyrolyzed gas. CO and NOx emissions significantly varied by combustion air distributions for staged combustion. The temperatures at Ports 1 and 2 were almost uniform along the measurement line, implying that the swirling flow from the secondary air injection enhanced flow uniformity. Meanwhile, the concentrations of CO at Port 1 exceeded the measurement limit (20,000 ppm) due to incomplete combustion. However, at Port 2, they were drastically reduced due to active combustion by additional air injection. At Port 3 where tertiary air was supplied, CO concentrations were under 100 ppm through completed combustion. As the tertiary air rate increased, the NOx concentration decreased, while the CO concentration at Port 3 increased slightly in case GF4. Acknowledgement This research was supported by the Strategic R&D Cluster Development Program of Seoul City 2006 and
2427
the Faculty Research Program 2007 of Kookmin University in Korea. References Cho, G.C., Choi, J.W., Park, H.K., Yoo, B.T., 2000. The evaluation of NOx emission factor from large combustion facilities in Seoul. Korean Journal of Sanitation 15, 78–83. Coghe, A., Solero, G., Scribano, G., 2004. Recirculation phenomena in a natural gas swirl combustor. Experimental Thermal and Fluid Science 28, 709–714. Fink, J.K., 1999. Pyrolysis and combustion of polymer wastes in combination with metallurgical processes and the cement industry. Journal of Analytical and Applied Pyrolysis 51, 239–252. Han, J.H., Jeong, K.K., Choi, J.H., Choi, S., 1996. A hot flow model analysis of MSW incinerator. International Journal of Energy Research 21, 899–910. Nasserzadeh, V., Swithenbank, J., 1991. Design optimization of a large municipal solid waste incinerator. Waste Management 11, 249–261. Nevers, N.D., 1995. Air pollution control engineering. McGraw-Hill International Ed., 209. Park, S., 2006. Pyrolysis Characteristics of RDF in Pyrolysis Chamber of a Pyrolysis Melting Incinerator, M.Sc., Thesis, Yonsei University, Seoul, Korea. Reed, R.J., 1978. North American Combustion Handbook, second ed. North American Mfg. Co., Cleveland, USA (pp. 16–17). Rogaume, T., Auzanneau, M., Jabouille, F., Goudeau, J.C., Torero, J.L., 2002. The effects of different airflows on the formation of pollutants during waste incineration. Fuel 81, 2277–2288. Ryu, C., Shin, D., Choi, S., 2002. Combined bed combustion and gas flow simulation for a grate type incinerator. Journal of Air and Waste Management Association 52, 189–197. Shin, D., Jang, S.D., Hwang, J., 2005. Combustion characteristics of paper mill sludge in a combustor with internally cycloned circulating fluidized bed. Waste Management 25, 680–685. Turns, S.R., 1996. An Introduction to Combustion, first ed. McGraw-Hill, 499. Yang, Y., Reuter, M.A., Hartman, D.T.M., 2003. CFD Modelling for control of hazardous waste incinerator. Control Engineering Practice 11, 93–101.