- Email: [email protected]
Plasma-aided solid fuel combustion
Proceedings of the
Proceedings of the Combustion Institute 31 (2007) 3353–3360
Combustion Institute www.elsevier.com/locate/proci
Plasma-aided solid fuel combustion E.I. Karpenko a, V.E. Messerle b, A.B. Ustimenko
b,*
a
b
Branch Centre of Plasma-Power Technologies of Russian J.S.Co. ‘‘United Power System of Russia’’, 33 Pushkin str., Gusinoozersk 671160, Russia Combustion Problems Institute, al-Farabi Kazakh National University, 172 Bogenbai batyra str., Almaty 050012, Kazakhstan
Abstract Plasma supported solid fuel combustion is promising technology for use in thermal power plants (TPP). The realisation of this technology comprises two main steps. The first is the execution of a numerical simulation and the second involves full-scale trials of plasma supported coal combustion through plasma-fuel systems (PFS) mounted on a TPP boiler. For both the numerical simulation and the full-scale trials, the boiler of 200 MW power of Gusinoozersk TPP (Russia) was selected. The optimization of the combustion of low-rank coals using plasma technology is described, together with the potential of this technology for the general optimization of the coal burning process. Numerical simulation and full-scale trials have enabled technological recommendations for improvement of existing conventional TPP to be made. PFS have been tested for boilers plasma start-up and flame stabilization in different countries at 27 power boilers steam productivity of 75–670 tons per hour (TPH) equipped with different type of pulverised coal burners. At PFS testing power coals of all ranks (brown, bituminous, anthracite and their mixtures) were used. Volatile content of them varied from 4 to 50%, ash—from 15 to 48% and calorific values—from 6700 to 25,100 kJ/kg. In summary, it is concluded that the developed and industrially tested PFS improve coal combustion efficiency and decrease harmful emission from pulverised coal-fired TPP. 2006 The Combustion Institute. Published by Elsevier Inc. All rights reserved. Keywords: Coal; Combustion; Thermochemical preparation; Plasma-fuel system
1. Introduction Plasma assisted coal combustion is a relatively unexplored area in coal combustion science and only a few references are available on this subject [1]. Coal-fired utility boilers face two problems, the first being the necessity to use expensive oil for start-up and the second being the increased commercial pressure requiring operators to burn a broader range of coals, possibly outside the *
Corresponding author. Fax: +73272 675141. E-mail address: [email protected] (A.B. Ustimenko).
specifications envisaged by the manufacturer’s assurances for the combustion equipment. Each of these problems results in a negative environmental impact. Oil firing for start-up increases the gaseous and particulate burden of the plant. The firing of poorer quality coals has two disadvantages: reduced flame stability performance necessitating oil support and its consequential emissions and cost implications; and reduced combustion efficiency due to a increased amounts of carbon in the residual ash, resulting in an increase of emissions per MW of power generated. Plasma-aided coal combustion represents a new effective and ecological friendly technology, which
1540-7489/$ - see front matter 2006 The Combustion Institute. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.proci.2006.07.038
3354
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
is equally applicable to alternative ‘green’ solid fuels. One of the prospective technologies is thermochemical plasma preparation of coals for burning (TCPPCB). This technology addresses the above problems in TPP. The realisation of the TCPPCB technology comprises two main steps. The first includes numerical simulations and the second involves full-scale trials of plasma supported coal combustion in a TPP boiler. For both the numerical study and full-scale trials, the boiler of 200 MW power of Gusinoozersk TPP (Russia) was selected. In the framework of this concept some portion of pulverised solid fuel (pf) is separated from the main pf flow and undergone the activation by arc plasma in a special chamber— PFS (Fig. 1). The air plasma flame is a source of heat and additional oxidation, it provides a high-temperature medium enriched with radicals, where the fuel mixture is heated, volatile components of coal arc extracted, and carbon is partially gasified. This active blended fuel can ignite the main pf flow supplied into the furnace. This technology provides boiler start-up and stabilization of pf flame and eliminates the necessity for additional highly reacting fuel. To demonstrate advantages of plasma-aided coal combustion technology the numerical simulations were performed using the Cinar ICE ‘CFD’ code [2]. Cinar ICE has been designed to provide computational solutions to industrial problems related to combustion and fluid mechanics. The Cinar code solves equations for mass, momentum and energy conservation. Physical models are employed for devolatilisation, volatiles combustion (fast un-premixed combustion), the combustion of char and the turbulence (k e). Comparison of the calculations with data generally reveals excellent agreement. The maximal discrepancies between measured and calculated furnace temperatures do not exceed 15%. Numerical simulation and industrial trials have enabled technological recommendations for improvement of existing conventional TPP to be made. PFS have been tested for boilers plasma start-up and flame stabilization in different countries at 27 power boilers steam productivity of 75–670 TPH equipped with different type of pulverised coal burners [3]. At PFS testing power coals of all ranks (brown, bituminous, anthracite and their mixtures) were used. Volatile content
Fig. 1. Sketch of PFS.
of them varied from 4 to 50%, ash—from 15 to 48% and calorific values—from 6700 to 25,100 kJ/kg. In summary, it is concluded that the developed and industrially tested PFS improve coal combustion efficiency and decrease harmful emission from pulverised coal-fired TPP. 2. Numerical simulation The PFS is a cylinder with the plasma generator placed on the burner (Fig. 2). In PFS, since the primary mixture is deficient in oxygen, the carbon is oxidised mainly to carbon monoxide. As a result, at the exit from the PFS a highly reactive mixture is formed of combustible gases and partially burned char particles, together with products of combustion, while the temperature of the gaseous mixture is around 1300 K. Further mixing with the secondary air, upon the introduction of the mixture into the furnace, promotes intensive ignition and complete combustion of the prepared fuel. The numerical experiments were performed for a cylindrical direct flow burner—PFS shown in Fig. 2. The efficiency of plasmatron is around 85%. All parameters of the PFS are presented in Table 1. ‘Tugnuiski’ bituminous coal (TBC) was used for the experiments. Its proximate and ultimate analyses and particle size distribution are presented in Table 2. From the available experimental data of the PFS operation [4], the measured composition of the gas phase at the exit of the PFS was (volume %): CO = 28.5; H2 = 8.0; CH4 = 1.5; CO2 = 2.0; N2 = 59.5; O2 = 0.0; others = 0.5, including NOX = 50 mg/nm3. The flame from the plasmatron feeds the PFS 0.35 m in axial direction downstream of the PFS inlet plane (Fig. 3). To enable the modelling exercise the plasma flame is assumed as a heat/mass source defined with the exit temperature of 2800 K and the mass flow of 54 kg/h. The thermal and chemical equilibrium approach was selected for the calculations of the PFS. For it, TERRA code was used [6]. This
Fig. 2. Schematic view of PFS.
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360 Table 1 Specification of PFS operating parameters PFS Length, m Inner diameter, m Thermal efficiency, %
2.35 0.25 90
Plasmatron Electric power, kW Plasma gas Mass flow, kg/h Inlet air temperature, K Outlet air temperature, K Inner diameter, m Outlet velocity, m/s
100 Air 54 298 2800 0.04 118.2
Primary air Air flow, kg/h Velocity, m/s Temperature, K Coal dust concentration, kg/kg
3500 20.0 80 0.50
Table 2 Specification of TBC Mass % Particle size distributiona Proximate analysis Moisture 14.00 Volat. matter 36.27 Fixed carbon 44.33 Ash
19.40
Ultimate analysis Carbon
61.7
Hydrogen Nitrogen Sulphur Oxygen
4.10 1.20 0.39 13.20
a
160 lm—10% 130 lm—10% 74 lm—20% 50 lm—40% 24 lm—20% Lower calorific value: 23,000 kJ/kg Coal feed rate: 1750 kg/h
Assumed particles size distribution.
3355
The numerical results for the radial temperature profiles at the PFS exit are presented in Fig. 4, while Fig. 3 shows the predicted temperature contours along the PFS axis. The numerical results are validated with the measured data only at the exit of the PFS. The radial temperature profile is shown for an axial location of 2.0 m from plasmatron axis (x = 2.35 m). The predicted profile is revealed to be axis-symmetric in accordance with the experimental profile. Although the measured profile shows a distinctive temperature local minimum at the chamber central line, which indicates the plasma flame penetration effect, in the case of predicted temperature profile this minimum is insignificant. It could be the reason of under-predicted penetration of the plasma jet into the co-flowing stream of air-coal mixture. In the real situation, it may be expected that the plasma jet will separate the air-coal mixture flow into two streams, leaving the central part of the flow with lower fuel concentration. The high-energy concentrated plasma jet, with high initial momentum, may act as a solid body [5] penetrating through the cross flow, while the coal particles trajectories are divided into two streams, showing two temperature maxima on both sides of the centre line. The mean calculated species concentrations of the gas phase composition at the exit of the PFS was (volume %): CO = 26.1; H2 = 12.1; CH4 = 0.1; CO2 = 0.8; N2 = 60.9; O2 = 0.0. The values of the temperature level and species concentrations were used as input values to numerically simulate the boiler working in the plasma regime. The PFS was incorporated in the furnace of a full-scale boiler with a steam productivity of 640 TPH (Gusinoozersk TPP, Russia). The schematic view of the boiler equipped with PFS and its primary dimensions are shown in Figs. 5 and 6. The furnace is characterised by two symmetrical combustion chambers, each having four tangentially directed main double burners in two layers. Each
Fig. 3. Predicted temperature contours along the PFS axial direction.
method included the formation of the libraries containing the values of species concentrations as a function of the mixture fraction and temperature level. Although the combustion is not a thermal equilibrium process, the application of this approach could be justified by the existence of charged species and radicals, which are highly active and probably act as a catalyst increasing the rate of chemical reactions. In addition to this, the high-energy input and maximum temperature level make the chemical reactions fast so that they are probably close to the equilibrium condition.
Fig. 4. Predicted temperature radial profiles at the exit of PFS. Line is calculation, d is experiment.
3356
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
of the main burners is divided in two sections. They are the section of air–fuel mixture supply and the section of secondary air supply. Combustion chambers are joined by a central section. The cooling chamber is above of the combustion chambers and then turning chamber follows. Fuel consumption on the boiler was 121.3 TPH and total amount of air in the boiler was 553,128 normal cubic meter per hour at temperature of the secondary air 350 C. Coefficient of air surplus on the fire-chamber’s outlet was 1.2. Initial data for the main burners are the same as for PFS shown in Table 1. Four PFS are mounted instead of four lower sections of the main double burners as it is shown in Fig. 5. During the period of boiler warm-up and flame stabilisation, the plasmatrons are operating. When the boiler performance is stabilised, the plasmatrons are switched off and PFS work
Fig. 5. BKZ 640-140 boiler furnace equipped with four PFS (top view).
Fig. 6. Scheme of the industrial furnace of BKZ 640-140 boiler.
as conventional pf burners. In the case of flame instabilities, the plasmatrons are easily switched on. The grid for the mathematical simulation is defined by 118 · 52 · 68 grid lines in three directions (x, y and z). The main results from numerical experiments for the furnace, velocity vectors, temperature profiles and oxygen concentrations, are presented in Figs. 7–9. The velocity vectors fields within the single semi furnace can be found in Fig. 7. The specific alignment of the burners in each corner of the combustion chamber generates the tangential flow of the fluid and particles, increasing the intensity of combustion. Comparing the fields for conventional (Fig. 7a) and plasma operated regime (Fig. 7b) we can see that in the plasma operated regime coal particles tend to move from the centre of the chamber towards the conventional burners. Note, the velocity vector fields presented in such way for the conventional and plasma operational regime can be compared just qualitatively as different scaling factors for vector quantities were applied. In Figs. 8 and 9, panel 1 presents the predicted values for the centre line along the furnace height, while panel 2 gives the values at the exit, along the furnace width. The numerical results represent the boiler performance for the standard operational regime and for operation in plasma regime. Values of boiler performance in standard regime at the exit from the furnace were validated. The measured averaged temperature at the exit is around 1400 K, and this value agrees with the numerical results, while the averaged concentration of measured oxygen is around 4%. In the plasma regime (Fig. 9) the temperature levels along the furnace height are lower, while rapid combustion of pf within the combustion chamber results from the introduction of the mixture of hot combustible gases and unburned char from the PFS. Comparison of the experimental data with the predictions in plasma operational regime (Fig. 9) shows satisfactory agreement. The following data were measured when the boiler power was 120 MW and the excess air factor was 1.24: concentration of oxygen (O2) in exhaust gas was 6.1%, NOX was 700 mg/nm3 (1431.5 ppm), unburned carbon was 0.8%, temperature in the body of flame was 1270 C and temperature in the furnace outlet was 1050 C. The concentration of carbon dioxide (CO2) in exhaust gas calculated through O2 concentration was 14%. As we can see the difference in temperature of the combustion products inside the furnace is not more than 17% and at the furnace outlet it is about 6%. The difference in concentrations of oxygen in the exhaust gas is about 30%. A possible explanation of the discrepancy is the 60% boiler load factor during the period of the measurements. There is no need to operate in plasma-aided regime at
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
3357
Fig. 7. Velocity vectors field within the combustion chamber at the burners’ level for the conventional (a) and plasma operational regime (b).
Fig. 8. Predicted temperature and oxygen level contours and profiles for the conventional operational regime (centre line, exit of the furnace).
Fig. 9. Predicted temperature and oxygen level contours and profiles for the plasma operational regime (centre line, exit of the furnace).
the 100% boiler load factor. But when the load factor decreases we need plasma for pf flame stabilization. 3. Industrial tests The complex of the experiment and computer simulation of pf plasma-aided combustion showed
realisability and efficiency of PFS for their use in coal-fired boilers. On the base of the fulfilled researches PFS were incorporated into the furnaces of industrial coal-fired TPP. The PFS have been tested successfully at 27 pulverised coal boilers at 16 TPP (Russia, Kazakhstan, Korea, Ukraine, Slovakia, Mongolia and China) with the steam productivity from 75–670 tons per hour since 1994 [3]. The boilers were fitted with different
3358
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
systems of pf preparation (such as the direct pf injection and systems with pf hopper). Total 70 PFS were mounted and tested on the boilers. Figure 5 illustrated a scheme of arrangement of four PFS on the power block 200 MW at Gusinoozersk TPP with the use of pulverised coal of high concentration (top view). Figure 10 illustrates a scheme of arrangement of PFS on the boiler BKZ-420 in Ulan-Bator TPP-4 (top view). Twelve burners are placed in three layers in the corners of the boiler. Two PFS were mounted on the low layer opposite each other. In 2–3 s from the PFS start the temperature of both pulverised coal flames achieved 1100– 1150 C. In an hour the flames temperature which were 7–8 m in length achieved 1260–1290 C. In accordance with operating instruction the boiler start-up duration was 4 h. All eight boilers of the TPP were equipped with PFS for fuel oil free boiler start-up. Figures 11 and 12 present results of experiments on NOX reduction and decrease of unburned carbon at plasma ignition of coals on the outlet of a furnace [7]. When a plasmatron operates in a regime of plasma stabilization of a flame, NOX concentration is reduced twice and at the same time amount of unburned carbon is reduced four times. NOX is reduced owing to two-stage pulverised coal combustion. The first one is PFS where ignition and partial gasification of coal in primary air is realised. The second stage is a furnace of a boiler where combustion of the products after the first stage in the secondary air is taken place. Fuel nitrogen is released with coal volatile inside PFS. It forms molecular nitrogen due to deficit of oxygen in air–fuel mixture plasma treated in PFS. In the second stage mainly thermal nitrogen oxides are formed. As known [8] fuel nitrogen is a main source of nitrogen oxide emission from furnace. And thermal NOX can reach 10–15% of all NOX emission in a furnace at temperature more than 1700 C. As regards unburned
Fig. 10. BKZ-420 boiler furnace equipped with two PFS (top view).
Fig. 11. Specific power consumption influence onto reduction of nitrogen oxides concentration at plasmaaided pulverised coal combustion.
Fig. 12. Specific power consumption influence onto reduction of unburned carbon at plasma-aided pulverised coal combustion.
carbon (Fig. 12), its decrease can be explained by the fuel reactivity increase due to formation of two-component high-reactive fuel from pulverised coal and heat explosion of coal particles at their interaction with arc plasma. Figures 13 and 14 represent a generalization of the data gotten as a result of full-scale boilers plasma-aided start-up and pf flame stabilization. We can see (Fig. 13) the specific power consumption for PFS decrease with coal volatile content increase, i.e., when coal reactivity rises. Note power consumption for plasmatron is within 2.5% from heat power of the pf burner, even for the anthracite coal with minimal reactivity. For example auxiliary power requirements of a boiler are 8–10%. Thus, if we know coal volatile yield using the curve from the figure we will determine electric power of plasmatron needed for PFS of specified coal consumption and known coal calorific value. Figure 14 shows that PFS relative heat power decreases from 25 to 10% with coal volatile (fuel
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
3359
three times higher ones for PFS [9]. Thus if we know coal volatile yield using the curve from the figure we will determine total heat power of PFS needed for boiler start-up of specified nominal coal consumption for the boiler. Then knowing coal consumption of the boiler burners we can find a number of PFS for this boiler plasma start-up. For instance, in accordance with Figs. 13 and 14, to start-up the power boiler of 200 MW using bituminous coal of 23,000 kJ/kg caloricity four PFS with 100 kW plasmatrons are needed (see Fig. 5). 4. Conclusions Fig. 13. Generalized experimental dependence of specific electric power consumptions on a plasmatron (e) versus coal volatiles content (Vdaf) gotten during PFS tests at sixteen different TPP. e = P/(Q Æ G), here P is a plasmatron electric power; Q is coal calorific value; G is the coal consumption through a PFS. Line is polynomial fit.
Fig. 14. Generalized experimental dependence of PFS specific heat power (R) versus coal volatiles content (Vdaf) gotten during PFS tests at 14 different TPP (1994–2004). R = (Gstart/Gnominal) Æ 100%, here Gstart and Gnominal is consumption of coal at plasma boiler’s start-up and nominal consumption of the coal for the boiler respectively. Line is polynomial fit.
reactivity) increase. Note, relative heat power of oil system for boilers start-up and pf flame stabilization is some 30% in existed TPP. In other words, even at incineration of a low-rank coal of 4–5% volatile content, relative heat power of the PFS is 5% less than one at conventional incineration of the coal, and in case of midband reactivity coal (Vdaf = 15–30%) and high reactivity one (Vdaf = 30–45%) relative heat power of the PFS is only 7–10%. It means that energy efficiency of PFS is 3–4 times higher of conventional oil system used in TPP for boilers start-up and pulverised coal flame stabilization. Note that oil system (oil nozzles and metal consuming, environmentally unfavourable oil equipment) operating costs are
• Developed, investigated and industrially tested plasma-fuel systems improve coal combustion efficiency, while decreasing harmful emission from pulverised-coal-fired thermal power plants. • PFS eliminate the need for expense gas or oil fuels on start-up. • PFS improve coal ignition and burnout without the need for such remedies as increasing the mill temperature, augmenting the excess air factor, or finer grinding. • The application of numerical modelling approaches of pulverised coal preparation for combustion within the plasma burner has shed new light on the possible chemical and physical mechanisms of the coal–plasma interaction. • Although the combustion process of pulverised coal may not be in thermal equilibrium, the present thermal equilibrium calculations resulted in predictions close to the experimental data. • Simulation of an industrial boiler in conventional and plasma operational modes reveal that during the operation of PFS results in stable ignition and intensive burning of the pf at reduced temperature, conditions which reduce the amount of nitrogen oxide formation. • Prior to the wider implementation of PFS, additional data relating to further coal types and their blends are ideally required.
Acknowledgments The authors gratefully acknowledge the European Commission for this work funding through the ISTC project and personally Professor F. Lockwood for his support and coordination of the research project. References [1] M. Sugimoto, K. Maruta, K. Takeda, O.P. Solonenko, M. Sakashita, M. Nakamura, Thin Solid Films 407 (2002) 186–191.
3360
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
[2] F.C. Lockwood, T. Mahmud, M.A. Yehia, Fuel 77 (12) (1998) 1329. [3] Z. Jankoski, E.I. Karpenko, F.C. Lockwood, V.E. Messerle, A.B. Ustimenko, in: Proceeding of the 8th International Conference on Energy for a Clean Environment ‘‘Clean Air’’, Lisbon, Portugal, 2005, CD (18.2), Book of Abstracts 69. [4] Z. Jankoski, F.C. Lockwood, V.E. Messerle, E.I. Karpenko, A.B. Ustimenko, Thermophysics and Aeromechanics 11 (3) (2004) 461–474. [5] F.J. Weinberg, in: F.J. Weinberg (Ed.), Advanced Combustion Methods, Academic Press, London, 1986.
[6] M. Gorokhovski, E.I. Karpenko, F.C. Lockwood, V.E. Messerle, B.G. Trusov, A.B. Ustimenko, Journal of the Energy Institute 78 (4) (2005) 157– 171. [7] V.E. Messerle, V.S. Peregudov, in: O.P. Solonenko, M.F. Zhukov (Eds.), Investigation and Design of Thermal Plasma Technology, vol. 2, Cambridge International Science Publishing, London, 1995. [8] D.H. Tike, S.M. Slater, A.F. Sarofim, J.C. Williams, Fuel 53 (1974) 120–125. [9] M.G. Drouet, Revue Generale d’Electricite (1) (1986) 51–56.
Proceedings of the Combustion Institute 31 (2007) 3353–3360
Combustion Institute www.elsevier.com/locate/proci
Plasma-aided solid fuel combustion E.I. Karpenko a, V.E. Messerle b, A.B. Ustimenko
b,*
a
b
Branch Centre of Plasma-Power Technologies of Russian J.S.Co. ‘‘United Power System of Russia’’, 33 Pushkin str., Gusinoozersk 671160, Russia Combustion Problems Institute, al-Farabi Kazakh National University, 172 Bogenbai batyra str., Almaty 050012, Kazakhstan
Abstract Plasma supported solid fuel combustion is promising technology for use in thermal power plants (TPP). The realisation of this technology comprises two main steps. The first is the execution of a numerical simulation and the second involves full-scale trials of plasma supported coal combustion through plasma-fuel systems (PFS) mounted on a TPP boiler. For both the numerical simulation and the full-scale trials, the boiler of 200 MW power of Gusinoozersk TPP (Russia) was selected. The optimization of the combustion of low-rank coals using plasma technology is described, together with the potential of this technology for the general optimization of the coal burning process. Numerical simulation and full-scale trials have enabled technological recommendations for improvement of existing conventional TPP to be made. PFS have been tested for boilers plasma start-up and flame stabilization in different countries at 27 power boilers steam productivity of 75–670 tons per hour (TPH) equipped with different type of pulverised coal burners. At PFS testing power coals of all ranks (brown, bituminous, anthracite and their mixtures) were used. Volatile content of them varied from 4 to 50%, ash—from 15 to 48% and calorific values—from 6700 to 25,100 kJ/kg. In summary, it is concluded that the developed and industrially tested PFS improve coal combustion efficiency and decrease harmful emission from pulverised coal-fired TPP. 2006 The Combustion Institute. Published by Elsevier Inc. All rights reserved. Keywords: Coal; Combustion; Thermochemical preparation; Plasma-fuel system
1. Introduction Plasma assisted coal combustion is a relatively unexplored area in coal combustion science and only a few references are available on this subject [1]. Coal-fired utility boilers face two problems, the first being the necessity to use expensive oil for start-up and the second being the increased commercial pressure requiring operators to burn a broader range of coals, possibly outside the *
Corresponding author. Fax: +73272 675141. E-mail address: [email protected] (A.B. Ustimenko).
specifications envisaged by the manufacturer’s assurances for the combustion equipment. Each of these problems results in a negative environmental impact. Oil firing for start-up increases the gaseous and particulate burden of the plant. The firing of poorer quality coals has two disadvantages: reduced flame stability performance necessitating oil support and its consequential emissions and cost implications; and reduced combustion efficiency due to a increased amounts of carbon in the residual ash, resulting in an increase of emissions per MW of power generated. Plasma-aided coal combustion represents a new effective and ecological friendly technology, which
1540-7489/$ - see front matter 2006 The Combustion Institute. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.proci.2006.07.038
3354
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
is equally applicable to alternative ‘green’ solid fuels. One of the prospective technologies is thermochemical plasma preparation of coals for burning (TCPPCB). This technology addresses the above problems in TPP. The realisation of the TCPPCB technology comprises two main steps. The first includes numerical simulations and the second involves full-scale trials of plasma supported coal combustion in a TPP boiler. For both the numerical study and full-scale trials, the boiler of 200 MW power of Gusinoozersk TPP (Russia) was selected. In the framework of this concept some portion of pulverised solid fuel (pf) is separated from the main pf flow and undergone the activation by arc plasma in a special chamber— PFS (Fig. 1). The air plasma flame is a source of heat and additional oxidation, it provides a high-temperature medium enriched with radicals, where the fuel mixture is heated, volatile components of coal arc extracted, and carbon is partially gasified. This active blended fuel can ignite the main pf flow supplied into the furnace. This technology provides boiler start-up and stabilization of pf flame and eliminates the necessity for additional highly reacting fuel. To demonstrate advantages of plasma-aided coal combustion technology the numerical simulations were performed using the Cinar ICE ‘CFD’ code [2]. Cinar ICE has been designed to provide computational solutions to industrial problems related to combustion and fluid mechanics. The Cinar code solves equations for mass, momentum and energy conservation. Physical models are employed for devolatilisation, volatiles combustion (fast un-premixed combustion), the combustion of char and the turbulence (k e). Comparison of the calculations with data generally reveals excellent agreement. The maximal discrepancies between measured and calculated furnace temperatures do not exceed 15%. Numerical simulation and industrial trials have enabled technological recommendations for improvement of existing conventional TPP to be made. PFS have been tested for boilers plasma start-up and flame stabilization in different countries at 27 power boilers steam productivity of 75–670 TPH equipped with different type of pulverised coal burners [3]. At PFS testing power coals of all ranks (brown, bituminous, anthracite and their mixtures) were used. Volatile content
Fig. 1. Sketch of PFS.
of them varied from 4 to 50%, ash—from 15 to 48% and calorific values—from 6700 to 25,100 kJ/kg. In summary, it is concluded that the developed and industrially tested PFS improve coal combustion efficiency and decrease harmful emission from pulverised coal-fired TPP. 2. Numerical simulation The PFS is a cylinder with the plasma generator placed on the burner (Fig. 2). In PFS, since the primary mixture is deficient in oxygen, the carbon is oxidised mainly to carbon monoxide. As a result, at the exit from the PFS a highly reactive mixture is formed of combustible gases and partially burned char particles, together with products of combustion, while the temperature of the gaseous mixture is around 1300 K. Further mixing with the secondary air, upon the introduction of the mixture into the furnace, promotes intensive ignition and complete combustion of the prepared fuel. The numerical experiments were performed for a cylindrical direct flow burner—PFS shown in Fig. 2. The efficiency of plasmatron is around 85%. All parameters of the PFS are presented in Table 1. ‘Tugnuiski’ bituminous coal (TBC) was used for the experiments. Its proximate and ultimate analyses and particle size distribution are presented in Table 2. From the available experimental data of the PFS operation [4], the measured composition of the gas phase at the exit of the PFS was (volume %): CO = 28.5; H2 = 8.0; CH4 = 1.5; CO2 = 2.0; N2 = 59.5; O2 = 0.0; others = 0.5, including NOX = 50 mg/nm3. The flame from the plasmatron feeds the PFS 0.35 m in axial direction downstream of the PFS inlet plane (Fig. 3). To enable the modelling exercise the plasma flame is assumed as a heat/mass source defined with the exit temperature of 2800 K and the mass flow of 54 kg/h. The thermal and chemical equilibrium approach was selected for the calculations of the PFS. For it, TERRA code was used [6]. This
Fig. 2. Schematic view of PFS.
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360 Table 1 Specification of PFS operating parameters PFS Length, m Inner diameter, m Thermal efficiency, %
2.35 0.25 90
Plasmatron Electric power, kW Plasma gas Mass flow, kg/h Inlet air temperature, K Outlet air temperature, K Inner diameter, m Outlet velocity, m/s
100 Air 54 298 2800 0.04 118.2
Primary air Air flow, kg/h Velocity, m/s Temperature, K Coal dust concentration, kg/kg
3500 20.0 80 0.50
Table 2 Specification of TBC Mass % Particle size distributiona Proximate analysis Moisture 14.00 Volat. matter 36.27 Fixed carbon 44.33 Ash
19.40
Ultimate analysis Carbon
61.7
Hydrogen Nitrogen Sulphur Oxygen
4.10 1.20 0.39 13.20
a
160 lm—10% 130 lm—10% 74 lm—20% 50 lm—40% 24 lm—20% Lower calorific value: 23,000 kJ/kg Coal feed rate: 1750 kg/h
Assumed particles size distribution.
3355
The numerical results for the radial temperature profiles at the PFS exit are presented in Fig. 4, while Fig. 3 shows the predicted temperature contours along the PFS axis. The numerical results are validated with the measured data only at the exit of the PFS. The radial temperature profile is shown for an axial location of 2.0 m from plasmatron axis (x = 2.35 m). The predicted profile is revealed to be axis-symmetric in accordance with the experimental profile. Although the measured profile shows a distinctive temperature local minimum at the chamber central line, which indicates the plasma flame penetration effect, in the case of predicted temperature profile this minimum is insignificant. It could be the reason of under-predicted penetration of the plasma jet into the co-flowing stream of air-coal mixture. In the real situation, it may be expected that the plasma jet will separate the air-coal mixture flow into two streams, leaving the central part of the flow with lower fuel concentration. The high-energy concentrated plasma jet, with high initial momentum, may act as a solid body [5] penetrating through the cross flow, while the coal particles trajectories are divided into two streams, showing two temperature maxima on both sides of the centre line. The mean calculated species concentrations of the gas phase composition at the exit of the PFS was (volume %): CO = 26.1; H2 = 12.1; CH4 = 0.1; CO2 = 0.8; N2 = 60.9; O2 = 0.0. The values of the temperature level and species concentrations were used as input values to numerically simulate the boiler working in the plasma regime. The PFS was incorporated in the furnace of a full-scale boiler with a steam productivity of 640 TPH (Gusinoozersk TPP, Russia). The schematic view of the boiler equipped with PFS and its primary dimensions are shown in Figs. 5 and 6. The furnace is characterised by two symmetrical combustion chambers, each having four tangentially directed main double burners in two layers. Each
Fig. 3. Predicted temperature contours along the PFS axial direction.
method included the formation of the libraries containing the values of species concentrations as a function of the mixture fraction and temperature level. Although the combustion is not a thermal equilibrium process, the application of this approach could be justified by the existence of charged species and radicals, which are highly active and probably act as a catalyst increasing the rate of chemical reactions. In addition to this, the high-energy input and maximum temperature level make the chemical reactions fast so that they are probably close to the equilibrium condition.
Fig. 4. Predicted temperature radial profiles at the exit of PFS. Line is calculation, d is experiment.
3356
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
of the main burners is divided in two sections. They are the section of air–fuel mixture supply and the section of secondary air supply. Combustion chambers are joined by a central section. The cooling chamber is above of the combustion chambers and then turning chamber follows. Fuel consumption on the boiler was 121.3 TPH and total amount of air in the boiler was 553,128 normal cubic meter per hour at temperature of the secondary air 350 C. Coefficient of air surplus on the fire-chamber’s outlet was 1.2. Initial data for the main burners are the same as for PFS shown in Table 1. Four PFS are mounted instead of four lower sections of the main double burners as it is shown in Fig. 5. During the period of boiler warm-up and flame stabilisation, the plasmatrons are operating. When the boiler performance is stabilised, the plasmatrons are switched off and PFS work
Fig. 5. BKZ 640-140 boiler furnace equipped with four PFS (top view).
Fig. 6. Scheme of the industrial furnace of BKZ 640-140 boiler.
as conventional pf burners. In the case of flame instabilities, the plasmatrons are easily switched on. The grid for the mathematical simulation is defined by 118 · 52 · 68 grid lines in three directions (x, y and z). The main results from numerical experiments for the furnace, velocity vectors, temperature profiles and oxygen concentrations, are presented in Figs. 7–9. The velocity vectors fields within the single semi furnace can be found in Fig. 7. The specific alignment of the burners in each corner of the combustion chamber generates the tangential flow of the fluid and particles, increasing the intensity of combustion. Comparing the fields for conventional (Fig. 7a) and plasma operated regime (Fig. 7b) we can see that in the plasma operated regime coal particles tend to move from the centre of the chamber towards the conventional burners. Note, the velocity vector fields presented in such way for the conventional and plasma operational regime can be compared just qualitatively as different scaling factors for vector quantities were applied. In Figs. 8 and 9, panel 1 presents the predicted values for the centre line along the furnace height, while panel 2 gives the values at the exit, along the furnace width. The numerical results represent the boiler performance for the standard operational regime and for operation in plasma regime. Values of boiler performance in standard regime at the exit from the furnace were validated. The measured averaged temperature at the exit is around 1400 K, and this value agrees with the numerical results, while the averaged concentration of measured oxygen is around 4%. In the plasma regime (Fig. 9) the temperature levels along the furnace height are lower, while rapid combustion of pf within the combustion chamber results from the introduction of the mixture of hot combustible gases and unburned char from the PFS. Comparison of the experimental data with the predictions in plasma operational regime (Fig. 9) shows satisfactory agreement. The following data were measured when the boiler power was 120 MW and the excess air factor was 1.24: concentration of oxygen (O2) in exhaust gas was 6.1%, NOX was 700 mg/nm3 (1431.5 ppm), unburned carbon was 0.8%, temperature in the body of flame was 1270 C and temperature in the furnace outlet was 1050 C. The concentration of carbon dioxide (CO2) in exhaust gas calculated through O2 concentration was 14%. As we can see the difference in temperature of the combustion products inside the furnace is not more than 17% and at the furnace outlet it is about 6%. The difference in concentrations of oxygen in the exhaust gas is about 30%. A possible explanation of the discrepancy is the 60% boiler load factor during the period of the measurements. There is no need to operate in plasma-aided regime at
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
3357
Fig. 7. Velocity vectors field within the combustion chamber at the burners’ level for the conventional (a) and plasma operational regime (b).
Fig. 8. Predicted temperature and oxygen level contours and profiles for the conventional operational regime (centre line, exit of the furnace).
Fig. 9. Predicted temperature and oxygen level contours and profiles for the plasma operational regime (centre line, exit of the furnace).
the 100% boiler load factor. But when the load factor decreases we need plasma for pf flame stabilization. 3. Industrial tests The complex of the experiment and computer simulation of pf plasma-aided combustion showed
realisability and efficiency of PFS for their use in coal-fired boilers. On the base of the fulfilled researches PFS were incorporated into the furnaces of industrial coal-fired TPP. The PFS have been tested successfully at 27 pulverised coal boilers at 16 TPP (Russia, Kazakhstan, Korea, Ukraine, Slovakia, Mongolia and China) with the steam productivity from 75–670 tons per hour since 1994 [3]. The boilers were fitted with different
3358
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
systems of pf preparation (such as the direct pf injection and systems with pf hopper). Total 70 PFS were mounted and tested on the boilers. Figure 5 illustrated a scheme of arrangement of four PFS on the power block 200 MW at Gusinoozersk TPP with the use of pulverised coal of high concentration (top view). Figure 10 illustrates a scheme of arrangement of PFS on the boiler BKZ-420 in Ulan-Bator TPP-4 (top view). Twelve burners are placed in three layers in the corners of the boiler. Two PFS were mounted on the low layer opposite each other. In 2–3 s from the PFS start the temperature of both pulverised coal flames achieved 1100– 1150 C. In an hour the flames temperature which were 7–8 m in length achieved 1260–1290 C. In accordance with operating instruction the boiler start-up duration was 4 h. All eight boilers of the TPP were equipped with PFS for fuel oil free boiler start-up. Figures 11 and 12 present results of experiments on NOX reduction and decrease of unburned carbon at plasma ignition of coals on the outlet of a furnace [7]. When a plasmatron operates in a regime of plasma stabilization of a flame, NOX concentration is reduced twice and at the same time amount of unburned carbon is reduced four times. NOX is reduced owing to two-stage pulverised coal combustion. The first one is PFS where ignition and partial gasification of coal in primary air is realised. The second stage is a furnace of a boiler where combustion of the products after the first stage in the secondary air is taken place. Fuel nitrogen is released with coal volatile inside PFS. It forms molecular nitrogen due to deficit of oxygen in air–fuel mixture plasma treated in PFS. In the second stage mainly thermal nitrogen oxides are formed. As known [8] fuel nitrogen is a main source of nitrogen oxide emission from furnace. And thermal NOX can reach 10–15% of all NOX emission in a furnace at temperature more than 1700 C. As regards unburned
Fig. 10. BKZ-420 boiler furnace equipped with two PFS (top view).
Fig. 11. Specific power consumption influence onto reduction of nitrogen oxides concentration at plasmaaided pulverised coal combustion.
Fig. 12. Specific power consumption influence onto reduction of unburned carbon at plasma-aided pulverised coal combustion.
carbon (Fig. 12), its decrease can be explained by the fuel reactivity increase due to formation of two-component high-reactive fuel from pulverised coal and heat explosion of coal particles at their interaction with arc plasma. Figures 13 and 14 represent a generalization of the data gotten as a result of full-scale boilers plasma-aided start-up and pf flame stabilization. We can see (Fig. 13) the specific power consumption for PFS decrease with coal volatile content increase, i.e., when coal reactivity rises. Note power consumption for plasmatron is within 2.5% from heat power of the pf burner, even for the anthracite coal with minimal reactivity. For example auxiliary power requirements of a boiler are 8–10%. Thus, if we know coal volatile yield using the curve from the figure we will determine electric power of plasmatron needed for PFS of specified coal consumption and known coal calorific value. Figure 14 shows that PFS relative heat power decreases from 25 to 10% with coal volatile (fuel
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
3359
three times higher ones for PFS [9]. Thus if we know coal volatile yield using the curve from the figure we will determine total heat power of PFS needed for boiler start-up of specified nominal coal consumption for the boiler. Then knowing coal consumption of the boiler burners we can find a number of PFS for this boiler plasma start-up. For instance, in accordance with Figs. 13 and 14, to start-up the power boiler of 200 MW using bituminous coal of 23,000 kJ/kg caloricity four PFS with 100 kW plasmatrons are needed (see Fig. 5). 4. Conclusions Fig. 13. Generalized experimental dependence of specific electric power consumptions on a plasmatron (e) versus coal volatiles content (Vdaf) gotten during PFS tests at sixteen different TPP. e = P/(Q Æ G), here P is a plasmatron electric power; Q is coal calorific value; G is the coal consumption through a PFS. Line is polynomial fit.
Fig. 14. Generalized experimental dependence of PFS specific heat power (R) versus coal volatiles content (Vdaf) gotten during PFS tests at 14 different TPP (1994–2004). R = (Gstart/Gnominal) Æ 100%, here Gstart and Gnominal is consumption of coal at plasma boiler’s start-up and nominal consumption of the coal for the boiler respectively. Line is polynomial fit.
reactivity) increase. Note, relative heat power of oil system for boilers start-up and pf flame stabilization is some 30% in existed TPP. In other words, even at incineration of a low-rank coal of 4–5% volatile content, relative heat power of the PFS is 5% less than one at conventional incineration of the coal, and in case of midband reactivity coal (Vdaf = 15–30%) and high reactivity one (Vdaf = 30–45%) relative heat power of the PFS is only 7–10%. It means that energy efficiency of PFS is 3–4 times higher of conventional oil system used in TPP for boilers start-up and pulverised coal flame stabilization. Note that oil system (oil nozzles and metal consuming, environmentally unfavourable oil equipment) operating costs are
• Developed, investigated and industrially tested plasma-fuel systems improve coal combustion efficiency, while decreasing harmful emission from pulverised-coal-fired thermal power plants. • PFS eliminate the need for expense gas or oil fuels on start-up. • PFS improve coal ignition and burnout without the need for such remedies as increasing the mill temperature, augmenting the excess air factor, or finer grinding. • The application of numerical modelling approaches of pulverised coal preparation for combustion within the plasma burner has shed new light on the possible chemical and physical mechanisms of the coal–plasma interaction. • Although the combustion process of pulverised coal may not be in thermal equilibrium, the present thermal equilibrium calculations resulted in predictions close to the experimental data. • Simulation of an industrial boiler in conventional and plasma operational modes reveal that during the operation of PFS results in stable ignition and intensive burning of the pf at reduced temperature, conditions which reduce the amount of nitrogen oxide formation. • Prior to the wider implementation of PFS, additional data relating to further coal types and their blends are ideally required.
Acknowledgments The authors gratefully acknowledge the European Commission for this work funding through the ISTC project and personally Professor F. Lockwood for his support and coordination of the research project. References [1] M. Sugimoto, K. Maruta, K. Takeda, O.P. Solonenko, M. Sakashita, M. Nakamura, Thin Solid Films 407 (2002) 186–191.
3360
E.I. Karpenko et al. / Proceedings of the Combustion Institute 31 (2007) 3353–3360
[2] F.C. Lockwood, T. Mahmud, M.A. Yehia, Fuel 77 (12) (1998) 1329. [3] Z. Jankoski, E.I. Karpenko, F.C. Lockwood, V.E. Messerle, A.B. Ustimenko, in: Proceeding of the 8th International Conference on Energy for a Clean Environment ‘‘Clean Air’’, Lisbon, Portugal, 2005, CD (18.2), Book of Abstracts 69. [4] Z. Jankoski, F.C. Lockwood, V.E. Messerle, E.I. Karpenko, A.B. Ustimenko, Thermophysics and Aeromechanics 11 (3) (2004) 461–474. [5] F.J. Weinberg, in: F.J. Weinberg (Ed.), Advanced Combustion Methods, Academic Press, London, 1986.
[6] M. Gorokhovski, E.I. Karpenko, F.C. Lockwood, V.E. Messerle, B.G. Trusov, A.B. Ustimenko, Journal of the Energy Institute 78 (4) (2005) 157– 171. [7] V.E. Messerle, V.S. Peregudov, in: O.P. Solonenko, M.F. Zhukov (Eds.), Investigation and Design of Thermal Plasma Technology, vol. 2, Cambridge International Science Publishing, London, 1995. [8] D.H. Tike, S.M. Slater, A.F. Sarofim, J.C. Williams, Fuel 53 (1974) 120–125. [9] M.G. Drouet, Revue Generale d’Electricite (1) (1986) 51–56.