ARTICLE IN PRESS
Journal of Electrostatics 63 (2005) 1105–1114 www.elsevier.com/locate/elstat
Streamer corona plasma for fuel gas cleaning: comparison of energization techniques S.A. Naira,, K. Yana, A. Safitrib, A.J.M Pemena, E.J.M. van Heescha, K.J. Ptasinskib, A.A.H. Drinkenburgb a
Faculty of Electrical Engineering, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands b Faculty of Chemical Engineering, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands Received 17 August 2004; accepted 28 February 2005 Available online 6 April 2005
Abstract Tar (heavy hydrocarbon or Poly Aromatic Hydrocarbon (PAH)) removal from biomass derived fuel gas is one of the biggest obstacles in its utilization for power generation. We have investigated pulsed corona as a method for tar removal. Our previous experimental results indicated an energy consumption of 200–250 J/L at an optimum gas temperature of 400 1C. One of the challenges in scaling up a process based on pulsed corona is the availability of high power pulsed power sources. To overcome this, a newer method of streamer corona generation by an alternative ‘‘DC/AC’’ power source is investigated. Energy consumption for tar removal process is used as a basis for comparing the two methods for its chemical efficiency. Results indicate the pulsed corona system and DC/AC system to have similar efficiency. r 2005 Elsevier B.V. All rights reserved. Keywords: Non thermal plasma; Pulsed corona; Tar; Biomass gasification; Electrostatic precipitators (ESP)
Corresponding author. Tel.: +31 40 247 4494; fax: +31 40 245 0735.
E-mail address:
[email protected] (S.A. Nair). 0304-3886/$ - see front matter r 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.elstat.2005.02.004
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 63 (2005) 1105–1114
1106
1. Introduction Clean up of biomass gasification gas remains as one of the challenges for the commercial utilization of this process. Fig. 1 gives a brief overview of the gas clean up problem. In previous works, we have investigated pulsed corona plasma as a method for tar abatement [1–5]. A distinct advantage of a plasma based gas cleaning system is the ability for simultaneous dust and tar removal. This concept has been demonstrated and a chemical mechanism for the process of tar removal by pulsed corona plasma has been outlined [4]. An optimum temperature for the process was determined to be 400 1C with an energy consumption of 200–250 J/L [6]. Nevertheless, the availability of large-scale pulse power sources still remains as one of the critical issues for adaptation to industrial situations. An important feature of pulsed corona plasma is the generation of electrons, ions and radicals. The use of ions to charge particles has led to worldwide industrial application of electrostatic precipitation. The use of radicals to convert one kind of compound to another one is the subject of many investigations. This technique is foreseen as the next generation gas cleaning system, where simultaneous reaction and precipitation can be achieved. The main technical difficulty for applying such pulsed corona plasma however, arises from simultaneous requirements on power rating, energy conversion efficiency, lifetime and cost. Industrial systems also bring up the issue of matching between the power source and reactor. As an alternative, a cost effective method of streamer corona generation using a combined DC/AC power source has been proposed and investigated [7,8]. A biomass gasification gas cleaning system based on the combined DC/AC system aims to overcome the technical Biomass gasification
Utilization
Exit of gasifier T ~ 800 oC CO2 (10 ~12 %), CO (18 ~20%), H2 (17 ~20%), CH4 (1~4 %), N2 (rest), Tars (Poly aromatic HC ‘s), Particulates, Alkali S, NH3
Power generation (Gas engine, Gas turbine) Biomass gasification gas clean-up
Fischer – Tropsch
Fuel Cell
• Particulates • Tar removal • S,NH3 removal (depending on biomass quality)
Fig. 1. Overview of biomass gasification system.
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 63 (2005) 1105–1114
1107
difficulties of a pulsed corona system, such as, availability of high power pulsed power sources. An important issue is, however, its chemical efficiency. The newly developed ‘‘DC/AC’’ power source (a DC voltage modulated with a high-frequency AC voltage) provide an ‘‘inexpensive’’ way of generating streamer corona or non thermal plasma [9]. The flexibility offered by this technique provide one of the easiest ways to upgrade existing ESP’s (ElectroStatic Precipitator). An efficiency or energy consumption similar to that of pulsed corona for use in a tar removal process by a DC/AC power source will result in a significant reduction in investment costs. From plasma generation point of view, both methods can be distinguished as [7] (a) Simultaneous streamer generation—for the case of pulsed power source In a pulsed corona system, CCD (Charged-Coupled Device) pictures with high spatial and time resolution have shown that on application of pulsed high voltages, many parallel streamers propagate almost simultaneously from the high-voltage wire to the ground electrode. The time delay between individual streamer formations (for the same pulse) is close to the time required for the voltage pulse to propagate along the high-voltage wire. (b) Random streamer generation—for the case of DC/AC system The DC voltage is at a level higher than the corona inception with a modified electrode configuration to facilitate streamer generation. The AC peak to peak voltage is usually higher than 1 kV. Two individual streamers occur randomly. The goal of the present investigation is to verify the effect of the two plasma generation concepts on the tar removal process. This is done by comparing the efficiency for naphthalene removal process in various gas compositions.
2. Power source The principle component of a pulsed corona system is the pulsed power source. One such source is described in the earlier investigation [10,11]. A pulsed power plasma system is rather insensitive to electrode mis-arrangements, although for a uniform streamer generation it is best to have the electrode centred along the axis of the reactor. For a DC/AC energized system, the electrode design as well as the spacing between the high voltage and ground electrode is quite critical. For plasma generation, the electrodes should be designed to have a low corona inception voltage [12,13]. As an example, two such geometries are shown in Fig. 2. Fig. 2(a) shows an electrode configuration for a wire-plate type geometry, and Fig. 2(b) for a wirecylinder type configuration. An additional problem is that corona is very sensitive to the gap distance, which can lead to spark breakdown and thereby loss of energy. A DC/AC corona plasma system is usually operated at a voltage level above the streamer inception voltage and below the breakdown voltage. The difference between the streamer inception and the breakdown voltage decreases at increasing
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 63 (2005) 1105–1114
1108
Fig. 2. (a) Electrode configuration for a wire-plate type reactor. (b) Electrode configuration for a wirecylinder type configuration.
Low voltage part
High voltage part
R 6 &C 6
R 1 &C 1
R4
L1 L2
L3 Filter
C0 Th3
L4
D1
Th1
R5
CL R3 & C3
C’h
D2 Th2 TR R 2 &C 2 &S
Fig. 3. Schematic diagram of the power source used for DC/AC corona.
temperature [14]. Hence, the modulation frequency has to be adjusted for each operating condition (temperature) for optimum performance. Based on this principle, pilot scale units have been developed and, as for the case of pulsed corona, the concept is being considered for a variety of applications [7]. It has a potential to replace pulsed-power sources for generating streamer corona. Nevertheless, for a 150 mm wire cylinder type reactor operating at a high temperature of T4400 1C, pulsed corona is an easy technique for non-thermal plasma generation. 2.1. Experimental set-up of the DC/AC system 2.1.1. Power Source Fig. 3 shows a schematic of the power source. The pulse transformer TR separates the low- and high-voltage parts of the circuit. The low-voltage part consists of a main
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 63 (2005) 1105–1114
1109
filter, a set of rectifiers, three air-core inductors L1 ; L2 and L3 , three thyristors Th1, Th2 and Th3, two energy storage capacitors C0 and C L , and the primary windings of the pulse transformer TR. The thyristors are switched consecutively in order to charge the high-voltage capacitor, C h (additional capacitance C 0h þ C Reactor ). Three RC snubbers and a silicon surge voltage suppressor S are used to avoid over-voltages on these thyristors. The high-voltage part consists of the secondary windings of the pulse transformer TR, two high-voltage diodes D1 and D2 , two damping resistor R4 , and R5 , an air-core inductor L4 , the capacitors, C 0h . In the first step, the low-voltage capacitor C L is resonantly charged via the energy storage capacitor C 0 , the thyristor Th1 and the inductor L1 , where C0bCL. During the second step, the high-voltage capacitor is resonantly charged via C L , L2 , TR, Th2, D1 , and L4 . When the charging voltage is higher than the inception voltage, streamer formation is initiated. The high-voltage diode D1 blocks the discharge to the transformer, thus maintaining the required DC level. The successive charging of the capacitor contributes to the AC component. Thus, a DC/AC corona plasma system can be realized. Fig. 4 indicates a typical voltage waveform at room temperature. The energization can be divided into two periods, viz. first one to charge the reactor to a peak voltage V p , and the second the voltage drop due to streamer formation. For the first stage, the reactor is resonantly charged and both capacitive and corona current co-exists. The reactor is energized by the power source. For the second period, the flow of energy from the power source to the reactor stops and the voltage drops due to corona discharge. The corona current can be calculated as I ¼ C h
dV . dt
(1)
20 18 16
Voltage (kV)
14 12
20
10
6
18 16 14 12
4
10 -2.72
8
Vp VDC -2.52
-2.32
2 0 -5
-3
-1
1
3
Time (ms) Fig. 4. Typical voltage waveform at room temperature.
5
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 63 (2005) 1105–1114
1110
The current can be further approximated as I ¼ kV ðV V c Þ. Thus, we have dV . dt If at time t ¼ 0, V ¼ V P , the voltage decay can be expressed as V ðtÞ V c V p V c kV c t ¼ exp , Ch V ðtÞ Vp kV ðV V c Þ ¼ C h
(2)
(3)
where V ðtÞ is the voltage on the reactor in the second stage, V c the corona inception voltage, V p the peak voltage, k is a coefficient (primarily dependent on the gas composition and ion mobility). In order to efficiently transfer the energy from the capacitor C L to the capacitor C h for plasma generation, the following conditions are adopted: 2 1 2C L V o
¼ 12C h ðV 2P V 2DC Þ
(4)
and C L V o ¼ NC h ðV P V DC Þ.
(5)
Substituting Eq. (5) in Eq. (4) NV o ¼ V P þ V DC .
(6)
Substituting Eq. (6) in (5), the following relation can be obtained for good matching: 2V P 2 CL ¼ N Ch 1 (7) NV o or in terms of V DC , 2V DC 2 CL ¼ N Ch 1 , NV o
(8)
where C L is the low voltage capacitance, C h the high-voltage capacitance (C 0h +capacitance of reactor, CR), N the transformer ratio (in the present case, N ¼ 60), V o the primary voltage, output from low voltage part, or input to the transformer, V DC the base DC voltage, V p the peak voltage. The reactor capacitance (CReactor) alone is quite small (typically 30–50 pf); hence to deliver higher energy into the reactor, an additional capacitance, C 0h (450 pf) is added in parallel to the reactor. However, there is a limitation on increasing the value of C L , due to breakdown within the reactor. The values used for the experiments are given in Table 1. 2.1.2. Reactor and the electrode arrangement The reactor used for the experiments is the same as mentioned in the earlier investigation [10], i.e. 0.16 m diameter and 1.3 m in length. The electrode configuration is as shown in Fig. 2(b). Due to breakdown limitations, the operating temperature is kept lower than 300 1C. The objective of the investigations as
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 63 (2005) 1105–1114
1111
Table 1 Parameters of the power source used for the experiments Temperature (1C)
C L (mF)
Peak voltage (V p ) (kV)
Energy per pulse (J/pulse)
200 300
1.5 1.0
16.1 14.1
0.114–0.115 0.088–0.090
mentioned earlier, is to compare the efficiency of removal of naphthalene by (i) pulsed corona (ii) DC/AC corona system. 2.1.3. Measurements The voltage measurements were done by means of a high-voltage probe (Tektronix HV probe) connected at the high-voltage electrode of the reactor, and the current measurements were done by means of a Pearson current transformer at the grounding path of the transformer. The product of the voltage and current waveforms are then integrated to obtain the energy per pulse. Chemical measurements, for naphthalene, are done in the same way as mentioned earlier, by FTIR (Fourier Transform Infrared Spectroscopy).
3. Experimental results and discussions The earlier investigations related to the chemical mechanisms have indicated that the primary process for tar removal is by CO2 decomposition [2,3]. Thus, to compare the performance and the effects of the plasma generating schemes, the energy consumption for naphthalene removal in a N2+CO2 gas mixture is studied. The results can, therefore, be translated into the case for the synthetic fuel gas by the kinetic model [4,5]. Fig. 5 shows the energy density requirements for naphthalene removal under dry reforming conditions at temperatures of 200 1C and 300 1C, respectively, for both the pulsed corona as well as the DC/AC system. At temperatures of 200 1C, the DC/AC system shows lesser energy requirements as compared to the pulsed corona system. However, at 300 1C, the energy density becomes almost identical. The exact reasons cannot be stated due to lack of fundamental studies related to the onset streamer production by DC/AC systems. Nevertheless, based on the information available so far, some explanation can be provided. The visual views of streamers produced in such a DC/AC system shows a more diffused structure emanating from the sharp points on the electrode, and such that to the naked eye the discharge looks uniform (Fig. 6). On the other hand, as is generally known, pulsed corona streamer appears to have a more filamentary nature [12,13]. In case of pulsed corona, diffusion of species out of the streamer channel plays a significant role in avoiding non-linear terminations [4]. The advantages of a more
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 63 (2005) 1105–1114
1112
100
Fraction Remaining (%)
90
200 K - Pulsed Corona
80
300 K - Pulsed Corona
70
200 K - DC/AC 300 K - DC/AC
60 50 40 30 20 10 0
0
50
100
150
200
250
300
Energy Density (J/L) Fig. 5. Comparison of the energy consumption for pulsed and DC/AC corona—removal of naphthalene (initial conc. 3–4 g/Nm3) from N2+CO2 (10%) at various temperatures.
Fig. 6. Stationary photo of DC/AC streamers produced in a wire plate configuration at room temperatures in air.
diffuse discharge were already seen during the experiments at increased temperatures where a decrease in the energy consumption was observed. Apart from this result, the energy per streamer in a pulsed corona case is higher (at least 10 times) than for the case of the onset streamers generated by DC/AC power source. Thus, the initial radical density is higher for the case of pulsed corona. A consequence of all of the above factors may be the apparent increase in radical yields with higher temperatures.
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 63 (2005) 1105–1114
1113
For the case of onset streamers, due to their diffused nature (small energy transfer per streamer and a self sustained high frequency), the radicals seem to be distributed more homogenously as compared to the case of pulsed corona. In other words, the radical mixing conditions appear to be better. Hence the kinetics of the process governs the removal process and the effects of diffusion become negligible. The process is dominated by linear termination mechanisms. At 200 1C, pulsed corona is severely hampered by radical-radical interactions, which decreases with increase in temperature. These effects are less severe for the DC/AC system at these temperatures (T ¼ 200 1C), hence, higher efficiency levels for removal can be seen. However, with an increase of temperature (T4200 1C), the non-linear effects tend to decrease in the case of pulsed corona, which contributes to higher radical utilization efficiencies. These factors being not significant in the case of the DC/AC system, hence the efficiency levels are similar. Thus the DC/AC system and the pulsed corona system, in fact have the same efficiency if the operating regime (linear or non-linear terminations) remains the same. For much higher temperatures, the process is completely dominated by linear terminations. Hence, we expect the energy consumption to be the same for both DC/AC and pulsed corona. In addition, it can also be seen that for the case of DC/AC, complete removal requires higher levels of energy than for the case of pulsed corona. The energy consumptions are similar for about 90–95% removal. Thus from an overall process point of view, a combination of two systems would be ideal.
4. Conclusion The aim of the investigation was to compare the chemical efficiency of the two energization methods: pulsed corona and DC/AC streamer corona generation. Although a difference in energy consumption is seen, both systems have the same chemical efficiency, if the operating regime (linear or non-linear terminations) remains the same, especially at higher temperatures (T43001C).
Acknowledgements The author acknowledges the financial support of SDE (Dutch Foundation for Sustainable Energy), the Dutch Energy Research Center, ECN and the Center for Sustainable Technology (TDO), Eindhoven University of Technology.
References [1] S.A. Nair, A.J.M. Pemen, K. Yan, F.M. Van Gompel, H.E.M. van Leuken, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, Tar removal from biomass derived fuel gas by pulsed corona discharges, Fuel Proc. Technol. 84 (1–3) (2003) 161–173.
ARTICLE IN PRESS 1114
S.A. Nair et al. / Journal of Electrostatics 63 (2005) 1105–1114
[2] S.A. Nair, A.J.M. Pemen, K. Yan, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, Chemical processes in tar removal from biomass derived fuel gas by pulsed corona discharges, Plasma Chem. Plasma Proc. 23 (4) (2003) 665–680. [3] A.J.M. Pemen, S.A. Nair, K. Yan, E.J.M van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, Pulsed corona discharges for tar removal from biomass derived fuel gas, Plasma Polym. 18 (3) 209–224. [4] S.A. Nair, A.J.M. Pemen, K. Yan, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, Tar removal from biomass derived fuel gas by pulsed corona discharges — a chemical kinetic study, Ind. Eng. Chem. Res. 43 (7) (2004) 1649–1658. [5] S.A. Nair, Corona plasma for tar removal, Ph.D. Thesis, Eindhoven University of Technology, 2004. [6] S.A. Nair, A.J.M. Pemen, K. Yan, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, Tar removal from biomass derived fuel gas by pulsed corona discharges — a chemical kinetic study II, Ind. Eng. Chem. Res., in press. [7] K. Yan, G.J.J. Winands, S.A. Nair, E.J.M. van Heesch, A.J.M. Pemen, From electrostatic precipitation to corona plasma system for exhaust gas cleaning, Proceedings of ICESP (International Conference on Electrostatic Precipitation) IX, May 2004, South Africa. [8] K. Yan, D. Higashi, S. Kanazawa, T. Ohkubo, Y. Nomoto, J.S. Chang, NOx removal from air streams by a superimposed AC/DC energized flow stabilized streamer corona, Trans. IEE Japan 118-A (1998) 948–953. [9] K. Yan, Inrichting voor het genereren van corona-ontladingen, NL 1024408, 30th September, 2003. Patent pending. [10] S.A. Nair, K. Yan, A.J.M. Pemen, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, A high temperature pulsed corona plasma system for tar removal from biomass derived fuel gas, J. Electrostat. 61 (2) (2004) 17–127. [11] A.J.M. Pemen, S.A. Nair, K. Yan, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, High temperature pulsed corona processing of fuel gas, J. Adv. Oxid. Tech. 7 (2) (2004) 123–127. [12] P.P.M. Blom, High power pulsed corona, Ph.D. Thesis, Eindhoven University of Technology, The Netherlands, 1997. [13] K. Yan, Corona plasma generation, Ph.D. Thesis, Eindhoven University of Technology, The Netherlands, 2001. [14] E.J.M. van Heesch, A.J.M. Pemen, K. Yan, S.V.B. van Paasen, K.J. Ptasinski, A.H.J. Huijbrects, Pulsed corona tar cracker, IEEE Trans. Plasma Sci. 28 (5) (2000) 1571–1575.
