ARTICLE IN PRESS
Journal of Electrostatics 61 (2004) 117–127
A high-temperature pulsed corona plasma system for fuel gas cleaning S.A. Naira,*, K. Yana, A.J.M. Pemena, G.J.J. Winandsa, F.M. van Gompela, H.E.M. van Leukena, 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, Netherlands b Faculty of Chemical Engineering, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, Netherlands Received 17 October 2003; accepted 4 February 2004
Abstract Tars in fuel gases create serious obstacles for gasification. This paper describes a pulsed corona plasma system for tar removal from the fuel gas. Experiments were performed under gaseous temperatures of up to 500 C. Effects of the gaseous temperature on the electrical matching and tar removal are reported. r 2004 Elsevier B.V. All rights reserved. Keywords: Pulsed corona; Biomass gasification; Tar removal; Pulsed power
1. Introduction Biomass is an important energy source for the future. One efficient way to utilize biomass is to convert it into a gaseous fuel. The producer gas can be utilized for generation of electricity or chemicals. The main issue for such applications, however, is the gas quality, which is often contaminated with heavy hydrocarbons (tars) and particulates. Conventional cleaning techniques are catalytic and thermal cracking, and scrubbing. Non-thermal plasma where tar removal is realized by plasma-induced chemical reactions has become an alternative [1,2]. Simultaneously, particles are *Corresponding author. Tel.: +31-40-247-3709; fax: +31-40-245-0735. E-mail address:
[email protected] (S.A. Nair). 0304-3886/$ - see front matter r 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.elstat.2004.02.002
ARTICLE IN PRESS 118
S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
collected inside the reactor under an electric force. Our previous feasibility studies were carried out with a corona reactor placed after a gasifier. Experiments have demonstrated that tars and particles can be removed simultaneously. It has also been considered that an ultra-short pulsed corona is one of the most ideal techniques for gaseous fuel cleaning and/or processing. The present paper describes a hightemperature laboratory pulsed corona plasma system for investigating the fuel gas cleaning.
2. Experimental set-up The whole experimental set-up consists of three distinct parts: a wire-cylinder type corona reactor, a pulsed-power source, and chemical and electrical diagnostic systems. 2.1. Corona reactor The corona reactor has a wire-cylinder configuration with a high-voltage electrode of 3 mm in diameter and a ground electrode of 160 mm in diameter, and 1.3 m in length. The wall thickness of the cylinder is 12 mm. Gases are circulated in and out of the reactor in a recycle loop driven by a high-temperature fan at variable speeds. It can be used up to 800 C and in the pressure range of 1–1000 mbar. The wall thickness of the loop is 8 mm. The layout of the system is shown in Fig. 1. The entire set-up is made of stainless steel (ASTM A-312). The internal gas flow velocity can be adjusted between 6 and 12 m/s, corresponding to a gas flow rate from 450 to 900 Nm3/h. The total circulation volume is 50 l. The entire set-up is heated electrically. 2.2. Pulsed-power source A schematic diagram of the main electric circuit of the pulsed-power source with a resistive load is indicated in Fig. 2. The pulse transformer TR separates low- and high-voltage parts of the circuit. The low-voltage part consists of a main 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 CL, and the primary windings of the pulse transformer TR. The thyristors are switched consecutively in order to charge a 10 m long pulse forming line (PFL) - RG218). As shown in Fig. 2, the two ends of the PFL are connected in parallel. For increasing the energy per pulse and the output peak power, a high-voltage capacitor can be added in parallel to the PFL. Three RC snubbers and a silicon surge voltage suppressor S are used for avoiding over voltage 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 PFL, a triggered spark-gap switch with an LCR trigger circuit [3], a Transmission line transformer (TLT), and magnetic cores placed around the TLT cables. The TLT consists of two coaxial transmission lines (RG218)
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
119
Fig. 1. (a) A photo of the experimental set-up and (b) schematic diagram of the set-up.
with a length of 2.5 m. At the switch side, both lines are connected in parallel, and at the reactor-side, the two lines are connected in series. Thus, the input and output impedances of the TLT are 25 and 100 O, respectively. The output characteristics are very much dependent on the TLT. A general guideline for design of the TLT cables
ARTICLE IN PRESS 120
S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
Fig. 2. Schematic diagram of the pulsed-power source.
Fig. 3. Schematic diagram and photo of low- and high-temperature feed through and an air heat exchanger.
and magnetic cores in terms of the pulse duration was reported earlier [4]. An energy efficiency of about 97% has been obtained for a TLT with micro-gap Metglas-type magnetic cores. Additional ferrites are placed on the TLT to absorb the remaining energy in the circuit after plasma quenching. The main functions of the TLT can be summarized as: (i) achieving a higher output impedance, (ii) doubling the output voltage, and (iii) protecting the switch against short-circuits and breakdowns. With regard to the temperature difference between the power source and the reactor, a heat exchanger is used in-between. Fig. 3 shows a schematic diagram and a
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
121
set-up photo. One low- and high-temperature feed-through are used to connect to the TLT and the reactor, respectively. The low-temperature feed-through is cooled by a forced downward airflow. In addition to the thermal considerations, the physical size of the heat exchanger is designed to match its characteristic impedance to the output impedance of the TLT. 2.3. Diagnostic system Both electrical and chemical diagnostic systems are installed on-site for real-time measurements. (i) Electrical measurements: Standard D/I (differentiating–integrating) systems are used for measurements of voltage and current pulses [5]. A coaxial capacitive sensor and a one-turn Rogowski coil are installed at the end of the TLT to measure the output voltage and current, respectively. Adequate precautions have been taken to prevent electro-magnetic interference noise. The voltage and current waveforms are recorded with an HP Infinium digital oscilloscope. Both the time resolved discharge power and the energy-per-pulse are calculated with those waveforms. (ii) Chemical measurement: Chemical components are measured with an Fourier transform Infrared (FTIR) spectrometer (Bruker-Vector 22). Two optical windows, as shown in Fig. 1, are installed in the recirculation loop for on-site FTIR measurements.
3. Characteristics of the pulsed-power source A four-step energy conversion process generates a high-voltage pulse. In the first step, the low-voltage capacitor CL is resonantly charged via the energy storage capacitor C0, the thyristor Th1 and the inductor L1, where C0 b CL. During the second step, the PFL is resonantly charged via CL, L2, TR, Th2, D1, and L4. Then the stored energy in the PFL is transferred to the load via the triggered spark-gap switch and the TLT. Before the low-voltage capacitor CL is recharged again, the third thyristor Th3 is used to correct the voltage polarity on CL via Th3, L3, and L1. In order to evaluate the characteristics of voltage rise time, pulse duration and peak voltage, experiments were performed with both resistive and corona reactor loads. (a) Resistive load: The output impedance of the pulsed-power source is represented by the output impedance of the TLT, i.e.100 O. For a matched resistive load and a pure PFL, the output peak voltage can be adjusted to values between 20 and 40 kV. After the spark-gap switch is fired, the PFL discharges into the TLT with an output voltage of half the charging voltage V on the PFL. Considering the 10 m long PFL, one can conclude that the pulse duration t is around 50 ns, and the energy per pulse is around 0.2–0.8 J/pulse. After voltage multiplication by the two-stage TLT, the output voltage becomes equal to the charging voltage V. When an additional capacitance is added in parallel to the PFL, the output peak voltage, current, pulse duration, and the energy per pulse increase. Figs. 4 and 5 show two typical output current waveforms without and with the additional capacitance. When the PFL is
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
122
400
Current(A)
300
200
100
0
-100 -50
0
50 Time (ns)
100
150
Fig. 4. Current waveform with a PFL and a matched resistive load (100 O).
400
Current(A)
300
200
100
0
-100 -50
0
50 Time (ns)
100
150
Fig. 5. Current waveform with a PFL and Ch and a matched resistive load (100 O).
used, the rise time and pulse duration are around 5 and 50 ns, respectively. The output peak current and voltage are about 300 A and 30 kV, respectively. When a 700 pF capacitance is added, the output peak current becomes 360 A. At the same time, the pulse width increases to about 80 ns, and the output waveform tends to change from a square to a double exponential function.
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
123
250
Current (A)
150
40
100
20
50 0 0 -20
-50 -100 -50
Voltage (kV)
60
200
0
50
100
150
200
250
-40 300
Time (ns) Fig. 6. Current and Voltage waveforms with a corona reactor at room temperature.
(b) Corona reactor: The impedance of a corona reactor is determined by two factors: the stray capacitance and the corona power itself. For energy efficient corona plasma energization one needs to optimize the peak voltage, the output impedance, the pulse duration, the rise time and the stray capacitance of the reactor. For the present system as shown in Fig. 2, the total stray capacitance mainly depends on the heat exchanger, the feed-through, and the corona reactor. When the applied voltage is below the corona inception voltage, there is no corona discharge inside the reactor. Voltage and current waveforms appear as a damped LC oscillation. The waveform data indicate that the stray capacitance is 66 pF in the present system. Fig. 6 shows typical voltage and current waveforms for the case of corona in air. Because the applied voltage is quite low, very low intensity corona plasma is generated. The corona reactor almost behaves as a capacitor, and the load to the TLT is similar to an open circuit. The current mainly is a capacitive current. A calculation based on the derived stray capacitance and on the voltage and current waveforms, results in a corona energy of only about 0.2 J/pulse. Most of the stored energy in the PFL is dissipated in the spark-gap switch and in the TLT. According to the previous experiments, a high voltage of around 85–100 kV is needed to efficiently energize the present reactor in air [4]. In the next section, we show that the increase of the gas temperature leads to a reduction of the minimum required peak voltage for realizing good matching.
4. Effect of gas temperature on the matching A good matching between the power source and the reactor can be realized by considering the three-stage energization: before, during and after corona plasma generation. It was concluded that for a given reactor, a minimum peak voltage is
ARTICLE IN PRESS 124
S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
Fig. 7. Typical output voltage and current waveforms with increasing the gaseous temperature.
required in order to make the impedance of the reactor equal to the output impedance of the generator [4]. The increase of gas temperature decreases the gas density. As a result, the required maximum peak voltage is reduced. In terms of the voltage and current waveforms, Fig. 7 shows typical effects of the gas temperature on the energization. The first peak current is mainly dominated by the capacitive current, which remains almost independent of the gas temperature. After the first peak, the corona current becomes larger when increasing the gas temperature. The increase of the gas temperature leads to decreased energy reflection and to improved
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
125
Fig. 8. Power waveforms from the corona reactor at different temperatures.
Fig. 9. Energy transfer to reactor with increase in temperature ðb ¼ ðEnergy transfer to streamer corona ðJ=PulseÞÞ=ðEnergy transfer to matched load ð0:5 J=PulseÞÞÞ:
energy transfer for corona generation. In terms of the output power, Fig. 8 illustrates the effects by comparing the waveforms under room temperature, 200 C, and 500 C. Fig. 9 indicates the effect on the ratio of corona energy transfer, which is defined as the ratio of corona energy to energy output under a matched load. With increase of temperature, the reactor impedance decreases and more energy is dissipated for streamer generation. Similar effects are reported [6] elsewhere, although till 200 C.
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
126
100 200 deg C 300 deg C 80 Fraction remaining(%)
400 deg C 60
40
20
0 0
50
100
150
200
250
300
Corona energy density (J/L) Fig. 10. Naphthalene removal from a gas mixture of N2+CO2 (10%)—Effect of temperature (initial concentration: 3B4 g/Nm3).
5. Effects of the gas temperature on tar removal With regard to the total energy consumption for fuel gas cleaning, both electrical energy conversion efficiency and chemical yield are critical for industrial implementation. A good matching between the reactor and the power source improves the overall electrical efficiency. A high chemical yield increases the chemical efficiency. In terms of the remaining naphthalene and the corona energy density, Fig. 10 shows the effects of the gas temperature on the chemical efficiency in a gas mixture of N2+CO2 (10%). When the temperature becomes higher than 400 C, thermal decomposition of naphthalene becomes very significant with the present set-up. At 400 C, about 50% of the naphthalene is decomposed thermally during a time of 20 min. With corona, 50% removal can be achieved with an energy density of 40 J/L at 400 C in about 3 min. The total experimental time with corona is 10 min, which indicates the decomposition seen in Fig. 10 at 400 C is a result of corona. Considering the proposed main reaction pathways [7–9], one can conclude that the increase of gas temperature may not only improve its related oxidation kinetics, but also improve the initial G-value for O radical generation via CO2 decomposition.
6. Conclusion A high-temperature pulsed corona plasma system not only improves the electrical energy conversion efficiency, but also the G-value for O radical generation.
ARTICLE IN PRESS S.A. Nair et al. / Journal of Electrostatics 61 (2004) 117–127
127
Acknowledgements The authors acknowledge the technical support and advice of Mr. A.W.M. van Iersel, Mr. F.A.J. Mertens, Mr. V. Lekx of Eindhoven University of Technology, H.P.E. Thieman & T.C.J. Geutjes of Vicoma B.V. and P. Leijendeckers of Montair Andersen B.V. in the construction of the set-up. The financial support of Dutch Foundation for Sustainable Energy (SDE), Dutch Energy Research Institute (ECN) and Center for Sustainable Technology (TDO) are acknowledged.
References [1] S.A. Nair, A.J.M. Pemen, K. Yan, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, Fuel Process. Technol. 84 (1–3) (2003) 161–173. [2] B.E.J.M. Van Heesch, G.U.A.J.M. Pemen, K. Yan, S.V.B. Van Paasen, K.J. Ptasinski, P.A.H.J. Huijbrechts, Plasma Sci. IEEE Trans. 28 (5) (2000) 1571–1575. [3] K. Yan, E.J.M. van Heesch, S.A. Nair, A.J.M. Pemen, J. Electrostatics 57 (1) (2003) 29–33. [4] K. Yan, Corona plasma generation, Eindhoven University of Technology, Eindhoven, The Netherlands, 2001. [5] E.J.M. Van Heesch, P.C.T. van der Laan, Power sources for electrical discharges, in: E.M. van Veldhuizen (Ed.), Electrical Discharges for Environmental Purposes, Fundamentals and Applications, Nova Science Publishers, New York, 2000. [6] R. Korzekwa, L. Rosocha, A high temperature pulsed corona reactor, in: Pulsed Power Conference, Tenth IEEE International, Vol. 1, 3–6 July 1995, pp. 138–143. [7] S.A. Nair, A.J.M. Pemen, K. Yan, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, Plasma Chem. Plasma Process. 23 (4) (2003) 665–680. [8] A.J.M. Pemen, S.A. Nair, K. Yan, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, Plasma Polym. 8 (2003) 209–224. [9] S.A. Nair, K. Yan, A.J.M. Pemen, E.J.M. van Heesch, K.J. Ptasinski, A.A.H. Drinkenburg, Ind. Eng. Chem. Res., in press.
