Accepted for 12th IEEE International Pulsed Power Conference, Monterey CA, 1999
EXPERIMENTAL PROGRAM OF THE PULSED CORONA TAR CRACKER E.J.M. van Heesch, A.J.M. Pemen, Keping Yan, S.V.B van Paasen, K.J. Ptasinski, Z. Matyáš, P.A.J.H. Huijbrechts, B.J. Hultermans, A. Nicoletti Eindhoven University of Technology, The Netherlands. P. Zacharias, ISET, Institute for Solar Energy Development, Kassel, Germany Mail: Eindhoven University of Technology, EUT, Div. EVT, P.O.Box 513, 5600MB Eindhoven, The Netherlands. Phone: +31 40 247 3993, Fax: +31 40 245 0735, email:
[email protected] Abstract We are concentrating on the development of pulsed corona discharges to crack heavy tar components (hydrocarbons) into lighter ones. The method has the advantage that it can operate at a high temperature, and be retrofitted to existing installations. The corona discharge is energized by 100 - 150 ns wide voltage pulses (100 kV) at a continuous repetition rate of 600 – 1000 pulses per second. The power dissipated by the corona discharges is 1.5 kW average and 50 MW peak in each pulse. To be cracked by discharges, the hydrocarbons of the tar mixture need to be gaseous and therefore, the corona reactor must operate at a high temperature. In the first phase of the experiments the reactor will run at a modest temperature of 150°C. The reactor is a 1 - 3 m long stainless steel cylinder, 0.25 m diameter with a corona wire along the axis. The pressure will be 1 atmosphere.
I. BIOGAS FROM WOOD GASIFICATION Biomass is a rather simple term for all organic materials that stem from plants, trees and crops. Literature quotes the contribution of biomass to the world’s energy supply as ranging from 10 to 14 %. Biomass orininates from wood farming and from forest and agriculture residues [1]. Several technologies like combustion, pyrolysis, gasification, digestion, and fermentation of biomass are found to make energy from biomass. In this work the issue of tar removal from gasification will be considered. Gasification is a thermochemical conversion of biomass into gaseous fuels by means of partial oxidation of the biomass at high temperatures [1]. Biomass gasification has a significantly higher biomass to power efficiency than biomass combustion. The low caloric gas that is produced in a gasifier with a low heating value (LHV) of typically 4-6 MJ/Nm3, which mainly consists of H2, CO, CH4, CO2, and N2, can be fired directly or be used for firing engines and gas turbine
cycles, and can also serve as syngas in the production of chemicals. Besides the high biomass to power efficiency, biomass gasification has a more compact and less costly gas cleaning equipment than biomass combustion units have, because the producer gas is cleaned before being combusted [1]. In other words, a lower gas flow has to be treated, which results in cheaper and smaller cleaning units. The process of gas cleaning, however, is still one of the major technical challenges with regard to the utilization of the producer gas in gas-fired engines or gas turbines. The technical boundary conditions for the operation of a gasification process are the gas quality requirements for the engine, Table 1 and the producer gas quality resulting from a circulating fluidized bed (CFB) or a downdraft gasifier, Table 2. Table 1. Gas quality requirements for satisfactory gas engine operation [1]. Material gas gas engine turbine Particles mg/Nm3 50 30 P. size 10 5 µm Tars mg/Nm3 100 40 Alkali mg/Nm3 none none Table 2. Producer gas quality from atmospheric, air blown gasifiers [1]. Material Downdraft CFB Moisture wt% 5-20 13-20 Particles g/Nm3 0.1-8 8-100 Tars g/Nm3 0.01-6 2-30 LHVdry MJ/Nm3 4.0-5.6 3.6-5.9 H2 Vol % 15-21 15-22 CO Vol % 10-22 13-15 CO2 Vol % 11-13 13-15 CH4 Vol % 1-5 2-4 CnHm Vol % 0.2-2 0.1-1.2 N2 Vol % rest rest The contaminants: SO2, HCl, NH3, and the particles can efficiently be removed with conventional cleaning
Accepted for 12th IEEE International Pulsed Power Conference, Monterey CA, 1999
quipment like scrubbers, cyclones, and bag filters. Since the developed tar removal systems do not meet the constraints of both low investment costs and high tar removal efficiencies, other tar removal systems must be developed.
Adapting pulsed high voltage to enable fairly homogeneous deposition of electric energy into the gas in time-concentrated form we can promote cracking of various undesired components [3,4,5]. The effects stem from gas discharge products such as electrons, UV
Thermodynamic equilibrium of tar in biogas for initial tar content of 10 g/Nm3 Phenanthrene Pyrene Naphthalene
1.E-03 Volume fraction
Pulse Breakdown
1.E-06 1.E-09 1.E-12 1.E-15 1.E-18
Total (peak) Voltage [kV]
1.E+00
140 120 100 80 60
DC Breakdown
40 Corona Inception
20
1.E-21 0
100
200 300 400 500 Temperature [°C]
600
0
700
Fig. 1:Thermodynamic equilibrium concentrations of tar components in biogas at different temperatures. Although the tar composition in the biogas is strongly dependent on the conditions in the gasifier, a model composition always contains the components naphtalene, pyrene, phenantrene and indene. The thermodynamic equilibrium conditions of the biogas depend on temperature and pressure and total C, O, H, N etc. content. The equilibrium conditions of the tar contaminants are displayed in Fig. 1 as a function of the temperature at atmospheric pressure. In most cases the pyrene concentration in real producer gas below 300°C is lower than the predicted model equilibrium concentration at these temperatures. Obviously the pyrene production rate at low temperature in real biogas is too low to reach equilibrium. However, pyrene production could be enhanced at low temperature by the addition of corona discharges to the biogas. In contrast, corona will enhance the destruction of pyrene at temperatures above 300 °C, because the equilibrium concentration of pyrene is strongly decreasing with increasing biogas temperature. Thus from the thermodynamical point of view the temperature of the corona tar cracker should be higher than 300 °C.
0
200
400
600
800
Fig. 2. Existence region for pulsed corona as a function of temperature (µs pulses, low repetition rate) radiation and radicals. In biogas the mix of radicals will be quite different from the mix in air-like gases. No oxygen radicals will be present. However, due to the moisture content of biogas we still we have hydroxyl radicals. In connection to our 1.5 kW (average), 100 kV pulse source, a corona tar cracker setup has been realized to
Tar Diagnostic
Stack
FTIR
Fan Out Reactor
GC FTIR
Heater
P, T, flow HighVoltage Pulses
Vacuum Pump
Fan
Stack
Cooler Air
II. PULSED CORONA APPLIED TO BIOGAS Biogas
We showed that pulsed corona runs excellently at temperatures of 800°C and higher, even under polluted conditions, see Fig.2 [2]. The diagram shows the applied peak corona voltage as a function of temperature for three cases: 1) corona ignition, 2) breakdown if DC voltage is applied and 3) breakdown if micro-second pulses are applied. If pulsed voltage is applied the existence region is sufficiently wide for high-temperature operation.
1000
Temperature [deg C]
N2
Fig. 3. The flow system of the laboratory corona tar cracker. treat synthetic biogas mixed with tar components. Figure 3 gives an overview. The corona reactor is a 3 m long wire-cylinder geometry with 0.25 m diameter. The gas is treated in a closed loop flow system. The mixture is synthetic. Later this year the reactor will be tested in the
Accepted for 12th IEEE International Pulsed Power Conference, Monterey CA, 1999
field at a real gasifier plant in The Netherlands. A 100 kV pulse source generates the energy to be transferred to the gas [3]. It produces short voltage pulses at a rate of 600 – 1000 pulses per second. Each pulse is 150 ns wide and transfers some 1.5 J to the gas. Well controlled energy transfer from source to corona was found to be possible at elevated temperatures.
First order chemical reactions are found to obey [1,3,5] : X = 1 – exp(-E/∆E),
(2)
where ∆E is an activation level specific for each type of chemical process. From various experimental programs and applications we obtained a list of specific activation energies for corona assisted chemical decomposition processes, see Table 3.
III. PULSED CORONA ENERGY TRANSFER The high voltage is generated by a heavy duty repetitive pulse source. Its final pulse forming network adapts the pulse to the load. The load is the gas discharge reactor. Energy is transferred in three steps: 1. Mains AC power converted into high voltage pulses 2. High-voltage pulses transferred to corona discharges 3. Discharge energy converted into chemical processes The efficiency of each conversion step is an important issue. For step 1, efficiencies of up to 90 % are well achievable. The efficiency of step 2 depends on the electrical matching between reactor and source impedance. The source impedance can be controlled but the impedance of the reactor is a problem. As the reactor is a pulsed corona discharge, the impedance depends on voltage and discharge development. Moreover, the discharge development itself is strongly affected by the impedance of the pulse source. Efficiencies of energy transfer from source to reactor may be as low as 30 % if matching is not well achieved. Currently we are developing more optimized configurations that achieve efficiencies of up to 80 %. Table 3. Specific activation energies for corona decomposition in mixtures with air component concentr. specific [ppm] energy [J/l] toluene 125-450 99 styrene 30-190 11 TCA 80-1000 135 pentane 80-1000 185 methane ≈ 1000
For the production of the high-voltage pulse three subsequent modules are used [3]: resonant charging of a 32 kV capacitor (a 1:60 pulse transformer is part of the charging inductor), a fast spark gap for 32 kV switching and finally a transmission line transformer (TLT) [6] to produce the 100 kV pulses. The transformer has a C-core with a thin lamination of 0.05 mm iron. It is a core-type transformer with two coil sets. The primary of each set has 10 thick flat copper turns. The secondary high voltage coils have 600 turns each and are equipped with field-shaping rings. Each high-voltage coil is epoxy insulated during a vacuum potting process. The two coil sets are connected in parallel. A low inductance spark gap (50 nH, coaxial with the high-voltage capacitor) discharges its coaxially mounted capacitor into the TLT. The reliability of the spark gap is excellent, after 109 pulses (total transferred charge 200 kC) only minor electrode wear is visible. The actual pulse forming line is a TLT (4 cables) with a voltage gain of 3.5.
V. ELECTROMAGNETIC COMPATIBILITY AND MEASURING SYSTEMS
. The efficiency of the final chemical conversion process cannot be expressed in terms of percentages, one has to use e.g. a unit such as removed fraction X of a substance as a function of energy density. This energy density E is the energy yield after step 2 dived by the gas volume to which it is applied. Here we also have to introduce the average power P into the discharge, and we also define the gas flow F passing the reactor. The energy transferred to each discharge is Q and the number of discharges per second is f. The following relations than apply: E = Q.f .t / (Ft) = P/F
IV. COMPONENTS OF THE HEAVY DUTY PULSE SOURCE
(1)
Fast pulsed currents are elegantly measured by Rogowski coils. The base of the corona reactor therefore is constructed as a one-turn Rogowski coil (a toroid of rectangular minor cross section) to measure the total current to the reactor. The high voltage is fed into the reactor via a HV feedthrough, which also contains a capacitive sensor to form a Differentiating-Integrating (DI) measuring system for the external (corona) voltage. In a DI-system it is the strong differentiated signal that is transported via the coaxial cable to the passive RC input section of an integrator at the wall of an EMC cabinet. The integrator restores the original waveform but it also acts as an effective EMC filter. Large common mode currents are allowed to flow on the signal cables from sensor to integrator and back via grounding systems. A sufficiently low transfer impedance of these cables leaves the measured signal undisturbed.
Accepted for 12th IEEE International Pulsed Power Conference, Monterey CA, 1999
Both positive and negative corona show similar behavior in a wire-cylinder geometry [4]. The voltage pulses are applied between the wire and the cylinder. After the start of each pulse, many narrow discharge channels, streamers, grow from the wire towards the cylinder, during the initial streamer phase, as illustrated in Fig.4 [4]. The 40 ns strongly inhomogeneous space charge at the streamer head creates a large E-field, which causes an enhanced ionization and further growth of the initial streamers towards the cylinder. Each 30 ns streamer carries a current of 0.1 - 1 A. The capacitance between the streamer head and the cylinder acts as a limiting impedance for the current through the initial 20 ns streamer. The conductive streamer phase starts after the arrival of the initial streamers at the cylinder. The development of the initial 10 ns streamer into a complete channel between the wire and Fig.4. Axial view along the cylinder results in a wire-cylinder reactor disappearance of the during streamer capacitance between the development. CCD streamer head and the images of 5 ns gating cylinder. This allows a much taken at 10 ns intervals larger current (1-10 A) to after the start of the flow through each streamer; voltage pulse [4]. many streamers in parallel carry the large current of the pulsed corona discharge.
VII. EXPERIMENTAL PROGRAM Within the European Joule program a cooperation has been established with partners from Germany (ISET, Alltech), Italy (JRC Ispra) and The Netherlands (EUT, Convex, Montair, Frigem). The aim is to develop a corona tar cracker for cleaning of gas from wood gasifiers. The initial phase of the program is being carried out in the laboratory using synthetic biogas at temperatures between 150°C and 300°C. During the first test runs we use nitrogen mixed with naphtalene or phenantrene. The analysis is performed with FTIR and gas chromatograph diagnostics. The results are shown in Fig. 5.
It shows the removed fractions of these tar components with and without water vapor added as a function of energy per kg of tar. The results are promising. The specific activation energy ∆E for phenantrene turns out to be as low as 20J/l when water vapor is present. After safety systems for biogas handling are ready we will perform tests on biogas. After optimization of the setup, later this year, it will be transported to a real downdraft gasifier. Results of these tests are expected by the end of 1999. 100 90 80 70 60 50 40 30 20 10 0
Fraction removed (%)
VI. DISCHARGE DEVELOPMENT
N2+naphtalene N2+naphtalene+H2O N2+phenanthrene N2+phenanthrene+H2O
0
5 10 15 Corona energy (kWh/kg in)
20
Fig. 5. Removal of tar components in 150°C atmospheric pressure nitrogen with and without water vapor added.
Acknowledgement These investigations were supported by the programs of the Dutch technology foundation STW and the European Commission.
VIII. REFERENCES [1] S.V.B. van Paasen, “Biogas Treatment with Pulsed Electric Fields”, EUT-SAI report, ISBN 90-5282-9403, Eindhoven, April 1999. [2] E.J.M. van Heesch et al, “Pulsed Corona Existence up to 850°C “, Proceedings of 6th Int. Symp. On HighVoltage Eng., New Orleans, paper 42.23, 4p., 1989 [3] H.W.M. Smulders et al, "Pulsed Power Corona Discharges for Air Pollution Control", IEEE Trans. Plasma Science, Vol. 26, No.5, October 1998. [4] P.P.M. Blom, “High Power Pulsed Corona”, Ph.D. thesis, Eindhoven University of Technology, February 1997. [5] L.A. Rosocha et al, “Treatment of Hazardous Organic Wastes Using Silent Discharge Plasmas”, Nonthermal Plasma Techniques for Pollution Control, Part B, Springer-Verlag, Berlin Heidelberg, 1993, pp. 281308. [6] I.A.D. Lewis and F.H. Wells, “Millimicrosecond Pulse Techniques”, 2nd Ed., Pergamon Press, London, 1959, pp. 109-111.
