Systems and method for ignition and reignition of unstable electrical discharges Number:7,417,385 from the United States Patent and Trademark Office (PTO) owispatent Systems and methods for ignition and reignition of unstable electrical discharges wherein a secondary electrode positioned is between a set of primary electrodes and a high voltage is applied between the secondary electrode and successive ones of the primary electrodes to produce pilot discharges that ionize a gas there between and thereby reduce the voltage necessary to ignite a primary discharge between the primary electrodes. Power is provided to the secondary electrode by a circuit which is independent of the circuit that supplies power to the primary electrodes and generates voltage pulses which are substantially higher than the voltage between the primary electrodes.<H discharges, electrical, Systems, and, met, 7417385.html, Patent, pto, 7417385

Title: Systems and method for ignition and reignition of unstable electrical discharges

Abstract: Systems and methods for ignition and reignition of unstable electrical discharges wherein a secondary electrode positioned is between a set of primary electrodes and a high voltage is applied between the secondary electrode and successive ones of the primary electrodes to produce pilot discharges that ionize a gas there between and thereby reduce the voltage necessary to ignite a primary discharge between the primary electrodes. Power is provided to the secondary electrode by a circuit which is independent of the circuit that supplies power to the primary electrodes and generates voltage pulses which are substantially higher than the voltage between the primary electrodes.

Patent Number: 7,417,385 Issued on 08/26/2008 to Czernichowski, et al.

Inventors: Czernichowski; Albin (Orleans, FR), Hnatiuc; Bogdan (Lasi, RO), Pastva; Peter (Pontoise, FR), Ranaivosoloarimanana; Albert (Nova Coladonia, FR)
Assignee: Ceramatec, Inc. (Salt Lake City, UT)

Appl. No.: 11/186,711

Filed: July 21, 2005

Related U.S. Patent Documents


Application Number	Filing Date	Patent Number	Issue Date
09995125	Nov., 2001	6924608

Foreign Application Priority Data


Nov 27, 2000 [FR]			00 15537

Current U.S. Class: 315/335 ; 219/383

Current International Class: H01J 17/36 (20060101)

Field of Search: 315/291,111.21,111.81,111.71,334,335 219/383

References Cited [Referenced By]

U.S. Patent Documents


3863107	January 1975	Mogensen et al.
4144444	March 1979	Dementiev et al.
4198590	April 1980	Harris
4588850	May 1986	Mueller et al.
4606799	August 1986	Pirklbauer et al.
4661763	April 1987	Ari et al.
4861446	August 1989	Blom et al.
4918590	April 1990	Ohtuka et al.
5043636	August 1991	Klopotek et al.
5711859	January 1998	Caramel et al.
5993761	November 1999	Czernichowski et al.
6007742	December 1999	Czernichowski et al.

Foreign Patent Documents


1059065	Jul., 1979	CA
378296	Jun., 1964	CH
2 049 269	Mar., 1971	FR
2593493	Jul., 1987	FR
2 639 172	May., 1990	FR
2724806	Mar., 1996	FR
2 775 864	Sep., 1999	FR
2172011	Sep., 1986	GB
172152	Jul., 1995	PL
112225	Jun., 1997	RO
WO-9426656	Nov., 1994	WO
WO PCT/GB94/01818	Mar., 1995	WO

Other References

International Search Report re PCT/US01/44307, dated May 17, 2002. cited by other .
Lesueur, et al., "Electrically Assisted Partial Oxidation of Methane", Int. J. Hydrogen Energy, vol. 19, No. 2, (1994), 139-144. cited by other .
Meguernes, et al., "Oxidization of CH4 by H2O in a gliding electric arc", 3rd European Congress on Thermal Plasma Processes, Aachen, Germany, Sep. 19-21, 1994, abstract No. 80; full text in VDI Berichte 1166, (1995), 495-500. cited by other .
Alemu, Office Action for U.S. Appl. No. 09995125 sent May 07, 2003, 1-6. cited by other .
Nave, Office Action for U.S. Appl. No. 09005647 sent Jan. 13, 1999, 1-7. cited by other .
Nave, Office Action for U.S. Appl. No. 09144318 sent Mar. 17, 1999, 1-10. cited by other .
Petit, PCT Written Opinion for PCT/US98/18027 sent Jul. 16, 1999, 1-4. cited by other .
Clement, PCT International Search Report for PCT/US98/18027 sent Jan. 04, 1999, 1-5. cited by other .
Mayekar, K. PCT International Search Report for PCT/US98/00393 sent May 04, 1995, 1-4. cited by other .
Jorgensen, et al., Abstract of FR2593493, (Dec. 10, 2007), 1. cited by other .
Abstract of FR2724806, DialogWeb DERWENT database search results for English-lanuage, printed on Jan. 25, 1999, 1-2. cited by other .
Esp@cenet Automated translation of Description and Claims of CH378296, (Dec. 10, 2007), 1-4. cited by other.

Primary Examiner: Owens; Douglas W.
Assistant Examiner: Alemu; Ephrem

Parent Case Text

CROSS-REFERENCE TO RELATED APPLICATIONS

This is a continuation application of U.S. patent application Ser. No. 09/995,125, filed Nov. 27, 2001 now U.S. Pat. No. 6,924,608, and claims foreign priority benefits under 35 U.S.C. 119(a)-(d) or 365(b) of French Application No. 00.15537 filed on Nov. 27, 2000, entitled, "Systems and Methods for Ignition and Reignition of Unstable Electrical Discharges" which are hereby incorporated by reference as if set forth herein in their entirety.

Claims

What is claimed is:

1. A gas flow chamber device comprising: a simultaneous arrangement of a plurality of primary power electrodes in the form of triads connected in a series to a single multi-phase transformer; a secondary high-voltage electrode positioned in a geometric center between each triad of primary power electrodes wherein the distance between the secondary electrode and each of the triad electrodes is shorter than a distance between the primary electrodes in each of the triads; a first circuit supplying power sequentially to the triads of primary electrodes without causing a discharge between said primary electrodes; and a second circuit supplying high voltage pulses forming an unstable discharge column gliding between the secondary electrode and alternate ones of the triad primary electrodes; the electrodes being enclosed in a gas-filled chamber wherein the gas is provided with a suitable flow rate.

2. The device of claim 1, wherein the first circuit is independent of the second circuit.

3. The device of claim 1, wherein the primary electrodes of each triad are symmetrically arranged about a central axis.

4. The device of claim 1, wherein the time between each in a rapid sequence of pulses is less than the duration of a discharge between a pair of primary electrodes.

5. The device of claim 1, wherein the duration of each pulse is sufficient to produce a single spark per pulse.

6. The device of claim 1, wherein a voltage applied by the second circuit between the secondary electrode and the primary electrodes is at least about 10 times greater than a voltage applied by the first circuit between the primary electrodes.

7. The device of claim 1, wherein the primary low power supply comprises a high frequency of about 150 kHz.

8. The device of claim 1, wherein the discharges form a conductive zone for an ionized gas between electrodes.

9. The device of claim 1, wherein an electrical current flows discharges from the initial triad to a final triad.

10. The device of claim 1, wherein the triads comprise an external resistance component limiting the current so as to provide reignition of the gliding discharges immediately upon extinction.

11. The device of claim 1, wherein the discharges comprise behavior dependant on hydraulic thrust of the gas flow, the rotation of the phases and the oscillation of the voltage imposed by the power supply frequency.

12. The device of claim 1, wherein the unstable fluctuating gliding discharges sustain each other by maintaining a continuous "electrical flame" in each triad.

13. The device of claim 1, wherein the single high-voltage power supply serves simultaneously all gliding discharges within the triad primary electrodes.

14. The device of claim 1, wherein all the high-voltage discharges are established in a series, the current entering at a first pole and exiting at a second pole flowing through a series of discharges which are initialed or reignited at any positions where the electrodes are proximal.

15. The device of claim 1, wherein at least one of the electrodes involved in causing self-ignition and reignition comprises an electrically resistive material, selected from a metal-ceramic composite.

16. A device comprising: a plurality of primary electrodes which are symmetric about a central axis, each electrode having a first end and a second end with the distance between the first ends of any two of said electrodes being less than the distance between the corresponding second ends of said two electrodes; a first circuit for supplying power to the plurality of primary electrodes; a secondary electrode positioned centrally between the plurality of primary electrodes; and a second circuit for supplying power between the secondary electrode and alternate ones of the primary electrodes.

17. A device comprising: a plurality of primary electrodes comprising which are symmetric about a central axis, each primary electrode having a first end and second end; a first circuit for supplying power to the plurality of primary electrodes; a secondary electrode positioned centrally between the plurality of primary electrodes, the secondary electrode having a first end and a second end, and wherein the secondary electrode is in electrical communication with only one end of each primary electrode; and a second circuit for supplying power between the secondary electrode and alternate ones of the primary electrodes.

18. A device comprising: a plurality of primary electrodes symmetric about a central axis, wherein a portion of each primary electrode is in direct electrical communication with a portion of every other electrode; a first circuit for supplying power to the plurality of primary electrodes; a secondary electrode positioned centrally between the plurality of primary electrodes; and a second circuit for supplying power between the secondary electrode and alternate ones of the primary electrodes.

Description

SUMMARY OF THE INVENTION

The present invention relates generally to the ignition and reignition of unstable electrical discharges between electrodes, and more particularly to systems and methods using an intermediate electrode to ignite and reignite discharges between a set of electrodes wherein it is desirable to maintain the discharges with a lower power than is necessary to ignite or reignite the discharges.

The ignition and maintenance of an unstable electrical discharge intended to glide along a pair of electrodes using relatively low power poses an interesting problem. In order to ignite the discharge between the electrodes, a high voltage is required. The voltage must be sufficient to cause breakdown of the impedance between the electrodes so that discharge (arcing) occurs. Once the discharge is established, however, it is desired to have the discharge continue at a relatively low power. This creates a need for complex power supplies to regulate the voltage and/or current between the electrodes.

The present invention provides an alternative to the complex power regulation schemes that have previously been necessary in gliding discharge systems. Rather than focus on the control of the voltage and current of the power supply feeding the discharge, the present invention focuses on reducing the need for such complex power supplies. This is achieved, in very basic terms, by providing an intermediate electrode which lies between a set of primary electrodes. Because the distance between the intermediate electrode and each of the primary electrodes is less than the distance between the primary electrodes themselves, less voltage is required to cause electrical breakdown and ignition of a discharge between the intermediate electrode and the primary electrodes. Once a discharge has been established between the intermediate electrode and each of a pair of primary electrodes, the discharges can effectively be joined to form a discharge between the pair of primary electrodes. Thus, the desired discharge can be achieved without having to deal with the higher threshold voltage that would have been required in the absence of the intermediate electrode.

This is only a brief, generalized description of the invention. The detailed description that follows will more clearly depict a preferred embodiment of the invention, as well as provide a more clear indication of the scope of the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

Other objects and advantages of the invention may become apparent upon reading the following detailed description and upon reference to the accompanying drawings.

FIG. 1 is a circuit diagram illustrating a power supply in accordance with the prior art.

FIG. 2 is a diagram illustrating the variations of voltage, current and instantaneous power in the power supply of FIG. 1.

FIG. 3 is a diagram illustrating the variations in current and voltage under the operating conditions of FIG. 2.

FIG. 4 is a diagram illustrating an alternative power supply in accordance with the prior art.

FIG. 5 is a diagram illustrating an electrode structure in accordance with the prior art.

FIG. 6 is a power supply configured for use with an electrode structure as shown in FIG. 5.

FIG. 7 is a diagram illustrating a power supply which is based on three single-phase transformers.

FIG. 8 is a diagram illustrating an ignition and reignition circuit which is set up independently from a main power circuit that supplies the primary electrodes of the present system.

FIGS. 9a and 9b are diagrams illustrating electrode structures which include a plurality of primary electrodes surrounding a central, intermediate electrode.

FIG. 10 is a diagram illustrating a power supply having a transformer comprising two low-voltage primary windings and one high-voltage secondary winding.

FIG. 11 is a diagram illustrating the electrical phenomena observed in the discharge corresponding to the power supply of FIG. 10.

FIG. 12 is a diagram illustrating a device for the simultaneous supply of four gliding discharges connected to a single high-voltage power supply.

FIG. 13 is a diagram illustrating a device for the simultaneous supply of nine power electrodes connected to a single three-phase transformer.

FIGS. 14a and 14b are diagrams illustrating electrode structures in accordance with one embodiment of the present invention.

FIGS. 15a and 15b are diagrams illustrating alternative electrode structures.

While the invention is subject to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and the accompanying detailed description. It should be understood, however, that the drawings and detailed description are not intended to limit the invention to the particular embodiment which is described. This disclosure is instead intended to cover all modifications, equivalents and alternatives falling within the scope of the present invention as defined by the appended claims.

DETAILED DESCRIPTION

The invention described herein proposes several power generators and electrical circuits to feed highly unstable high-voltage discharges.

One of these discharges, referred to as GlidArc, was previously proposed for multiple industrial applications. Several GlidArc discharges can be interrelated within a single device. Therefore, the invention described herein also proposes generators and circuits to feed certain structures with multiple discharges.

For a plasma-chemical process, such as the destruction of molecules of airborne pollutants or the conversion of a gas containing hydrocarbons, the beneficial action of "cold" electrical discharges has long been demonstrated in the scientific literature. A specific nonequilibrium plasma generator was designed, see BF 88.14932 (2639172), by H. Lesueur, A. Czernichowski, and J. Chapelle, to process significant flows of gas circulating at very high flow rates near a system of stationary electrodes. It was observed that such dual electrode module (since then, referred to as GlidArc-I) was capable of developing an output of up to approximately 5 kW before this simple discharge was transformed into a thermal source considered inappropriate for the processing of gases. Thus, in order to process a significant volume of gas, it was necessary to use a battery consisting of several modules, each of which was fitted with a gas acceleration system located near the electrodes, and with a power supply.

In order to avoid such acceleration of the gas for some applications, a new principle was designed: electric discharges gliding along mobile electrodes, see BF 98.02940 (2775864), by A. Czernichowski and P. Czernichowski. This device, called GlidArc II, contains a minimum of two electrodes, at least one of which must be mobile. As before, the multiple electrode structures were designed for the processing of significant gas flows in multi-stage systems, where each discharge is fed by a specific electrical generator.

The original power supplies of these GlidArc-I or -II discharges are based on a direct or alternating current power supply and current limited by series impedance. This impedance must limit the strong surge during the ignition phase. The absence of such impedance would produce a dead short circuit, along with all of its adverse consequences for the device and the power supply system. Three types of impedance can be considered: resistance, which produces a significant loss of energy if it dissipates outside of the reactor in the form of Joule's heat, which is of little use for the process, capacitance, which is discharged very violently once that the ignition path is established and, therefore, changes the nature of the discharge, which becomes excessively thermalized and thus inappropriate for the "cold" plasma-chemical process, series self-inductance, which transforms a voltage generator into a current generator, which appears to be appropriate.

We originally decided on self-inductance. This simple assembly makes the high ignition voltage that is required for the quasi-cyclical operation of the GlidArc readily available. In fact, the limitation of current by the inductive effect does not appear to create any technological problems. Furthermore, "leak" transformers are commercially available. These transformers (e.g., 15 kV no-load voltage and 15 kVA power) are capable of withstanding dead short circuits, adapting to the variable load, and tolerating a significant surge. However, they show a very poor power factor (sometimes expressed as cos .phi.) of the order of 0.1 to 0.2, which must be offset in order to increase the power factor to about 1, through the use of parallel capacitors. This involves an additional investment, without however solving the problem of the low power transmitted to the GlidArc in relation to the installed power (from 10 to 20%), thus resulting in a high investment cost. We have eliminated or, at the very least, mitigated these defects in the new generators and power supply circuits, which constitutes the subject of this invention.

The detailed description of this GlidArc discharge, which is extremely unstable by design, should make it possible to solve the problems related to its supply and thus understand the characteristics to look for in a higher-performance power supply for industrial scale reactors.

The principle of the GlidArc (both I and II) is based on a quasi-periodic ignition--spreading--extinction sequence of a series of electrical discharges with limited current. We recommend the use of currents lower than 5 Amps in order to remain within the range of a "self-sustained discharge," which has not yet been clearly defined and is still poorly known to science, as it is comprised between "luminescent discharges" and "electric arcs."

At least two electrodes are in contact with the discharge. The legs of the discharge (i.e. galvanic contacts communicating with an electrical power supply of the discharge) glide over these electrodes to prevent their thermal erosion and/or chemical corrosion. The gliding of the legs of the discharge is caused by a quick movement of a flow (gas, vapor, with or without powder or droplets, etc.) across the electrodes (GlidArc-I), or by the mechanical movement of at least one of the electrodes (GlidArc-II). Regardless of the movement's origin, the discharge column spreads fairly quickly and, as the distance between the electrodes is not constant, it increases as the legs move. We also observe that it is somewhat difficult to move the legs of the discharge, and that the column, which is very long compared to the position of the legs, is what causes the legs to jump towards another position, thus shortening the column . . . This increase of the distance between the electrodes, which causes the quasi-progressive spreading of the discharge, is complemented by very quick fluctuations of the column that is moving across a flow that is often turbulent. These fluctuations are similar to the meanders of a river, with the same bends, short circuits, and deviations from the old bed, except that they occur in a very short time.

Moreover, the discharge column may change its diameter following a periodic oscillation of the current feeding the discharge, e.g. an alternating current that goes several times through a value of zero without causing the column to disappear. The column can also change its diameter following electrical current oscillations caused by active components of the power supply circuit . . .

According to the very principle of the GlidArc, we do not attempt to reduce the quasi-progressive change or the "meandering" of the length of the column, nor do we limit the fluctuations of its diameter by eliminating the oscillations of the electrical current . . . On the contrary: we cause and/or maintain all of these column instability phenomena in order to obtain a medium characterized by a significant electrical and dynamic nonequilibrium of a quasi-random flow--as this enables us to obtain a highly thermodynamically nonequilibrium medium that is appropriate for the treatment of the material constituting the flow in close contact with the electrical discharge.

All of these instability features must be accepted, maintained, or even reinforced by an electrical power supply. This should be some type of black box that "sees" the discharge on one side and, on the other side, is connected to an industrial power supply (e.g. 400 V three-phase mains). Such transmission box must be as simple (for reasons of economy, durability, etc.) and as performing (transformation efficiency, filtering of electrical discharges not compatible with the mains, etc.) as possible.

To simplify, let us consider in detail the life cycle of such discharge between two electrodes only (several electrodes in a multiphase structure can also be involved in a more complex discharge). Naturally, the two electrodes (referred to as power electrodes) are set distant from each other, otherwise there would be a dead short circuit. The shortest distance between the electrodes must be at least several millimeters, otherwise it would be very difficult to adjust this distance with accuracy, as the electrodes and their supports are placed inside a reactor, such as a chemical reactor, and therefore are not very accessible. Furthermore, we wish to prevent the slight wear of the electrodes, or the roughness that may develop on their surface, from causing a relatively significant change of said distance in relation to the initial setting. It is at this shortest distance that we observe an electric ignition when the voltage applied to the electrodes exceeds the dielectric breakdown voltage in the flow comprised between the electrodes. Immediately after this breakdown, a small volume of plasma formed between the electrodes is carried by the movement of the gas (GlidArc-I) or by the movement of one electrode in relation to the other (GlidArc-II; in this case, the movement may be helped by the flow of gas or other diluted material). The rate of travel of the discharge depends mainly on the flow rate and/or rate of mechanical displacement of one (or both) electrode(s). The discharge column begins to spread since, according to the very principle of the GlidArc, the distance between the electrodes increases in the direction of flow (e.g. the electrodes are diverging). At the same time, the voltage at the terminals of the electrodes increases, in an attempt to offset the loss of energy through the column that is growing longer. During this phase, the discharge (or rather a quasi-arc) is in a state of near thermodynamic equilibrium, meaning that at each point of the plasma, the temperature of the electrons is close to the temperature of the gas. This state results from the high frequency of collisions between electrons and molecules; the electrical power supplied per unit of length of the discharge is sufficient to offset the radial losses suffered by the column due to thermal conduction. This balancing phase continues while the discharge keeps spreading until the power that can be supplied by the power generator feeding the discharge reaches its maximum value. From that point, while the thermal conduction losses keep increasing, the discharge enters its thermal nonequilibrium phase and a significant drop is observed in the temperature of the gas. However, the temperature of the electrons remains very high. Following the drop in gas temperature, the heat losses decrease, and the length of nonequilibrium plasma can then continue to grow until the heat losses exceed the power available in the discharge. Then, the discharge is extinguished and a new discharge is established at the spot where the two electrodes are closest, and the cycle of ignition, life, and extinction is repeated.

Therefore, in order to operate, the GlidArc reactor needs special power generators. The generator must supply a voltage high enough to ignite the charge and, then, when the voltage of the discharge drops, it must supply a limited power. Thus, its current-voltage characteristic must "drop" quickly after the ignition.

The second phase of the discharge's life, i.e. thermal and electrical nonequilibria during which up to 80% of the power is injected, is especially interesting for the purposes of stimulating a chemical reaction. The active discharges thus created in the GlidArc devices can sweep almost the entire flow. In the GlidArc-I device, the flow of material (e.g. gas) moves across the column at a slightly lower velocity than the flow that is pushing it. In the GlidArc-II, it is no longer necessary to accelerate this flow near the electrodes, as the velocity of travel of the discharge is determined by the movement of one electrode . . . Thus, almost all of the flow is exposed to the electrons, ions, radicals, and particles energized by the discharge. This makes it possible to obtain the desired chemical effect. Following a quick scattering and aerodynamic turbulence, these active species, which have a relatively long life, even manage to spread over the space that is not directly touched by the discharges. These phenomena also contribute to the extraordinary activity of these GlidArc discharges.

The nature of the current and voltage of the GlidArc is such that even their measurement requires special attention. In particular, the significant and quick variations of voltage (10.sup.-10 V/s) and current (10.sup.-8 A/s) during both the ignition and the extinction of each discharge, cause electrical interference. Of course, these same phenomena can be sensed by a non-protected generator feeding such electrical discharge.

It is usually possible to establish mathematical models to describe the physical phenomena and the properties of electrical discharges. These take into account the evolution in time and space of the specific parameters of the plasma, such as diffusion, electrical conductivity, thermal conductivity, viscosity, etc. Thus, there are 3 types of models: microscopic (energy balance of all levels of all components), intermediate (energy balance in the discharge column described by the Elenbaas-Heller equations, which can be simplified by taking into account the radiation and convection phenomena), or yet simplified further by maximum reductions of the energy balance (Cassie, Mayr, or Brown models where the plasma constitutes the variable conductance electrical discharge) . . .

However, in spite of our efforts over long years of research, we were not able to propose an analytical description that would provide an adequate representation of GlidArc type discharges. We have hundreds of records showing the high temporal resolution Volt-Ampere characteristics provided by high-speed digital oscilloscopes connected to computers, the characteristics in different gases flowing under different flow rates, pressures, temperatures, for discharges between electrodes of different sizes and materials, fed by various power supplies . . . but, unfortunately, we cannot use them to design a power supply that would be sufficiently compatible with such sources of instability . . . well maintained for the "chemical" reasons. Therefore, it was necessary for us to invent new power supplies that could accommodate said GlidArc discharges.

In general, a GlidArc can be supplied with rectified direct current, single-phase alternating current, three-phase alternating current, or multiphase alternating current. As mentioned above, the GlidArc operates in a discharge state, compared to a conventional electric arc, with relatively high voltages (several kilovolts) and weaker currents (a few amperes). Thus, for the same electrical power, the intensity of the currents is much lower than in a conventional plasma torch. The voltage increases following the extension of the discharge channel. This extension is due to one or several causes, such as: the high turbulence of the medium where the discharge develops, the distance between the electrodes, the non-thermal conduction of the current through the medium.

In broad outline, the electrical power supply of a GlidArc must perform two functions: 1) ignite the discharges, and 2) deliver the electrical power into the discharge.

The following description will explain the mode of operation of the GlidArc discharge in relation to power supplies that have been previously used or described in the literature. To this effect, we will mention problems related to these power supplies, which will enable us to better position our new power supplies and circuits (assemblies), which are the subject of this invention.

FIG. 1 shows a mode of operation of the GlidArc-I that was previously described in BF 88.14932 (2639172). The direct current supply consists of two generators (G1) and (G2) connected in parallel to the terminals of two electrodes. The generator (G1) delivers the voltage necessary to ignite the discharges (.about.5 kV) for a current limited to 1 A. The generator (G2) delivers the power necessary to maintain the discharge while it is spreading. The voltages and currents can be limited to values of up to 800 V for the voltage, and 60 A for the current (which is unusually high for a purely thermal application). A resistance (R) adjustable between 0 and 25 .OMEGA., and a self-inductance (S) of 25 mH are connected in series between the positive terminal of the generator (G2) and an electrode, in order to limit both the direct current component and the current variations. Furthermore, a cutoff and protective diode (D) is placed in series with the resistance (R) and the inductance (S) in order to protect (G2) from the voltage delivered by (G1). The diode (D) will become conducting only when the voltage in the terminals of the electrodes is lower than or equal to the voltage measured at the terminals of the generator (G2) (immediately after the ignition of a discharge). The limitation of the current by the resistance (R) and inductance (S) makes it possible to maintain the discharge state below the arc state that does not allow for the proper operation of the device. The negative terminals of the generators (G1) and (G2) are interconnected and constitute the negative terminal of the power supply, which is connected to the other electrode.

FIG. 2 shows the variations of voltage at the terminals of the two electrodes, the variations of current, and the variations of the instantaneous power, respectively, which are plotted in relation to time for an average output of 9.5 kW and an airflow of 120 m.sup.3(n)/h. The air is channeled by a cylindrical conduct with an inside diameter of 85 mm, where two steel electrodes are attached. This recording was obtained with a digital oscilloscope. It shows a sequential process; the life of a discharge is approximately 6 ms, the mean current is 20 A, and the mean voltage is 480 V. The duration of a quasi-period can be extended or shortened according to the linear speed of the gas in the area between the electrodes, the nature of the flow, and the geometry of the GlidArc.

This FIG. 2 shows that, at the time of the ignition of the discharge, the dielectric breakdown voltage, which is a function of the shortest distance between the electrodes, should be in the order of several kilovolts while the current intensity does not need to be high. Unfortunately, it was not possible to lower this voltage by bringing the electrodes closer together, as these must remain separated at least by a few millimeters due to mechanical reasons. In fact, metal scales or deposits of any origin could produce short circuits.

Therefore, the gliding discharges have variable characteristics from the time that they are ignited to their extinction, with, in particular, energy dissipation values that increase over time (and which may reach values comparable to those of the arc state). In FIG. 3, we plotted the "cloud" of experimental points originating from the current-voltage characteristic (shown in FIG. 2), which corresponds to the preceding operating conditions. This characteristic highlights the turbulent and discontinuous operation of this discharge. This is precisely the type of operation that makes it possible to obtain a relatively cold (or warm) plasma that is in highly thermodynamic nonequilibrium.

Therefore, our observations indicate a significant drop in voltage between the electrodes immediately after the ignition. Although this voltage increases along the path of the discharge between the diverging electrodes, it is never as high as the voltage achieved between the electrodes at the time of the first breakdown. In fact, the voltage required by the successive breakdowns is not as high as that required for the first breakdown, unless there is an extended interruption causing a partial deactivation of the ions that are present between the electrodes and which facilitate the successive reignitions. Finally, the mean voltage between the electrodes is comprised between a few hundred volts and 2 kV, depending on the nature of the gas, its temperature and pressure, the distance between the electrodes, the shape of the electrodes, etc. By definition, this voltage is much too low in relation to the voltage required for the ignition and, therefore, it appears that a "conventional" continuous voltage power supply would be difficult to apply. Thus, the power supply shown in FIG. 1 presents several drawbacks: the use of the resistance (R) to limit the current in the main power circuit causes substantial Joule losses in the form of heat unnecessarily dissipated outside of the GlidArc, the mean current is too high and the mean voltage too low to obtain a true nonequilibrium plasma source for some chemical conversions; this puts us rather in the area of an electric arc, two continuous power sources must be obtained (G1) and (G2) while the power distribution system is always alternating 50 (or 60) Hz, it is difficult to feed several electrodes from a single generator of this type.

Another type of electrical power supply was used in our numerous laboratory-based experiments. It is based on a system of "leak" or "lighting" transformers (single-phase 50 Hz, 230 V primary current, 10 kV secondary current, 1 kVA, inductive limitation of secondary current of 0.15 A). These are special single- or multiphase transformers with an increased magnetic resistance between the primary winding and the secondary winding (i.e., by separation). Several single-phase transformers can be interconnected within a three-phase circuit (system) to feed 3 or 6 electrodes, at different power levels (transformers placed in parallel) for "open circuit" effective voltages of 10 kV (or 5 kV) between each pair of electrodes set opposite each other, or 17 kV (8.5 kV) between adjoining electrodes (24.5 kV or 12.2 kV peak). This type of power supply is not optimal for potential industrial applications. The efficiency of these transformers is low (10-20%) because they operate for the most part under voltages that are much lower than their open circuit voltage. We also observed some loss of energy reflected by the heating of these transformers. This loss was measured in a "dead short circuit" state for two typical situations: 3 transformers, 3 kVA installed, power loss=0.58 kW, 6 transformers, 6 kVA installed, power loss=0.90 kW.

Instead of "leak" transformers, it is possible to use "rigid" transformers and separate self-inductances placed in series. In order to increase the output of such power supply, the power transformer could be linked to several pairs of electrodes connected in parallel. In this case, each branch must be separated from the secondary circuit by a series inductance. These inductances are used to charge their respective branches with a significant voltage drop (80-90% of transformer's rated voltage). Thus, the reactive power losses cannot be prevented.

Therefore, this type of power supply for GlidArc discharges has several drawbacks. In particular, their reactive power requirements are high because the initial voltage required to ignite the discharge is high. An electric field of at least 3 kV per mm of spacing between the electrodes is already required for a reliable ignition between the electrodes and in a gas (such as air) circulating at atmospheric pressure. This value is even greater for higher pressures or gases such as H.sub.2S or SO.sub.2 that capture free electrons. The ratio between the open circuit voltage and the mean voltage of the discharge in operation is quite high, meaning that the installed (reactive) power is much greater than the effective (active) power. In most cases, the latter should occasionally reach up to several tens or hundreds of kW for industrial applications, although we observed that only a small fraction of the "installed" power is actually transmitted towards the discharge. It rarely exceeds 30%, even for a GlidArc that has been optimized in terms of material flow and distance between the electrodes (which, as mentioned above, should be at least a few millimeters, otherwise the adjustment would be inaccurate or altered by the possible deposit of substance treated in a GlidArc reactor). Some capacitors were sometimes connected at the power intake in order to correct a very poor power factor. After the ignition of the discharge under the "open circuit" voltage applied to the electrodes and exceeding the dielectric breakdown voltage, this high open circuit voltage no longer helps in maintaining the discharge. However, a "leak" transformer must be build in order to support this voltage. Therefore, the solution providing for the separation of the ignition function from the discharge maintenance function, like the one presented in FIG. 1, appears to be the most beneficial.

Another solution to the power supply problem was proposed by J. E. Harry in a patent WO95/06225. FIG. 4 summarizes this solution, where an additional electrode (2) is placed between the two primary electrodes (1). The use of this third high voltage electrode, separated from the main power supply (Ap) which has a lower voltage, would make it possible to increase the separation between the power electrodes. The two primary electrodes (1) are fed by a main alternating current generator (Ap). An ignition electrode (2) fed with rectified current drawn from an auxiliary power supply (Aa) with an output of less than 500 W is positioned in an asymmetrical manner between these two electrodes. The two power supplies are connected by a common point (P), so that the dielectric breakdown voltage is exceeded between the electrode (2) and one of the two electrodes (1). A relatively powerful spark (with a current of approximately 0.1 A) can thus be generated, causing the ionization of the gas near these electrodes. This is sufficient to establish a main discharge between two electrodes (1). Thus the open circuit voltage of the main generator (Ap) could be reduced by half. However, FIG. 4 shows the presence of a resistance (Rp) in series in relation to the main power circuit; therefore, it constitutes a source of energy loss in the form of Joule's heat dissipated outside of the GlidArc device.

Another solution to the discharge ignition problem was proposed in a Romanian application No. 112225B (1994) by E. Hnatiuc and B. Hnatiuc. The solution presented in FIG. 5 consists in placing two auxiliary electrodes (A.sub.1) and (A.sub.2) between the primary electrodes (E.sub.1) and (E.sub.2). These auxiliary electrodes are independently fed from an additional power supply that is similar to that used for the electronic ignition of an automobile, see FIG. 6. It is a high voltage, low output power supply. This power supply enables the ignition of a "pilot" electrical discharge that pre-ionizes the space between the primary electrodes (E.sub.1) and (E.sub.2), and provides for the ignition of the main discharge at much lower supply voltages. This makes it possible to increase the energy output of the power supply up to 70%. The operation of this GlidArc-I device is controlled and adjusted through the modification of the phase of control pulses applied to the control grid of a thyristor (T) placed in the primary of an induction coil (BS) of which the secondary is connected to the auxiliary electrodes (A.sub.1) and (A.sub.2). The control pulses are generated by an integrated circuit. The electrical power supply assembly also contains a reactance coil (R) in series to limit the current in the main circuit.

However, for some applications, it could be difficult to add two auxiliary electrodes in the ignition area for the GlidArc-I reactor. Furthermore, this principle cannot be used to feed the GlidArc II type reactor. The adjustment of the distance between the primary electrodes (changing the performance of the device) and the simultaneous adjustment of the position of the auxiliary electrodes present significant technological problems.

Another electrical power supply for the GlidArc was proposed in a Polish patent PL301836A1 (1994) by T. Janowski and D. Stryczewska. FIG. 7 shows this solution, which is based on three single-phase transformers (Tr1), (Tr2), and (Tr3) supplied with 230 V by three phases (e1), (e2) and (e3) of the star-connected system, 50 Hz, 400 V. Thus, the three primary electrodes of the GlidArc are fed a three-phase current of medium voltage up to approximately 2 kV, with the possibility of adjusting this voltage (and, therefore, the dissipated power) within a range of approximately 10%. Three capacitors (C1), (C2) and (C3) are installed upstream from the power supply in order to correct the power factor. These main transformers have an inductive nature, which is marked by the series inductances (z1), (z2) and (z3). A fourth transformer (Tr4) recovers a very low pulsation due to the near magnetic saturation of the cores of the main transformers, between the floating node of the main circuit of (Tr1), (Tr2) and (Tr3), and the neutral of the electrical network. Thus, the primary of the power supply system has a low voltage with a triple frequency (150 Hz) which is then transformed by (Tr4) to a level of the order of 12 kV. This high voltage ignites a 20 mA discharge, thus performing the pre-ionization in the area where the three primary power electrodes are closest (approximately 2 mm). At this moment, the voltages generated by the transformers (Tr1), (Tr2) and (Tr3) act as a relay, by supplying the electrical power required to sustain the GlidArc discharges that develop between the primary electrodes, according to the rotation of the electric field. During the operation of the main discharges, the secondary of the transformer (Tr4) suffers a short circuit through these discharges.

Nevertheless, the system shown in FIG. 7 requires the use of a specific transformer operating as a near-saturated magnetic core, as it is the non-linearity of the magnetic feature of the core that produces an AC voltage of 150 Hz between the common point of the primary windings and the neutral. Without this voltage, it would not be possible to generate a high ignition voltage.

This invention proposes below several other new electric generators and specific circuits to improve the power supply of a very unstable high-voltage and relatively low current discharge such as GlidArc-I or GlidArc-II.

Ignition and Reignition Electrode Set in the Geometric Center of Two or More Power Electrodes and Supplied Independently From the Main Power Circuit

As shown in FIG. 8, the ignition and reignition circuit (3) and (4) is set up independently from the main power circuit that supplies the primary electrodes (1) of a very unstable electric discharge. This assembly is especially suitable for GlidArc-I type devices. It comprises an external transistorized ignition and reignition system with an additional electrode (2) set in the geometric center of two or more primary power electrodes. For example, the supply (V.sub.D) of the transformer (3) is 33 V, while the separation capacity (C.sub.S) is 2 nF. This assembly makes it possible to use commercial power transformers that do not need to be specifically built to provide for the saturation of the magnetic cores in order to generate a non-linear effect of which the purpose is to act as ferromagnetic amplifiers.

During the opening of the power transistor ("high level" of oscillator), the electric current intensity (I.sub.D) increases according to the exponential distribution law: I.sub.D=I.sub.0(1-e.sup.-1/.tau..sup.L) (1) defined by the time constant: .tau..sub.L=L.sub.1/(R.sub.1+R.sub.DS+R.sub.V) (2) and by the balance current: I.sub.0=V.sub.D/(R.sub.1+R.sub.DS+R.sub.V); (3) where L.sub.1 is the inductance of the primary winding of the transformer, R.sub.1 the ohmic resistance of the winding, R.sub.DS the "drain-source" resistance of the transistor, and R.sub.V the internal resistance of the power supply (V.sub.D). The secondary winding of the high-voltage pulse transformer (3) contains many more coils than the primary winding. Therefore, the quick variations of the magnetic flux in the core produce a strong electromotive force in the secondary circuit. Upon the interruption of the primary circuit ("high level".fwdarw."zero" transition of oscillator), the induced voltage (U) can be expressed according to the following formula (without taking into account the parasitic capacitance of the circuit):

.times..times..times.dd.epsilon..times. ##EQU00001##

Thus, the amplitude of the voltage (U) can be governed by:

The rate of variation of the current intensity (I.sub.D); it is given by the dynamic characteristic of the transistor used;

The amplitude of current intensity (I.sub.D) during the interruption of the primary circuit; as it happens, said amplitude can be controlled by the opening time of the transistor, according to formula (1).

The capacitor (C.sub.S) separates the ignition circuit from the main power supply circuit: it prevents the electric current of the GlidArc main power supply from flowing, after the ignition, through the pulse transformer. Therefore, the ignition voltage (U.sub.A) is reduced to the following value:

.times. ##EQU00002## where (C.sub.P) represents the parasitic capacitance of the cable. In order to maintain (U.sub.A) at the maximum level, it is necessary to ensure that (C.sub.P)<<(C.sub.S), meaning that the cable must be shortened as much as possible, and its insulation and path must be properly sized.

Because of the parasitic capacitance (C.sub.T) of the winding of the transformer, the secondary circuit resembles an RLC oscillating circuit of which the performance depends on the quality

.times..times. ##EQU00003## of the circuit (R.sub.2-resistance of secondary winding of transformer). A theoretical model of this type of oscillatory circuit with attenuation provides that if Q>1/2 (which was true in our experiments), the output voltage (U) is in the form of frequency oscillations f.sub.0=1/2.pi. {square root over (L.sub.2C.sub.T)}, of which the envelope is attenuated with a time constant of approximately L.sub.2/R.sub.2. By modifying the high-voltage pulse repetition frequency, it is possible to modify the state of the electric discharge connecting this ignition electrode with a power electrode:

If the time between two pulses is greater than the relaxation time of the oscillations, the discharge appears in the form of individual sparks, with a time separation between them.

If the time between two pulses is less than the relaxation time of the oscillations, there are no more barriers between the sparks. The discharge thus becomes continuous and resembles an alternating current luminescent discharge with a frequency f.sub.0.

This last state does not appear to be beneficial for the ignition and reignition of a very unstable high-voltage electric discharge, such as a GlidArc, since the pulse transformer remains in a quasi-permanent short circuit. On the other hand, the time between two individual sparks must be significantly lower than the duration of a GlidArc cycle (ignition-extinction-reignition), in order to minimize the dead time between two discharges. Therefore, it is preferable to adjust the parameters of the RLC oscillatory circuit so that Q.apprxeq.1/2. This provides for the fastest transmission of the electromagnetic energy of the circuit into the discharge.

During our power supply optimization tests described herein, we observed a new fact related to the shape of the ignition electrode (2). Contrary to the oblique shape proposed by J. E. Harry in FIG. 4 (taken from his patent), we propose a highly pointed shape, which is presented in FIG. 9b. It resembles the frame of a partially open umbrella, or a star (top view) with each branch extending towards one of the primary electrodes. This shape makes it possible to ignite discharges between electrodes that are significantly more distant from each other than those shown in FIG. 9a.

In fact, the distance between the primary electrodes (1) of the GlidArc-I should not vary too much from the diameter of the flow inlet nozzle. For example, for a large volume of gas, this diameter may reach several centimeters. Therefore, the distance between the electrodes must be adjusted according to this diameter and, as a consequence, the ignition voltage of the GlidArc increases. A system that may solve this problem is based on the use of an additional ignition and reignition electrode (2) placed in the ignition area, in the geometric center between the electrodes (1), of which the shape is shown in FIG. 9b. This additional electrode receives a very high voltage (several tens of kV), which is superimposed on the electric potential of the primary electrodes (1) by a few kV. This high voltage can be supplied, for example, by a generator presented in FIG. 8. Consequently, the spark is ignited in the electric field that rotates successively between each of the primary electrodes (the example provided in FIG. 8 shows six electrodes, each of which is connected to a 50 or 60 Hz six-phase generator) and the ignition and reignition electrode (2), thus covering the entire ignition area, in spite of minor differences in the distances between the electrodes. These very short electric discharges (typically lasting a few tens .mu.s--depending on the nature of the ignition circuit) form a conducting zone for the ionized gas between the electrodes, which creates a current path for the main circuit, thus igniting the GlidArc-I. Furthermore, during the operation of the GlidArc-I, the ignition occurs in an automatic and selective manner: the electrode without discharge and, therefore, under a higher electric potential than the other electrodes, is the first to be short-circuited by a spark. Considering that, in this case, the main power supply may be designed for lower output voltages, its performance increases significantly.

The shape of the ignition and reignition electrode shown in FIG. 9b was designed after taking into consideration four different aspects:

Ignition Aspect

The ignition and reignition electrode is shaped like a star (top view), with each of n branches (where n is the number of phases of the main power supply; FIG. 9 shows a six-phase circuit) extending towards one of the primary electrodes (1), which have such distance between them that the main discharge could never self-ignite without the electrode (2) activated by the ignition and reignition circuit. After the ignition of the GlidArc-I, this electrode acts like a short-circuit bridge between the primary electrodes: these very unstable discharges glide over the central electrode in the gas flow (FIG. 9b), until they meet in the middle of the electrodes. This phenomenon can be obtained because of the diverging shape (side view) of the central electrode (2.) Thereafter, the discharges spread freely between the primary electrodes (1) until they are extinguished.

Aspect of Gas Flow

The shape of the ignition and reignition electrode (2) is also adapted to the flow that runs around it. The flow runs between the branches of the star and allows the discharges to glide over the electrode without creating a flow diversion area. Thus, this shape of the electrode (2) also provides for the thermal exchange with the flow and keeps this electrode from overheating.

Thermal Aspect

The shape must also guarantee a thermal balance between the different parts of the electrode (2): this means that the electrode that heats up the quickest on the surface making contact with the discharge must be strong enough to allow for a thermal flow between the different branches. The electrical power dissipated in the central electrode (2) can be calculated according to the following formula: P.sub.EA.apprxeq.bn(U.sub.CI+A.rho.I.sup.2), (6) I--electric current of GlidArc through one electrode, in Amperes; U.sub.C--cathodic potential drop of discharge plasma, in Volts, given by the plasma-forming gas and the electrode material used; specific resistance of electrode material in .OMEGA.m; A--geometric factor of electrode in m.sup.-1; n--number of primary electrodes (and phases feeding them); b--factor representing the fraction of the life cycle (ignition-primary unstable discharge-extinction-reignition) of the GlidArc during which the electrical current runs through the ignition electrode. Its value can be calculated as:

.apprxeq..times..times..times..times..times..times..times..times..times..t- imes..times..times. ##EQU00004## In our tests k.apprxeq.0.1.

The first term of the sum (6) represents the portion of electrical power due to the discharge plasma. This power dissipates on the surface of the branches of the star; therefore, a good heat dissipation towards the volume of the electrode must be provided. The second term represents the losses in the material of the electrode due to the Joule effect. It may be ignored in the case of metal materials with a very low .rho.. On the other hand, for conducting refractory materials, this term can be quite significant. In fact, the dissipation of electrical power in the ignition electrode is offset by the thermal exchanges with the flow.

Aspect of Electric Field

The minimum intensity of the electric field in a gas, from which an independent discharge is ignited, is determined by the nature of the gas and the concentration of gas molecules (Paschen's law). For a distance d between two electrodes, the maximum value of the intensity of the electric field E.sub.MAX.sup.R varies according to the minimum radius of curvature R of the electrodes. If we take an electric field between flat electrodes E.sub.MAX.sup..infin.=U/d for R>>d as reference, the influence of R can be determined according to the following formula:

.infin..ident.d.function.d ##EQU00005##

With d=5 mm and for R=1 mm: E.sup.R=2.8. For R=0.1 mm, E.sup.R increases to 13. Therefore, it is highly advisable to design the ignition and reignition electrode with a shape featuring tips characterized by a relatively small radius of curvature (tenths of mm). However, when they are exposed to electric discharges with high current densities, these tips can wear out during their use. Therefore, it is preferable to use metals that have a high melting point or refractory materials-electrical conductors.

B. Self-Contained Ignition and Reignition Device and Circuit Feeding Two Power Electrodes

Another solution proposed in FIG. 10 pertains to the use of a special transformer as a power supply. The transformer comprises two low-voltage primary windings (P.sub.1) and (P.sub.2) and one high-voltage secondary winding (S). The aim of the two primary windings is to superimpose the effects produced by each primary winding onto the secondary winding (S). The first power winding (P.sub.1) is connected to the mains supply, e.g. 220 V. However, the mains supply is separated by a filter (F). The second ignition winding is designed to be fed pulses of adjustable amplitude and phase. This winding has a rated voltage of 24 V, but it can withstand higher voltages of up to 200 V, for short periods of time. The filter (F) of the mains supply stops the spreading of the pulses induced from the winding (P.sub.2) into the winding (P.sub.1), which could