WO2008136698A1

WO2008136698A1 - Method and apparatus for an efficient electrohydrodynamic flow control of a gas

Info

Publication number: WO2008136698A1
Application number: PCT/RU2007/000225
Authority: WO
Inventors: Thomas Hammer; Andrey Mikhailovich Bartenev
Original assignee: Siemens Aktiengesellschaft
Priority date: 2007-05-04
Filing date: 2007-05-04
Publication date: 2008-11-13

Abstract

In a method for flow control of a gas with means of Electrohydrodynamic (EHD) actuating the flow the EHD-actuation generates a small discrete discharge region and a larger drift region respectively, which are both controlled independently in view of time and in view of locations(loci). An Apparatus for such a method including means for a dielectric barrier discharge (DBD) has at least three electrodes (1-3), of which a first voltage (U1(t)) between the first and second electrodes (1, 2) generates a discharge region and a second voltage (U2(t)) between the second and a third electrode (2, 3) generates a drift region (10).

Description

METHOD AND APPARATUS FOR AN EFFICIENT ELECTROHYDRODYNAMIC

FLOW CONTROL OF A GAS

The invention is related to a method for an efficient Electrohydrodynamic flow control of gases and an apparatus therefore. The invention is used in known turbo-machines or other devices .

Technical Problem:

Turbo-machines such as compressors or turbines are optimized for certain operation conditions characterized by thermal and transport properties of the gas mixture applied (density, specific heat, viscosity, heat conductivity) , temperature, and mass flow. If turbo-machines are operated at other than the optimum operation conditions, the efficiency decreases, and even under optimized operation conditions some inherent losses cannot be avoided: Mixing losses occur due to mixing of the flow fields at the trailing edge of a turbine blade caused by the velocity difference between pressure side and suction side of the turbine blade profile. The shear stresses in the boundary layers generate boundary layer losses. In the worst case boundary layer separation may occur. Other losses are clearance gap losses at rotating blades and end-wall losses, which may also occur at the stator. Further due to interactions between rows of blades (wake formation) cascade losses are generated.

Another problem related to turbo-machinery is heat transfer, especially at the vanes and blades of gas turbines exposed to high temperature gas flows. Here film cooling is utilized in order to protect the surface from overheating, which again is optimized for a certain range of operation conditions. Turbo-machines for industrial application such as electric power plants are characterized by high mass-flow densities and high energy densities and therefore often operate at atmospheric pressure and above.

Further heat transfer plays an important role in fossil fuel fired boilers and in the walls of gas turbine combustors. Again the formation of a stable thin boundary layer at the surface of the heat exchanger is of importance for the efficiency of the heat transfer.

Known solutions of the problems:

In all these examples mechanical actuators are required in order to adapt the turbo-machine to different operation conditions. However, the inherent losses can hardly be influenced by mechanical actuation. Recently the application of electro-hydrodynamic flow control has been suggested for the reduction of flow losses at air foils of aero-engines, for the modification of the heat flow at a combustor head, and for the reduction of sound emission from aero-engines. Electro-hydrodynamic flow control is based on momentum transfer of charged particles accelerated in an electric field to the surrounding neutral gas, which e.g. under certain conditions forms the so-called ionic wind.

Investigations to latter objects are presented at 43

"Aerospace Science Meeting", January 10, 2004, Rano, Nevada

(US) or published in "Journal of Physics D: Applied

Physics" 38 (2005), p. 3635 to 3642 and additionally partly covered by WO 2002/081304 Al.

The actuators suggested in this application are either based on dielectric barrier discharges (DBD) known from ozonizers or on DC-Corona discharges (DC-CD) known from electrostatic precipitators . Unfortunately these actuators are limited in their power density because both in the case of the DBD and of the DC-CD the charge density and the electric field must not exceed certain limits.

In the case of the DBD the electric field at the dielectric barrier needs to be below the dielectric breakdown field, and thus the charge transferred per half cycle of the applied AC-voltage (or per pulse in the case of a pulsed DBD) is limited to: Q = E_bmax x d_b x C_b (1) , where Ebmax is a maximum electric field below the breakdown limit of the barrier material, d_b is the thickness of the barrier material, C_b is the capacity of the dielectric barrier. Since the barrier capacity is given by:

with the relative permittivity of the barrier material ε_b, the vacuum permittivity ε₀, the surface area of the barrier Ab. The maximum charge transferred per half cycle cannot be influenced by the thickness of the barrier material . Increasing the barrier surface and thus the total cross section of dielectric barrier discharge filaments through which the electrical current is flowing increases the total charge, however, the charge density as well as the current density related to the discharge cross section remains constant. Thus the power density also remains constant. Further in the case of the DBD the charge transferred across the discharge gap integrated over a complete cycle of the applied voltage must necessarily be zero. Thus momentum transfer can only be achieved in that there is an asymmetry between electron- and ion-flow: Because ions have a larger mass than electrons they can transfer momentum much more efficiently to the neutral gas than ions . Further in the case of a DBD the reduced discharge gap being the product of discharge gap d_g and pressure p is limited to a few mm x bar. Thus the energy transferred from the electric field to the neutral gas flow, which is proportional to the product of the average electric field force <Q x E> and the gap d_g along which this force is exerted, is limited.

Finally, if full use shall be made of the momentum transfer from the ions generated during a half cycle of an

AC-excited DBD, an electric field forcing the ion drift to the dielectric barrier or electrode surface needs to be applied to the discharge gap for the full ion transit time.

This limits the maximum frequency of the AC-voltage and thus the average electric power supplied to the DBD actuator to:

where μi_on is the ions mobility, and E_av is an average value of the electric field acting on the ions. In the case of a DC-CD the applied voltage and thus also the current density for given design parameters are limited by the spark breakdown voltage .

In contrast to a DBD the DC-CD can be operated at large reduced discharge gaps of some 100 mm x bar, however, the power density which can be achieved without sparking is much lower than that of a DBD .

Thus none of the EHD-actuators being described so far can advantageously be applied to industrial turbo- machinery. The focus of the current invention is therefore an improved technology for the manipulation of gas flows being in contact with solid surfaces, which is not based on movable mechanical parts and avoids the disadvantages of DBD- or DC-CD-actuators. The invention allows electric flow control near surfaces and can be applied to any gas, especially to air, steam, and exhaust gases of combustion processes.

The invention is defined by a method with the steps of claim 1. An apparatus is defined in claim 11. Further realisations of the inventive method and the related apparatus are given by the dependent claims . Finally claim 21 to 25 give preferred uses of the inventive method and apparatus in a turbine and other devices .

Object of the invention is the electrohydrodynamic actuation of a gas flow by means of EHD-actuator, which other than the DC-CD or the DBD-actuator makes use of separately controlled ion formation in an electrical gas discharge and momentum transfer to the gas flow in an ion drift zone. Thus this EHD-actuator transfers momentum and energy much more efficiently to the gas flow than a DBD- or a DC-CD-actuator.

The basic idea of the invention is that in a first step in a limited volume being substantially smaller than the volume of interest for EHD-actuation large electron-ion pair densities are generated efficiently by means of a very short gas discharge pulse, which is characterized by an electric field exceeding the gas discharge breakdown value substantially. In a second step following charge generation an electric field is generated by means of an electrode structure consisting of at least one anode and one cathode in the complete volume of interest for EHD-actuation, whereby the electric field is large enough to separate the electron-ion pairs and to force the electrons to drift to the anode (s) and the ions to drift to the cathode (s) of the electric field generating electrode structure without causing spark-formation between the electrodes. In a preferred solution the electric field in the second step is kept below the gas discharge breakdown limit of the electrode structure. In this case the second step can be considered as a non self-sustained gas discharge. Advantages of the Current Invention: The current invention combines the advantage of a short time, small volume gas discharge plasma, that charges can be generated very efficiently at a high density, with the advantage of a DC-corona or a non self-sustained gas discharge, that charges can be transported over a large distance. Thus there is a large flexibility in the working point of the EHD-actuator allowing more flexible operation and better increase of efficiency of turbine machinery or heat exchangers . In addition to these improvements the invention allows to build turbines and heat exchangers more compact, because the improved properties of the boundary layer, e.g. at the surface of a turbine blade, allow to reduce the number of stages of a gas turbine or of an axial compressor. More advantages and details of the invention are given in the description of preferred embodiments in respect to the drawings. There is shown schematically:

Figure 1 - the method performed in a set-up for direct electrical gas discharge generation between two electrode tips and a third wire electrode for charge separation and momentum transfer,

Figure 2 - the method performed in a set-up for direct electrical gas discharge generation between parallel wire electrodes and a third wire electrode for charge separation and momentum transfer,

Figure 3 - the method performed in a set-up for dielectric barrier discharge generation between two electrode tips and a third wire electrode for charge separation and momentum transfer, and an enlarged view of the electrode set-up for DBD-generation in figure 3A,

Figure 4 - the method performed in a set-up for dielectric barrier discharge generation between two parallel wire electrodes and a third wire electrode for charge separation and momentum transfer,

Figure 5 - a view onto a wall including four electrodes for generating the discharge region and a drift region respectively, Figure 6 - the cross section of the set-up of Figure 5,

Figure 7 - a first graph presenting the voltages over time applied to the electrodes of the electrical gas discharge and to the ion extracting cathode, respectively, Figure 8 - a second graph presenting the voltages over time applied to the electrodes of the electrical gas discharge and to the ion extracting cathode, respectively,

Figure 9, 10 - the use of the embodiments of one of the figures 1 to 8 to the blade of an axial compressor or of a turbine,

Figure 11 - the use of printed electrodes on an insulating film of an conductive body and

Figure 12, 13 - the use of the embodiments of one of the figures 1 to 8 to a heat exchanger. In figure 1 there are two electrodes, a first electrode 1 and a second electrode 2 which are supplied by a high voltage source 11 with a time dependent voltage Ui (t) . The voltage amplitude Ui₀ is set such that between the electrode tips a self-sustained electrical gas discharge plasma is generated. If d_g is the gap between the electrode tips and n_gas is the number density of the gas molecules filling the discharge gap, then preferred values of the reduced electric field being approximated by

Eav/n_gas = U₁₀/ (n_gas x d_g) (4.1) are in the range of IxIO^"19 Vm² to 5xlO^'19 Vm². The region 5 between the electrodes 1 and 2 is called "discharge region" .

In a certain distance da_r to the opposed electrode tips there is a third, linearly extended electrode 3 supplied by a voltage source 12 with another time dependent voltage U₂ (t) normally having a negative polarity. Thus electrode 3 normally is a cathode. The voltage amplitude U_2O is set such that it causes separation of the charges generated in the discharge region 5, and drift of the positive ions with an electrical current 7 to the cathode without generating a self-sustained gas discharge. The region 6 between the electrodes 1/2 and the electrode 3 is called "drift region" . Preferred values of the reduced electric field in the drift region 6 being approximated by E_av/n_gas = U₂₀/ (n_gas x d_dr) (4.2) are in the range of IxIO^"20 Vm² to IxIO^"19 Vm². For the efficient application of the method it is essential that the discharge region 5 and the drift region 5 being in direct contact to the discharge region 5 are well defined by the placement of the electrodes 1, 2 and 3, and that the gas discharge generation in the discharge region 5 and the momentum transfer to the gas flow by means of ion drift in the drift region 6 are controlled separately with respect to reduced electric field and time by different voltages Ui (t) an U₂ (t) .

The shape of the electrode tips 1, 2 may be arbitrary. They may be rounded or sharp edged and can simply be formed by cut wires or thin rods mounted with insulating clamps, by pasting thin conductive foils pasted on an insulating surface, or by conductive layers printed on an insulating surface using conductive ink.

In figure 2 the electrical gas discharge plasma generating means of figure 1 are changed in that way that the first and the second electrode 21 and 22 are parallel wires, strips, or rods extending over a certain length being substantially larger than the discharge gap between the electrodes. This means that the discharge region 5 located between the electrodes 21 and 22 is a linear source of ions supplied to the drift region 6. Thus this set-up is advantageous if a gas flow needs to be controlled near to a wide surface such as a wing of an airplane or a turbine blade .

In figure 3 the electrical gas discharge plasma generating means of figure 1 are changed in that way that a dielectric barrier discharge (DBD) is generated in the discharge region 5. Electrode 31 has a coating 34 of a dielectric material acting as a dielectric barrier, whereas electrode 32 is uncoated. Details of the DBD-electrode setup are shown in the enlarged view of figure 3A. Electrode 32 is connected to ground and for the charge separation thus acts as the anode, the cathode 33 being connected to the voltage supply 12 is the same as in figure 1. The advantage of this set-up compared to that shown in figure 1 is that more intense electrical gas discharge plasmas can be formed without the risk of sparking. Thus larger ion currents can be supplied to the drift region, resulting in a larger momentum transfer.

In figure 4 the electrical gas discharge plasma generating means of figure 2 are changed in that way that a linearly extended DBD region 5 is formed: Electrode 41 has a coating 44, whereas electrode 42 is uncoated. Again electrode 42 and cathode 43 define the drift region 10 being much wider than that generated by the set-up given in figure 3. Compared to the set-up in figure 2 more intense gas discharge plasmas can be formed without the risk of sparking. Thus larger ion currents can be supplied to the drift region, resulting in a larger momentum transfer.

In figure 5A and figure 5B a device with four electrodes applied to a surface is shown. There is a wall 50 of electrically insulating material acting as a dielectric barrier, into which flat, stripe-shaped gas discharge electrodes 51 and 52 are embedded. On top of the surface being in contact with the flowing gas two stripe- shaped drift electrodes 53 and 54 are placed. The electrodes 51 and 52 are connected to the voltage-source 11 supplying U_x (t). The voltage-source 12 supplying U₂ (t) is connected to the electrodes 53 and 54. The electrodes are placed such that the discharge region 5 and the drift region 6 overlap with the discharge region being near to the anode (53) side of the drift region.

In figure 6 and figure 7 graphs representing of the voltages over time are shown: The abscissa is the time t in arbitrary units, and the ordinate is the voltage U in arbitrary units, respectively. There are graphs 61, 62 for U₁ (t) and U₂ (t) in the case of a direct electrical gas discharge utilized for ionization, and graphs 71, 72 for U₁ (t) and -U₂ (t) in the case of a DBD utilized for ionization, respectively. In the case of a direct discharge of figure 6 short high-voltage pulses are required in order to achieve high reduced electric field strength for efficient ionization at one hand and to avoid sparking at the other hand. In the case of a DBD short high-voltage pulses may also be used, however, it is much easier to generate these electrical gas discharges by applying an AC- voltage to the electrodes . Of course this could be done by applying the voltage continuously, but then the amplitude needs to be kept low in order to adapt the ionization rate to the maximum possible ion removal rate due to charge separation and drift. Since this would reduce the efficiency of ion generation substantially, the preferred solution is the application of short sine- or square-wave packages having an amplitude which is optimized for ion generation. Of course other wave shapes than sine-wave or square-wave may be utilized for the generation of wave packages .

The time interval between two high-voltage pulses or wave packages usually is long compared to the duration of the high-voltage pulses or wave packages . Preferred values for the duration of the high-voltage pulses are in the range of 1 ns to 10 μs, whereas the duration of wave packages has preferred values ranging from 10 ns to 100 μs . Preferred values for the intervals between pulses or wave packages range from 10 μs to 10 ms .

The described method and apparatuses can be used in gas turbines, whereby gas flow induced losses at turbine blades are reduced or also in steam turbines, whereby gas flow induced losses at turbine blades are reduced. Another use is in an axial compressor, whereby gas flow induced losses at compressor blades are reduced. Also in wings boundary layer separation is suppressed. Finally there is use in heat exchangers, whereby the heat transfer per unit surface area of metal tubes is increased.

In figure 9 and figure 10 the gas flow around the blade of an axial compressor or of a turbine is shown. Without EHD-actuation in figure 9 mixing-losses occur, because at the trailing edge of the blade the flow velocities from the pressure- and from the suction-side of the profile are different. Figure 10 shows the blade 90 of figure 9 with means for EHD-actuation. Due to momentum transfer to the boundary layer gas flow on the suction side of the profile mixing losses are avoided. In the same way boundary layer separation can also be avoided.

Since a lot of structures, to which EHD-actuation may be applied, are metallic, in figure 11 an example of an electrode set-up attached to a conductive body 105 is presented. On the surface of the conductive body 105 a first insulating layer 110 is prepared e.g. by plasma spraying of alumina in the region where the EHD-actuator shall be placed. At least one discharge electrode 111 is prepared by pasting a thin metal foil or printing a conductive pattern on top of that layer 110. Then this layer 110 carrying electrode structures is covered by another insulating layer 110', on top of which two electrodes 112 and 113 are prepared in the same way as the electrode 111. Electrode 112 is connected to ground and works both for discharge generation and drift activation. In order to get a smooth surface the complete dielectric insulated electrode-setup can be placed in a cut-out region 106 of the conductive body 105. In figure 12 there is presented a heat exchanger with tubes 120 and a structure of separation parts 130. The gas flows around the tubes 120. Because of the boundary layer separation the heat transfer may be inefficient. Figure 13 shows the tubular heat exchanger with means for EHD- actuation included. There are two electrodes 121, 122 and a voltage power supply 125 for AC and a voltage power supply 125 for DC. Here it is assumed, that the tubes 120 of the heat exchanger are metallic, and the structure 130 carrying the tubes 130 is insulating or has an insulating surface. Due to EHD-actuation at the surface of the structure carrying the tubes re-attachment of the gas-flow to the tube-surface is achieved. Thus the heat transfer is improved, and the heat-exchanger can be built more compact.

So far the different examples are described in their structure. Many changes are possible. More details of the function of the devices are given below.

The gas discharge required for generation of electron- ion pairs in a limited volume can be excited periodically by application of a pulsed DC-voltage or of a short high- frequency voltage pulse U₁Ct) to a pair of electrodes with electrode 1 and electrode 2 generating the electric field required for gas discharge breakdown in this limited volume (e.g. "discharge region"), only. In the case of pulsed DC- voltage this may either be a pulsed corona discharge or a pulsed dielectric barrier discharge. The barrier discharge may be formed either as a volume discharge or as a surface discharge . In the case of excitation by means of a short high- frequency voltage pulse this may also be a dielectric barrier discharge, or it may be a capacitively coupled radio-frequency discharge, which is distinguished from the dielectric barrier discharge in that the surface charging of the barrier material is not essential for the gas discharge characteristics. Thus the dielectric breakdown of the barrier material is no longer a limiting factor to the power density.

In each case both the reduced electric field in the gas discharge region and the pulse duration of the AC- or

DC-voltage will be controlled such that the efficiency of ion generation is maximized. If the reduced electric field is near to the value required for gas discharge breakdown, electronic loss mechanism such as vibrational and electronic excitation of molecules will prevail, and if the reduced electric field is more than five times of the gas discharge breakdown value, the efficiency of ionization as a function of the plasma input energy per pulse will saturate, however, the efficiency of the power coupling between electric power supply and electrical gas discharge will drop drastically. Thus the product of these efficiencies will maximize at a certain reduced electric field value. Further the duration of the electrical gas discharge pulse is of importance for the efficiency of ion generation: At the beginning of each pulse the ion density in the gas discharge region will increase exponentially as a function of time because ions are generated due to electron collision ionization at a rate being proportional to the electron (or ion) number density and a rate coefficient depending on the reduced electric field. Since ion losses are caused by ion drift and diffusion to surfaces, especially to the electrodes, and by ion-electron recombination, an increase of the ion loss rate with increasing ion number density can be assumed. Thus after some time a quasi-steady state ion concentration would be achieved resulting in an increase of energy consumption without any increase in ion density if too long pulses were applied. The optimum pulse duration is reached when the ion density approaches 50-70 % of its saturation value which of course depends on the reduced electric field in the discharge gap. The voltage U₂ (t) driving the drift in the second step may be applied as a DC-voltage all the time, or it may be a time-dependent voltage being switched on after the gas discharge voltage of the first step has been switched off and being switched off before the gas discharge voltage is been switched on. In the case of time dependence this voltage may be bipolar, as long as by some means such as the design of the electrode arrangement or by the time- dependence of the voltage itself a predominant ion-flow direction is achieved. The duration of the drift voltage pulse being only slightly smaller than the period between two gas discharge pulses should be such that most of the ions are removed from the gas discharge region. However, a small fraction of the electrons and ions may remain in the gas discharge region in order to facilitate ionization in the next gas discharge pulse. This fraction is preferably between 0.01 % and 1 % of the maximum ion density.

The preferred case, however, is that of a unipolar applied voltage. The drift may be along a surface, it may, however, be also a drift between electrodes placed somewhere in the volume of interest. Further the direction of the electric field causing the drift may include any angle with the direction of the electric field applied to the gas discharge.

As described above the gas discharge may especially be perpendicular as shown in Figures 1 and 3 , or it may be parallel as shown in Figures 2, 4, 5, 8, 9, and 10 respectively. Normally it is not bordered exactly and can be frayed in respect to the geometry of the electrodes.

Claims

1. A method for control of a gas flow with means electrohydrodynamically (EHD) actuating the flow, whereby the EHD-actuation is induced by ion generation in a small volume due to a spatially confined electrical gas discharge followed by momentum transfer to the gas flow due to charge separation and ion drift in a spatially extended drift region, where the ionization and the momentum transfer are controlled independently of each other.

2. The method of claim 1, whereby the electrical gas discharge is generated periodically by a pulsed DC- or AC-voltage applied to a pair of electrodes such that the electric field generated in the gas discharge region exceeds the gas discharge breakdown field substantially,

- the pulse duration of the electrical gas discharge is controlled such that the efficiency of ion generation is maximized,

- charge separation and ion drift are induced by a constant or a slowly varying voltage applied to an electrode structure consisting of at least one anode and one cathode in the complete volume of interest for the EHD- actuation, the pause between two successive gas discharge pulses is controlled such that the main fraction of the ions generated in the discharge region is removed to the cathode before the next electrical gas discharge pulse is applied.

3. The method of claim 2, whereby the electric field in the drift region is large enough to separate the electron-ion pairs and to force the electrons to drift to the anode (s) and the ions to drift to the cathode (s) and low enough to avoid self-sustained gas discharge formation between these electrodes .

4. The method of claim 3, whereby the reduced electric field applied to the gas discharge region is kept between lxlCT¹⁹ Vm² and 5xlO^~19 Vm² and the reduced electric field applied to the drift region is kept between IxIO^"20 Vm² and IxIO^"19 Vm².

5. The method of claim 4, whereby the duration of the gas discharge pulse is between 10 ns and 100 μs and the period between two successive gas discharge pulses utilized for momentum transfer due to ion drift is between 10 μs and 10 ms.

6. The method of one of the preceding Claims 1 to 5 , whereby the electrical gas discharge is a pulsed corona discharge .

7. The method of one of the preceding Claims 1 to 5, whereby the electrical gas discharge is a dielectric barrier discharge excited by a DC- or an AC-voltage pulse.

8. The method of one of the preceding Claims 1 to 5, whereby the electrical gas discharge is a capacitively coupled radiofrequency discharge.

9. The method of one of the preceding Claims 1 to 8 , whereby the electrical gas discharge is a surface discharge.

10. The method of one of the preceding Claims 1 to 8, whereby the electrical gas discharge is a volume discharge burning in contact to a surface guiding the gas flow.

11. Apparatus for the method of claim 1 or one of the claims 2 to 10, with means for generating an electrical gas discharge and means for exerting forces on the charges generated in the electrical gas discharge, whereby there are at least three electrodes defining a discharge and a drift region and a power supply generating two time- dependent voltages, the first voltage (U_3. (t)) applied between the first and second electrode (1, 2) for generation of an electrical gas discharge in the discharge region and the second voltage (U₂Ct)) applied between the second and the third electrode for charge separation and generation of an ion drift in the drift region ( ) .

12. Apparatus of claim 11, whereby at least one electrode of the first or second electrodes (1, 2; 21, 22; 31, 32; 51-54) is coated with a dielectric barrier (23, 33) .

13. Apparatus of claim 12, whereby the first electrode (31) is a rod with the discharge barrier (33) around it.

14. Apparatus of claim 12 , whereby at least on electrode of the first and second electrode (1, 2; 21, 22, 31, 32; 51-54) is a flat electrode (21, 22) with a dielectric coating (23) thereon attached to an insulating surface .

15. Apparatus of claim 11, whereby there is are four electrodes (51-54) , two of them defining the discharge region (5) , and the other two defining the drift region (10) .

16. Apparatus of claim 15, whereby the two electrodes (52, 52) defining the discharge region are attached to an insulating surface and coated by a dielectric barrier (50) .

17. Apparatus of claim 11 or one of the following claims 12 to 16, whereby the first voltage U₁ (t) and the second voltage U₂ (t) generated by the power supply are pulsed DC-voltages.

18. Apparatus of claim 11 or one of the following claims 12 to 16, whereby the first voltage Ui (t) is a pulsed AC-voltage with a frequency between 20 kHz and 120 MHz and the second voltage U₂ (t) is a pulsed DC-voltage.

19. Apparatus of claim 17 or claim 18 equipped with an electronic control unit (ECU) , a pressure transducer, and a temperature sensor, whereby the control unit controls the reduced electric fields resulting from the first voltage Ul (t) and the second voltage U₂ (t) depending on gas pressure and temperature.

20. Apparatus of claim 11 or one of the following claims 12 to 19 for controlling the gas flow near to the surface of a conductive body, whereby on a fraction of the surface of the body one or more insulating layers are provided carrying the electrodes.

21. Use of the method of claim 1 or one of the claims 2 to 10 with an apparatus of claim 11 or one of the claims 12 to 20 in gas turbines, whereby gas flow induced losses at turbine blades (50) are reduced.

22. Use of the method of claim 1 or one of the claims 2 to 10 with an apparatus of claim 11 or one of the claims 12 to 20 in steam turbines, whereby gas flow induced losses at turbine blades (50) are reduced.

23. Use of the method of claim 1 or one of the claims 2 to 10 with an apparatus of claim 11 or one of the claims

12 to 20 in an axial compressor, whereby gas flow induced losses at compressor blades (50) are reduced.

24. Use of the method of claim 1 or one of the claims 2 to 10 with an apparatus of claim 11 or one of the claims 12 to 20 at wings, whereby boundary layer separation is suppressed.

25. Use of the method of claim 1 or one of the claims 2 to 10 with an apparatus of claim 11 or one of the claims 12 to 20 in heat exchangers, whereby the heat transfer per unit surface area of metal tubes (111) is increased.