EP1990097A2 - Wall-flow filter with unlimited creep rupture behaviour stability - Google Patents

Abstract

The method involves channeling an exhaust gas stream through a channel (5) of a porous ceramic body, where the channel runs in a longitudinal direction of the ceramic body. A voltage is applied to the ceramic body for generating an electrical field in the channel. Unipolar voltage impulses are selected such that a ratio of drift velocity of a sooty particle in a field direction of the electrical field to flow velocity of the gas flow in the channel is greater than or equal to a ratio of twice of a channel height (h) to a channel length.

Description

The invention relates to a method for operating a filter arrangement for separating soot particles from an exhaust gas flow, in which the exhaust gas flow is passed through passages of the ceramic body extending in the longitudinal direction of a porous ceramic body, according to the preamble of claim 1.

In this type of filter system, the exhaust gas flow passes through pores of the walls of the channels of the ceramic body, which are open only on one side, whereby the soot particles are retained. In this so-called "wall-flow filter", the deposition of the soot particles thus takes place mechanically. Different systems are known for the degradation of the deposited soot particles in the channels of the ceramic body, for example by means of a plasma generated in the channels of the ceramic body ("plasma-generated filter systems"). For this purpose, a voltage is applied to the ceramic body for generating an electric field in the channels of the ceramic body, which is oriented in each case normal to the axis of the channels on parallel to the channels extending electrodes. Electric field strengths of about 1 kV / cm in the channels of the honeycomb body are usually sufficient to generate a plasma in the channels, which converts deposited soot particles into gaseous substances.

The wallflow filters in use today, with which the exhaust gas from diesel-powered internal combustion engines is cleaned of soot particles, the biggest disadvantage is their limited creep strength. By oil ash, the porous walls of the ceramic body are continuously added, and there is an irreversible pressure build-up, which makes a very expensive change of the filter unit after about 150,000 km to 200,000 km necessary. Moreover, the pressure build-up also increases the fuel consumption significantly.

This problem will be particularly critical in the future, as the increasing proportion of biodiesel in diesel fuels accelerates this irreversible pressure build-up. The phosphate content in biodiesel is converted by the high regeneration temperature necessary in today's systems to phosphorus, which in addition to the oil ash leads to a very rapidly increasing, irreversible increase in pressure, which no longer enables the mileage achievable today from 150,000 km to 200,000 km. This difficulty currently applies only to a lesser extent to the wall-flow filter regenerated with an alternating current plasma, but becomes more effective when the proportion of biodiesel in the fuel exceeds the 10% limit.

It is therefore the object of the invention to increase, by a suitable method, the creep rupture strength of filter assemblies based on the use of a closed-sided ceramic body.

These objects are achieved by the features of claim 1. Claim 1 relates to a method for operating a filter arrangement for separating soot particles from an exhaust gas flow, wherein the exhaust gas flow is passed through extending in the longitudinal direction of a porous ceramic body channels of the ceramic body, wherein the exhaust gas flow through pores of the walls of the only one-sided open channels Passing ceramic body, and applied to parallel to the channels extending electrodes, a voltage to the ceramic body for generating an electric field in the channels of the ceramic body, which is oriented substantially normal to the axis of the channels. According to the invention, it is provided that prior to the introduction of the exhaust gas flow into the channels of the ceramic body, the soot particles are charged by means of a further electrode arrangement, and the voltage applied to the electrodes extending parallel to the channels is unipolar voltage pulses. wherein the unipolar voltage pulses are selected so that at a given channel height h in the field direction and given channel length L the Ratio of the generated by the voltage pulses drift velocity c of the charged soot particles in the field direction of the electric field generated in the channels to the flow velocity v of the gas flow in the channels greater than or equal to the ratio of twice the channel height h to the channel length L is.

The measures according to the invention therefore provide for the separation of the soot particles in the channels of the ceramic body not exclusively to be mechanical, but to use the electric field built up in the channels for the soot erosion also for the deposition of the soot particles, for which purpose a further electrode arrangement for previous charging of soot particles is provided. However, this measure alone is not yet sufficient to solve the problem of reduced creep strength described above. According to the invention is therefore also proposed to make the deposition of the soot particles in the wall-flow filter and the subsequent regeneration of the deposited soot through a unipolar plasma field on one wall side, while the second, opposite wall side always remains pure and therefore no clogging of pores, and so that no pressure build-up can take place. The pressure that tends to be higher as a result of the long-term halving of the filter surface must be taken into account when designing the filter size and the channel dimensions.

With this method according to the invention it is achieved that a wall of the channels remains completely pure, and opposes the gas penetrating it a consistently low resistance. Although the second wall receives faster higher Russbelag and faster irreversible pore closure by oil ash and other incombustible deposits, but the pressure loss through a completely clean filter wall is so much lower that the initially lower pressure drop of a normal wall-flow filter after a few 1000 km already caught up.

The flow velocity v of the gas flow in the channels and the drift velocity c, which depends on the diameter d of the soot particles, can easily be determined. The flow velocity v is measured anyway in the course of the regulation of the filter in order to regulate the applied voltage. The drift velocity c (d) of the soot particles of diameter d is again given by the mobility κ (d), the charge number z (d) and the field strength E perpendicular to the channel axis, the latter being given by the unipolar voltage pulses, so that for the Drift velocity c (d) for particles of diameter d gives: $$\mathrm{c}\left(\mathrm{d}\right)\mathrm{=}\mathrm{\kappa }\left(\mathrm{d}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)\mathrm{,}\mathrm{e}$$

The mobility κ (d) and the charge number z (d) can be read for different diameters d from well-known tables. The charge of the soot particles quantified by the charge number z (d) is due to the previous charge by means of the discharge electrode. The field strength E is given in a known manner directly by the applied voltage pulses. On the basis of the features according to the invention, it is thus possible to derive the field strength E necessary in the respective application, and thus the voltage to be applied to the electrodes. The setting according to the invention of the drift velocity c, the flow velocity v, the channel height h and the channel length L will be explained in more detail below.

Claim 2 gives a preferred choice for the duration of the unipolar voltage pulses and suggests that the unipolar voltage pulses have a duration below 20μs. Particularly advantageous are according to claim 3 unipolar voltage pulses with a duration between 5μs and 15μs. According to claim 4, the time interval between two unipolar voltage pulses between 50μs and 150μs.

Another difficulty in the approach to optimal separation efficiency results from the very different It is therefore proposed according to claim 5, a control in connection with the method according to the invention, in which the differential pressure of the exhaust gas stream is measured on the ceramic body, and above a predetermined value of the differential pressure, the control of the unipolar voltage pulses as a function of the flow velocity the exhaust gas flow takes place in the channels of the ceramic body, and below this predetermined value, the control only takes place as a function of the flow rate of the exhaust gas flow in the channels of the ceramic body, if the differential pressure has a tendency increasing over time, and otherwise the control independent of the differential pressure he follows. By means of this measure, as will be explained in more detail below, the soot distribution in the channel can be steered, and in particular also kept constant at different gas velocities, by at least temporarily reducing the field amplitude, preferably with short but very high amount of soot, while maintaining the field amplitude in FIG the remaining time directly proportional to the flow velocity in the channel controls.

• Fig. 1 eine schematische Darstellung eines Querschnitts eines Keramikkörpers zur Entfernung von Russpartikel aus einem Abgasstrom,
• Fig. 2 eine Detailansicht von Fig. 1, wobei insbesondere die Anordnung der Kanäle ersichtlich ist,
• Fig. 3 eine Darstellung zur Erläuterung der Geschwindigkeitskomponenten des Abgasstromes in einem Kanal des Keramikkörpers, und
• Fig. 4 eine Darstellung zur Erläuterung der Regelung der Spannungsimpulse im Zuge des erfindungsgemäßen Verfahrens.

The invention will be explained in more detail below with reference to the accompanying drawings. It show here the

• Fig. 1 a schematic representation of a cross section of a ceramic body for removing soot particles from an exhaust gas stream,
• Fig. 2 a detailed view of Fig. 1 in particular, the arrangement of the channels is evident,
• Fig. 3 a representation for explaining the velocity components of the exhaust gas flow in a channel of the ceramic body, and
• Fig. 4 a representation for explaining the regulation of the voltage pulses in the course of the inventive method.

The Fig. 1 and 2 each show a schematic representation of a cross section of a ceramic body 7, which is a honeycomb body. In each case, a ceramic body 7 is shown with a convex, namely elliptical circumferential line, but it could also have other cross-sectional shapes, such as a trapezoidal shape. The ceramic body 7 has channels 5, which extend in the longitudinal direction of the ceramic body 7, and are open at one end face of the ceramic body 7, and closed on the respective opposite side. Thus, the exhaust gas flow enters through a channel 5 which is open at the inlet side but closed at the outlet side and must be open to leave the ceramic body 7 through the internal wall of the respective channel 5 to the adjacent channel 5 which is closed at the inlet side but open at the outlet side is, step through.

According to the embodiment of the Fig. 1 and 2 For example, the electrodes 1, 2 are each formed by a group of electrode channels 4, in each of which an electrical coating 6 is introduced at least partially along its axial extent. As in particular from the Fig. 2 it can be seen, the groups of electrode channels 4 are each formed by adjacent electrode channels 4, so that a flat electrode surface 1,2 is defined by each group of electrode channels 4. However, other embodiments of the electrodes 1,2 are also possible.

Like from the Fig. 1 it can be seen, the flat electrode surfaces 1,2 each extend horizontally and parallel to each other. The distance between two adjacent electrode surfaces 1 and 2 is preferably less than 40 mm, about 15-25 mm. As a result, a homogeneous electric field can be ensured between the electrode surfaces 1 and 2, in particular in those regions of the room which are located within the each of two adjacent electrode surfaces 1,2 limited space region of the ceramic body 7 are located, which is referred to below as a homogeneous field region. The region 3 of the ceramic body 7 lying outside the homogeneous field region has in the embodiment according to FIG Fig. 1 and 2 over a denser structure, in order to additionally increase their structural load capacity.

Two adjacent electrode surfaces 1 and 2 are each contacted opposite polarity, wherein in the Fig. 1 approximately the electrode surface 1 is grounded, and the electrode surface 2 is supplied with bipolar voltage pulses.

wobei h die Kanalhöhe und t die Durchlaufzeit des Gases durch einen Kanal der Länge L ist: $$\mathrm{v}\mathrm{=}\mathrm{L}\mathrm{/}\mathrm{t}$$

The process according to the invention will now be explained in more detail below. First, the function of the filter assembly in a late equilibrium state is considered, as it will occur after about 50,000 km to 100,000 km, and in which a wall side by oil ash and phosphorus is already largely added. Assuming that all the exhaust gas flowed into the duct 5 has to escape through the second, unloaded side, the flow velocity v of the exhaust gas normal to the free wall of the duct will be the upper estimate of the velocity component v _x $${\mathrm{v}}_{\mathrm{x}}\mathrm{=}\mathrm{H}\mathrm{/}\mathrm{t}\mathrm{=}\mathrm{H}\mathrm{,}\mathrm{v}\mathrm{/}\mathrm{L}$$

where h is the channel height and t is the transit time of the gas through a channel of length L: $$\mathrm{v}\mathrm{=}\mathrm{L}\mathrm{/}\mathrm{t}$$

So that the soot is not entrained by the gas flow in the wrong direction, the drift velocity c (d) of the charged carbon particles with the diameter d in the field direction must be greater than or equal to the velocity component v _x against the field direction, ie $$\mathrm{c}\left(\mathrm{d}\right)\mathrm{\ge }{\mathrm{v}}_{\mathrm{x}}$$

With the mobility κ (d), the charge number z (d) and the field strength E _x perpendicular to the channel axis, the drift velocity c (d) results for the Rus particles of diameter d $$\mathrm{c}\left(\mathrm{d}\right)\mathrm{=}\mathrm{\kappa }\left(\mathrm{d}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)\mathrm{,}\mathrm{e}$$

zu $$\mathrm{k}\left(\mathrm{d}\right)\mathrm{.}\mathrm{z}\left(\mathrm{d}\right)\mathrm{.}\mathrm{E}\mathrm{\ge }\mathrm{h}\mathrm{.}\mathrm{v}\mathrm{/}\mathrm{L}$$

This results in the condition for the deposition of the Russteilchen on the right side with $${\mathrm{v}}_{\mathrm{x}}\mathrm{=}\mathrm{H}\mathrm{,}\mathrm{v}\mathrm{/}\mathrm{L}$$

to $$\mathrm{k}\left(\mathrm{d}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)\mathrm{,}\mathrm{e}\mathrm{\ge }\mathrm{H}\mathrm{,}\mathrm{v}\mathrm{/}\mathrm{L}$$

oder $$\mathrm{c}\left(\mathrm{d}\right)\mathrm{/}\mathrm{v}\mathrm{\ge }\mathrm{2.}\mathrm{h}\mathrm{/}\mathrm{L}$$

oder $$\mathrm{\kappa }\left(\mathrm{d}\right)\mathrm{.}\mathrm{z}\left(\mathrm{d}\right)\mathrm{.}\mathrm{E}\mathrm{\ge }2\mathrm{h}\mathrm{.}\mathrm{v}\mathrm{/}\mathrm{L}$$

da das Teilchen im ungünstigsten Fall den doppelten Weg im Gas zurücklegen muss, wenn es ganz in der Nähe der nicht Russ tragenden Wand einströmt.For the equal sign, as the drift velocity c (d) and the velocity component v _x cancel out, the soot particle practically does not move with respect to the distance to the channel walls. If, on the other hand, the particle is to reach the right wall, then the invention must apply $$\mathrm{c}\left(\mathrm{d}\right)\mathrm{\ge }\mathrm{2}{\mathrm{v}}_{\mathrm{x}}$$

or $$\mathrm{c}\left(\mathrm{d}\right)\mathrm{/}\mathrm{v}\mathrm{\ge }\mathrm{Second}\mathrm{H}\mathrm{/}\mathrm{L}$$

or $$\mathrm{\kappa }\left(\mathrm{d}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)\mathrm{,}\mathrm{e}\mathrm{\ge }2\mathrm{H}\mathrm{,}\mathrm{v}\mathrm{/}\mathrm{L}$$

because in the worst case, the particle has to travel the double path in the gas when it flows in close to the non-soot-bearing wall.

Since Russteilchen depending on their diameter d different mobilities κ (d) and allow different saturation charges z (d), resulting in the electric field E and different drift velocities c (d). Table 1 shows two for one DC field achievable mean electric field strengths (3kV / cm and 4 kV / cm) corresponding drift velocities c: Table 1: Particle diameter κ (d) / cm ² / Vs / (mobility) z (d) (loading number) κ (d) .z (d) κ (d) .z (d) .E (E = 3 kV / cm) κ (d) .z (d) .E (E = 4 kV / cm) 10 nm 6.3. 10 ^-3 0.3 1.9. 10 ^-3 5.7 cm / s 7.6 cm / s 20 nm 1.3. 10 ^-3 1 1.3. 10 ^-3 3.9 cm / s 5.2 cm / s 100 nm 1.0. 10 ^-4 10 1.0. 10 ^-3 3.0 cm / s 4.0 cm / s 1 μm 5.4. 10 ^-6 300 1.6. 10 ^-3 4.8 cm / s 6.4 cm / s 10 μm 5.4. 10 ^-7 11,000 6.0. 10 ^-3 18.0 cm / s 24.0 cm / s

Table 2 shows the necessary for the present conditions channel heights h and average flow times t of the exhaust gas through the channels 5 of the unipolar wall-flow filter: Table 2: Particle diameter κ (d) .z (d) .E (3 kV / cm) h = 0.8 mm t ≥ h = 0.6 mm t ≥ κ (d) .z (d) .E (4 kV / cm) h = 0.8 mm t ≥ h = 0.6mm t ≥ 10 nm 5.7 cm / s 28 ms 21 ms 7.6 cm / s 21 ms 16 ms 20 nm 3.9 cm / s 41 ms 31 ms 5.2 cm / s 31 ms 23 ms 100 nm 3.0 cm / s 53 ms 40 ms 4.0 cm / s 40 ms 30 ms 1 μm 4.8 cm / s 33 ms 25 ms 6.4 cm / s 25 ms 19 ms 10 μm 18.0 cm / s 9 ms 7 ms 24.0 cm / s 7 ms 5 ms

With a throughput time of t = 20 ms, exhaust gas quantities of 400 kg / h at 550 ° C with acceptable filter cross sections can fulfill these requirements. This exhaust gas mass flow of 400 kg / h corresponds to a full load operation of a supercharged diesel engine of the lower middle class. A deposition field E of 4 kV / cm can thus at a channel height h of 0.6 mm, even at full load all particles except for Size range around 100nm along a separation wall. Since particles with a diameter of 100 nm no longer contribute to the number of particles and still nothing to the particle mass, filtering with these parameters is perfectly acceptable, since regeneration under plasma proceeds continuously and rapidly at these temperatures.

Thus it can be seen that particles with a diameter of about 100 nm reach the lowest drift velocity c. For these, necessary flow-through times t are given in Table 3 for some channel heights h, so that these particles can reach the correct channel wall: Table 3: Channel height (mm) 3 kV / cm 4 kV / cm 0.6 40 ms 30 ms 0.8 53 ms 40 ms 1.0 66 ms 50 ms

The difficulty in operating a soot filter is always in the low load range and is particularly critical when the deposition of the filter must be designed for high engine performance. The higher motorized mid-range cars differ between full load and city operation in the gas flow by about a factor of 10. If the filter is now designed so that all particles at full load over the filter length L (usually 20 cm to 25 cm) are deposited, so this deposition takes place in City operation at the first 20 mm to 25 mm instead. Thus, and especially at low engine temperatures, the regeneration of the filter can be locally overwhelmed. The situation becomes particularly critical when the N0x emission is exchanged for more soot emission.

abgeleitet werden, indem im Fall der Gültigkeit des Gleichheitszeichen die mittlere Gasgeschwindigkeit v in den Kanälen 5 ersetzt, wobei q der freie Gesamtquerschnitt aller Kanäle 5 darstellt $$\mathrm{v}\mathrm{=}\mathrm{V}\mathrm{/}\mathrm{q}$$

V ist dabei der Gasvolumenstrom. Dadurch ergibt sich die Regelgleichung für die Abscheidefeldstärke E als Funktion des Gasvolumenstroms V $$\mathrm{E}\mathrm{=}\mathrm{\beta }\mathrm{.}\mathrm{V}$$

mit der Regelkonstanten β $$\mathrm{\beta }\mathrm{=}\mathrm{2}\mathrm{h}\mathrm{/}\mathrm{q}\mathrm{.}\mathrm{L}\mathrm{.}\mathrm{\kappa }\left(\mathrm{d}\right)\mathrm{.}\mathrm{z}\left(\mathrm{d}\right)$$

The separation field can therefore be controlled proportionally to the gas volume flow, so that the soot deposition always extends approximately over the entire filter length L. The corresponding rule equation can be derived from the relationship $$\mathrm{\kappa }\left(\mathrm{d}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)\mathrm{,}\mathrm{e}\mathrm{\ge }2\mathrm{H}\mathrm{,}\mathrm{v}\mathrm{/}\mathrm{L}$$

are derived by substituting, in the case of the validity of the equal sign, the mean gas velocity v in the channels 5, where q represents the total free cross section of all the channels 5 $$\mathrm{v}\mathrm{=}\mathrm{V}\mathrm{/}\mathrm{q}$$

V is the gas volume flow. This results in the control equation for the Abscheidefeldstärke E as a function of the gas flow rate V $$\mathrm{e}\mathrm{=}\mathrm{\beta }\mathrm{,}\mathrm{V}$$

with the control constant β $$\mathrm{\beta }\mathrm{=}\mathrm{2}\mathrm{H}\mathrm{/}\mathrm{q}\mathrm{,}\mathrm{L}\mathrm{,}\mathrm{\kappa }\left(\mathrm{d}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)$$

If the temperature dependence of the mobility κ (d, T) and of the gas volume flow V is also taken into account, then the equation results for the control equation $$\mathrm{e}\left(\mathrm{d}\mathrm{T}\right)\mathrm{=}\mathrm{V}\left(\mathrm{d}\mathrm{T}\right)\mathrm{.2}\mathrm{H}\mathrm{/}\mathrm{q}\mathrm{,}\mathrm{L}\mathrm{,}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)$$

For d, the diameter can be taken with the smallest product of κ (d) .z (d), or those whose soot particles are to be deposited just within the filter length L.

If now the deposition field E with the gas volume flow V and the temperature correction V (T) / κ (d, T) is controlled, there is the further difficulty that with decreasing Abscheidefeldstärke E, the plasma temperature and thus the regeneration rate is lower, and ultimately the conversion of the soot completely disappears. At low temperatures and high soot accumulation, therefore, the deposition field E must not be regulated below a predetermined limit. This problem is particularly serious in short Full load accelerations with a cold engine that is cold for longer partial load operation.

According to the invention, this problem can be solved by setting a correspondingly high deposition field E with a high amount of soot, but switching between high field and low field depending on the regeneration state of the filter with low sootfall, wherein the data "high sootfall" and "Low soot" and gas mass flow from the engine computer via a CAN bus to the processor of the filter to be transferred, or the change is controlled directly from the engine computer, which then calculates soot amount and soot distribution in the filter itself.

• Intervall I:

  v > v₁

  Die Regelgleichung lautet: $$\mathrm{E}\left(\mathrm{T}\right)\mathrm{=}\mathrm{h}\mathrm{.}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{.}\mathrm{L}\mathrm{.}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{.}\mathrm{z}\left(\mathrm{d}\right)$$

  

• Intervall II:

  v < v₁

  Die Regelgleichung lautet bei dv/dt ≥ 0 ebenfalls: $$\mathrm{E}\left(\mathrm{T}\right)\mathrm{=}\mathrm{h}\mathrm{.}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{.}\mathrm{L}\mathrm{.}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{.}\mathrm{z}\left(\mathrm{d}\right)$$

  

  v < v₁

  Die Regelgleichung lautet bei dv/dt < 0 dagegen: $$\mathrm{E}\left(\mathrm{T}\right)\mathrm{=}{\mathrm{E}}_{\max }$$

  

If v is the average flow velocity of the exhaust gas in the channels 5 of the honeycomb body 7, the control principle according to the invention results from dividing the velocity v into at least two intervals, where the normal flow equation applies to the larger flow velocity, whereas for the lower flow velocity only the normal equation applies if the derivative of the velocity after time is positive or nearly zero, ie the velocity increases with time or remains constant. If, on the other hand, the derivative of the velocity over time is negative, ie if the flow velocity decreases with time, the field E is driven with the highest value E _max permitted under the constraints of plasma flow and temperature:

• Interval I:

  v> v ₁

  The rule equation is: $$\mathrm{e}\left(\mathrm{T}\right)\mathrm{=}\mathrm{H}\mathrm{,}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{,}\mathrm{L}\mathrm{,}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)$$

  

• Interval II:

  v <v ₁

  The rule equation is also at dv / dt ≥ 0 : $$\mathrm{e}\left(\mathrm{T}\right)\mathrm{=}\mathrm{H}\mathrm{,}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{,}\mathrm{L}\mathrm{,}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)$$

  

  v <v ₁

  The rule equation is at dv / dt <0, however: $$\mathrm{e}\left(\mathrm{T}\right)\mathrm{=}{\mathrm{e}}_{\mathrm{Max}}$$

  

Of course, the processor of the soot filter can also perform this control itself. The simplest reaction is obtained by measuring the differential pressure preferably occurring at the filter itself (p ₁ -p ₂ ), which in a good approximation is directly proportional to the mean flow velocity v in the channels 5 of the honeycomb body 7 and its increase or decrease is good Conclusion on the soot emission of the engine allowed.

• Intervall I:

  (p₁ - p₂) > Δp₁

  Die Regelgleichung lautet: $$\mathrm{E}\left(\mathrm{T}\right)\mathrm{=}\mathrm{h}\mathrm{.}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{.}\mathrm{L}\mathrm{.}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{.}\mathrm{z}\left(\mathrm{d}\right)$$

  

• Intervall II:

  (p₁ - p₂) < Δp₁

  Die Regelgleichung lautet bei
  d(p₁ - p₂) /dt ≥ 0 ebenfalls $$\mathrm{E}\left(\mathrm{T}\right)\mathrm{=}\mathrm{h}\mathrm{.}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{.}\mathrm{L}\mathrm{.}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{.}\mathrm{z}\left(\mathrm{d}\right)$$

  

  (p₁ - p₂) < Δp₁

  Die Regelgleichung lautet bei
  d(p₁ - p₂) /dt < 0 dagegen $$\mathrm{E}\left(\mathrm{T}\right)\mathrm{=}{\mathrm{E}}_{\max }$$

  

This results in the conversion of the control principle according to the invention by dividing the differential pressure (p ₁ -p ₂ ) occurring at the filter or another corresponding flow resistance into at least two intervals, where the normal control equation applies to the larger differential pressure, while for the lower differential pressure (p ₁ - p ₂ ) only the normal control equation applies if the derivative of the differential pressure after the time is positive or nearly zero, ie the differential pressure increases with time or remains constant. If, on the other hand, the derivative of the differential pressure after the time is negative, ie if the differential pressure decreases with time, the field E is driven with the highest value E _max permitted under the secondary conditions of the plasma flow and the gas temperature:

• Interval I:

  (p ₁ -p ₂ )> Δp ₁

  The rule equation is: $$\mathrm{e}\left(\mathrm{T}\right)\mathrm{=}\mathrm{H}\mathrm{,}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{,}\mathrm{L}\mathrm{,}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)$$

  

• Interval II:

  (p ₁ -p ₂ ) <Δp ₁

  The rule equation is at
  d (p ₁ -p ₂ ) / dt ≥ 0 also $$\mathrm{e}\left(\mathrm{T}\right)\mathrm{=}\mathrm{H}\mathrm{,}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{,}\mathrm{L}\mathrm{,}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)$$

  

  (p ₁ -p ₂ ) <Δp ₁

  The rule equation is at
  d (p ₁ - p ₂ ) / dt <0 $$\mathrm{e}\left(\mathrm{T}\right)\mathrm{=}{\mathrm{e}}_{\mathrm{Max}}$$

  

A very expedient and efficient implementation of these control principles can be carried out according to the invention so that the "controlled variable" (p ₁ -p ₂ ), by which the "control value" E (T) is to be set accordingly, together with the "control value" E (FIG. T) is read as at least two-column "rule table" in the processor and processed by him. This approach requires the least amount of computing power and develops the highest control speed to match voltage and field in real time to engine emissions. If implied dependencies of further variables are also to be taken into account, it may be particularly advantageous to read into the table at least one second controlled variable, in this case the temperature, thereby creating the possibility of adding a further control level for very different temperatures in addition to the pressure difference to get. Depending on the fineness of the subdivision for this second controlled variable, we obtain further columns in which the processor always finds the corresponding manipulated variable. The rules table could look like this: Table 4 Control value 1 Control value 2 Control value 1 for T <T ₁ Control value 2 T ₂ <T <T ₁ Control value 3 for T ₂ <T (p ₁ -p ₂ ) _n T _n E _n1 E _n2 E _n3 (p ₁ -p ₂ ) _{n + 1} Tn _{+ 1} E _{n1 + 1} E _{n2 + 1} E _{n3 + 1} , , , , , , , , , , , , , , ,

wobei die fehlende Normierung für den Regelvorgang unerheblich ist. Somit ergibt sich bei $$\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{p}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}}\mathrm{-}\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{p}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}\mathrm{-}\mathrm{1}}\ge 0\mathrm{ebenfalls}$$

$$\mathrm{E}\left(\mathrm{T}\right)\mathrm{=}\mathrm{h}\mathrm{.}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{.}\mathrm{L}\mathrm{.}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{.}\mathrm{z}\left(\mathrm{d}\right)$$

und mit $$\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{p}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}}\mathrm{-}\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{p}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}\mathrm{-}\mathrm{1}}<0\mathrm{dagegen}$$

$$\mathrm{E}\left(\mathrm{T}\right)\mathrm{=}{\mathrm{E}}_{\max }$$

Similarly, the processor proceeds to find the derivative of the control value by time by replacing the differential quotient d (p ₁ - p ₂ ) / dt by the difference between two successive incoming control values, ie $$\mathrm{d}\left({\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="imgf0001.png"\; state="\{\{state\}\}">p</figure-callout>}}_{\mathrm{2}}\right)\mathrm{/}\mathrm{dt}\mathrm{=}\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="imgf0001.png"\; state="\{\{state\}\}">p</figure-callout>}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}}\mathrm{-}\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="imgf0001.png"\; state="\{\{state\}\}">p</figure-callout>}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}\mathrm{-}\mathrm{1}}$$

where the missing normalization is irrelevant for the control process. Thus arises at $$\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="imgf0001.png"\; state="\{\{state\}\}">p</figure-callout>}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}}\mathrm{-}\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="imgf0001.png"\; state="\{\{state\}\}">p</figure-callout>}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}\mathrm{-}\mathrm{1}}\ge 0\mathrm{also}$$

$$\mathrm{e}\left(\mathrm{T}\right)\mathrm{=}\mathrm{H}\mathrm{,}\mathrm{V}\left(\mathrm{T}\right)\mathrm{/}\mathrm{q}\mathrm{,}\mathrm{L}\mathrm{,}\mathrm{\kappa }\left(\mathrm{d}\mathrm{T}\right)\mathrm{,}\mathrm{z}\left(\mathrm{d}\right)$$

and with $$\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="imgf0001.png"\; state="\{\{state\}\}">p</figure-callout>}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}}\mathrm{-}\mathrm{(}{\mathrm{p}}_{\mathrm{1}}\mathrm{-}{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="imgf0001.png"\; state="\{\{state\}\}">p</figure-callout>}}_{\mathrm{2}}{\mathrm{)}}_{\mathrm{n}\mathrm{-}\mathrm{1}}<0\mathrm{on\; the\; other\; hand}$$

$$\mathrm{e}\left(\mathrm{T}\right)\mathrm{=}{\mathrm{e}}_{\mathrm{Max}}$$

Of course, according to the invention, the filter processor itself can calculate and carry out the voltage control necessary for the optimum separation of the soot, if the motor processor has the necessary signals, preferably temperature, gas volume flow or mass flow, preferably also EGR rate and injection quantity, via a signal system, preferably via CANBUS. provides. In this case, there is also a further advantage in that the smaller filter processor determines the necessary voltage changes more quickly, and thereby can adjust the deposition field E, which is only slowly controllable by capacitances on the high-voltage side, in time for the new soot accumulation.

After the plasma in the filter anyway by a microprocessor as a function of temperature, residual oxygen and preferably humidity and soot amount is regulated, it is advantageous if this processor also calculates the amount and distribution of the deposited soot, and then controls its deposition field E and in particular the temporal distribution between deposition field strength and regeneration field strength. Although the information can be determined by the filter's own filter as well as by the filter's own analysis of the current-voltage characteristics, it is particularly advantageous if from the engine control processor, preferably via a CANBUS, corresponding data such as exhaust gas mass flow, injection quantity, EGR rate, residual oxygen and Russemission be fed to the filter processor.

This inventive method for the operation of the filter in the low load range can be further modified according to the invention by the necessary for its implementation device consists in a two-flow design of the filter, wherein in the lower power range, only a partial filter is acted upon with the exhaust gas. With repeated high, but short soot attack with a cold engine (so-called "housewife cycle") can preferably be switched back and forth between the two sub-filters to regenerate the soot in the off system at higher field, and in the acted upon with the exhaust stream sub-filter at preferably lower field over the entire filter to collect.

With the method according to the invention, the creep rupture strength of filter arrangements based on the use of a ceramic body with channels closed on one side can thus be decisively increased.

Claims (5)

Method for operating a filter arrangement for separating soot particles from an exhaust gas stream, in which the exhaust gas stream is passed through channels (5) of the ceramic body (7) running in the longitudinal direction of a porous ceramic body (7), the exhaust gas flow through pores of the walls being only one-sided open channels (5) of the ceramic body (7) passes, and at parallel to the channels (5) extending electrodes (1,2) a voltage to the ceramic body (7) for generating an electric field (E) in the channels (5) of the ceramic body (7) which is oriented substantially normal to the axis of the channels (5), characterized in that prior to introducing the exhaust gas flow into the channels (5) of the ceramic body (7), charging of the soot particles by means of another Electrode arrangement is carried out, and at the voltage which is applied to the parallel to the channels (5) extending electrodes (1,2) to han unipolar voltage pulses delt, wherein the unipolar voltage pulses are selected so that at a given channel height (h) in the field direction and given channel length (L) the ratio of the drift velocity (c) generated by the voltage pulses of the charged soot particles in the field direction of in the channels (5) generated electric field to the flow velocity (v) of the gas flow in the channels (5) is greater than or equal to the ratio of twice the channel height (h) to the channel length (L). A method according to claim 1, characterized in that the unipolar voltage pulses have a duration of less than 20μs. A method according to claim 2, characterized in that the unipolar voltage pulses have a duration between 5μs and 15μs. Method according to one of claims 1 to 3, characterized in that the time interval between two unipolar voltage pulses is between 50μs and 150μs. Method according to one of claims 1 to 4, characterized in that the differential pressure of the exhaust gas stream on the ceramic body (7) is measured, and above a predetermined value (Δp ₁ ) of the differential pressure, the control of the unipolar voltage pulses in dependence on the flow velocity (v) of the Exhaust gas flow in the channels (5) of the ceramic body (7), and below this predetermined value (Δp ₁ ), the control only in response to the flow rate of the exhaust gas flow in the channels (5) of the ceramic body (7), if the differential pressure (Δp) has a tendency to increase with time, and otherwise the control is independent of the differential pressure (Δp).

Images