Search All Patents:

---

Register or Login To Download This Patent As A PDF

---

United States Patent Application	20110138779
Kind Code	A1
NEUMAYER; Roland	June 16, 2011

---

---

---

Foreign Application Data

---

Date	Code	Application Number
Dec 12, 2009	DE	10 2009 058 089.1

---

Claims

---

1. A measurement method for use in operation of an SCR exhaust gas aftertreatment system of a vehicle, the SCR exhaust gas aftertreatment system having an SCR catalytic converter, a urea introducing device, a first NO.sub.x sensor arranged in an exhaust gas flow upstream of the SCR catalytic converter and of the urea introducing device, and a second NO.sub.x sensor arranged in one of the SCR catalytic converter and the exhaust gas flow downstream of the SCR catalytic converter, the measuring method comprising the acts of: determining a first signal by the first NO.sub.x sensor; determining a second signal by the second NO.sub.x sensor; and using the first and second signals to determine a linearity indication measuring a linear correlation between the first and second signals.

2. The method according to claim 1, wherein a correlation coefficient of the first and second signals is determined as the linearity indication.

3. The method according to claim 1, wherein the first signal is a signal that is time delayed with respect to a sensor signal of the first NO.sub.x sensor.

4. The method according to claim 3, wherein the time delay corresponds approximately to a time delay of the exhaust gas flow between a position of the first NO.sub.x sensor and a position of the second NO.sub.x sensor.

5. The method according to claim 4, further comprising the act of: determining the time delay as a function of an exhaust gas volume flow.

6. The method according to claim 1, further comprising the act of: evaluating a sensor signal of the second NO.sub.x sensor as a function of the linearity indication to provide one of an NO.sub.x indication and an NH3 indication.

7. The method according to claim 2, further comprising the acts of: differentiating between an NO.sub.x slip and an NH3 slip as a function of the correlation coefficient, wherein a correlation coefficient close to 1 is indicative of an NO.sub.x slip and a correlation coefficient close to 0 is indicative of an NH3 slip.

8. The method according to claim 1, further comprising the act of: regulating introduction of urea into the SCR exhaust gas aftertreatment system in accordance with use of the linearity indication.

9. The method according to claim 7, further comprising the act of: regulating introduction of urea into the SCR exhaust gas aftertreatment system in accordance with use of the linearity indication.

10. The method according to claim 8, wherein the regulating act evaluates a sensor signal of the second NO.sub.x sensor using the linearity indication.

11. The method according to claim 8, wherein the linearity indication is used for at least one of determining an NH3 actual fill level and determining an NH3 desired fill level.

12. The method according to claim 9, wherein the linearity indication is used for at least one of determining an NH3 actual fill level and determining an NH3 desired fill level.

13. A device for use in an SCR exhaust gas aftertreatment system of a vehicle, the exhaust gas aftertreatment system having an SCR catalytic converter and a urea introducing device arranged upstream of the SCR catalytic converter, the device comprising: a first NO.sub.x sensor arrangable in an exhaust gas flow upstream of the SCR catalytic converter and of the urea introducing device; a second NO.sub.x sensor arrangable either in the SCR catalytic converter or in the exhaust gas flow downstream of the SCR catalytic converter; a linearity indication determining unit receiving first and second signals related to outputs of the first and second NO.sub.x sensors, the linearity indication determining unit providing a linearity indication based on the first and second signals, the linearity indication being a measure for a linear correlation between the first and second signals.

14. The device according to claim 13, wherein the linearity indication determining unit is operatively configured for determining a correlation coefficient of the first and second signals.

15. The device according to claim 13, further comprising: a delay element for determining the first signal by delaying a sensor signal output of the first NO.sub.x sensor, a delay time corresponding to approximately the delay time of the exhaust gas flow between a position where the first NO.sub.x sensor is arrangable and a position where the second NO.sub.x sensor is arrangable with respect to the exhaust gas aftertreatment system.

16. The device according to claim 14, further comprising: a delay element for determining the first signal by delaying a sensor signal output of the first NO.sub.x sensor, a delay time corresponding to approximately the delay time of the exhaust gas flow between a position where the first NO.sub.x sensor is arrangable and a position where the second NO.sub.x sensor is arrangable with respect to the exhaust gas aftertreatment system.

17. An SCR exhaust gas aftertreatment system for a vehicle, comprising: an SCR catalytic converter; a urea introducing device arranged upstream of the SCR catalytic converter; a first NO.sub.x sensor arranged in an exhaust gas flow upstream of the SCR catalytic converter and of the urea introducing device; a second NO.sub.x sensor arranged either in the SCR catalytic converter or in the exhaust gas flow downstream of the SCR catalytic converter; a linearity indication determining unit receiving first and second signals related to outputs of the first and second NO.sub.x sensors, the linearity indication determining unit providing a linearity indication based on the first and second signals, the linearity indication being a measure for a linear correlation between the first and second signals.

18. The SCR exhaust gas aftertreatment system according to claim 17, further comprising: a closed loop control circuit for regulating the urea introduction, wherein the closed loop control system includes the linearity indication determining unit, the closed loop control circuit being operatively configured to regulate the urea introduction with simultaneous use of the linearity indication.

---

Description

---

CROSS-REFERENCE TO RELATED APPLICATION

[0001] This application claims priority under 35 U.S.C. .sctn.119 from German Patent Application No. DE 10 2009 058 089.1-13, filed Dec. 12, 2009, the entire disclosure of which is herein expressly incorporated by reference.

BACKGROUND AND SUMMARY OF THE INVENTION

[0002] The invention relates to SCR exhaust gas aftertreatment systems (SCR--selective catalytic reduction).

[0003] A method for reducing nitrogen oxide (NO.sub.x) emissions in diesel engines of motor vehicles is the so-called SCR method--that is, the selective catalytic reduction of nitrogen oxides. In order to bring about the reaction, ammonia (NH3) is used. The products of the reaction are water (H.sub.2O) and nitrogen (N.sub.2). The ammonia that is used for the SCR reaction is introduced in the form of an aqueous urea solution (typically 32.5% urea) into the exhaust gas system upstream of the SCR catalytic converter--for example, injected by use of a metering pump or injector. This solution of urea and water yields ammonia and CO.sub.2 through a hydrolysis reaction. The ammonia reacts with the nitrogen oxides in the exhaust gas in a special SCR catalytic converter.

[0004] SCR catalytic converters can store only a certain amount of NH3 as a function of their size. The urea metering should correspond on average to the urea required to reduce the nitrogen oxide emissions. At the same time it must be noted that the nitrogen oxide emissions of the engine are a function of the respective speed and the respective torque of the engine, so that the urea metering should be adjusted to match. If the urea metering is too low, then the result is a decline in the effectiveness of the nitrogen oxide reduction. This state is also referred to as the NO.sub.x slip--that is, the SCR catalytic converter allows too much nitrogen oxide to pass through. If, however, the urea metering is too high, then the resulting ammonia does not react with the nitrogen oxide in an adequate amount owing to the oversupply of ammonia. In this case ammonia can pass into the environment, a state that can lead to a perceptible odor. In this case the phenomenon is also known as the NH3 slip.

[0005] FIG. 1 shows the main components of a conventional SCR exhaust gas aftertreatment system 5 for use in a motor vehicle. The SCR exhaust gas aftertreatment system 5 includes a first NO.sub.x sensor 1 in the exhaust gas flow upstream of a urea metering device 3 and of an SCR catalytic converter 4. The first NO.sub.x sensor 1 is used to measure the nitrogen oxide emissions of the engine. A second NO.sub.x sensor 2 is arranged in the exhaust gas flow downstream of the SCR catalytic converter 4. This second NO.sub.x sensor 2 can also be arranged in the SCR catalytic converter 4 itself (not illustrated). The second NO.sub.x sensor 2 measures both the NO.sub.x slip and also the NH3 slip.

[0006] The current NO.sub.x sensors that are used in vehicles cannot differentiate between NO.sub.x and NH3. That is, NO.sub.x sensors have a so-called NH3 cross sensitivity. For this reason it is not possible to distinguish directly between an NH3 slip and an NO.sub.x slip from the sensor signal, measured downstream of the SCR catalytic converter 4. This drawback represents a major deficiency for correctly regulating the urea metering.

[0007] The document WO 2009/036780 A1 discloses an SCR catalytic converter with an NH3 fill level monitoring system that determines the NH3 fill level in two different ways by use of two NO.sub.x sensors, where the respective errors--for example, clue to the cross sensitivity of the second sensor to ammonia--are at least partially compensated.

[0008] The object of the invention is to provide a measurement method, which makes it possible to differentiate between the NO.sub.x slip and the NH3 slip despite the cross sensitivity of the second NO.sub.x sensor to NH3--that is, makes it possible to differentiate whether the substance detected by the second NO.sub.x sensor is NO.sub.x or NH3. Furthermore, the object of the invention is to provide a corresponding device. Furthermore, the object of the invention is to provide a regulating method that is used by this measurement method for regulating the urea metering.

[0009] A first aspect of the invention relates to a measurement method for an SCR exhaust gas aftertreatment system of a vehicle. Such an SCR exhaust gas aftertreatment system includes an SCR catalytic converter, a urea introducing device, a first NO.sub.x sensor, which is arranged in the exhaust gas flow upstream of the SCR catalytic converter and of the urea introducing device, and a second NO.sub.x sensor, which is arranged in the SCR catalytic converter or in the exhaust gas flow downstream of the SCR catalytic converter.

[0010] According to the method, a first signal is determined by the first NO.sub.x sensor. This first signal can also be a time delayed signal, which will be discussed in more detail below in the description. In addition, a second signal is determined by the second NO.sub.x sensor. Based on the first and the second signal, a linearity indication is determined that is a measure for the linear correlation between both signals.

[0011] The linearity indication can be used to differentiate between the NO.sub.x slip and the NH3 slip, which will be discussed in more detail below. The first signal, which is determined by the first NO.sub.x sensor, indicates the NO.sub.x emission of the engine before the urea injection. The second signal, determined by the second NO.sub.x sensor, indicates both the NO.sub.x emissions downstream of the SCR catalytic converter and also the NH3 emissions downstream of the SCR catalytic converter. For the NO.sub.x emissions, measured by the second NO.sub.x sensor, there is a high degree of linear correlation to the first signal, measured by the first NO.sub.x sensor. That is, in this case there exists a high correlation between the first and the second signal. In this case the correlation can be increased, if the first signal and the second signal are time synchronized with respect to each other by suitably delaying the sensor signal of the first sensor. If, in contrast, primarily NH3 is present at the second NO.sub.x sensor, then the correlation between the two signals is low.

[0012] Preferably a correlation coefficient between the two signals is calculated as the linearity indication. By determining the correlation coefficient of the two sensor signals it is possible to obtain a dimensionless measure for the degree of linear correlation between the two signals. It is assumed in the following that the amount of the correlation coefficient can be a maximum value of +1, but it is not mandatory within the scope of the invention that the correlation coefficient has to be normalized to a maximum value of 1. If the correlation coefficient has a value of +1, then there exists a totally positive linear correlation between the two signals. If the correlation coefficient has a value of 0, then the two features are not at all linearly dependent on each other.

[0013] Therefore, an NO.sub.x slip can be inferred from a correlation coefficient exhibiting a value close to +1; and an NH3 slip can be inferred from a correlation coefficient exhibiting a value close to 0. Therefore, despite the cross sensitivity of the second NO.sub.x sensor to NH3, it is possible with this method to distinguish between an NO.sub.x slip and an NH3 slip.

[0014] Preferably in order to obtain the first signal, the sensor signal of the first NO.sub.x sensor is time delayed, in order to at least partially compensate for the time delay of the exhaust gas flow (that is, the NO.sub.x emissions) between the position of the first NO.sub.x sensor and the position of the second NO.sub.x sensor. This time delay is equivalent to approximately a dead time and would falsify the results, if the time delay were not considered in the calculation. Hence, the time synchronization increases even more the degree of linearity in the case of an NO.sub.x slip. In order to totally compensate in essence for the influence of the time delay, the time delay of the sensor signal of the first NO.sub.x sensor is selected preferably in such a way that this time delay corresponds approximately to the time delay of the exhaust gas flow between the position of the first NO.sub.x sensor and the position of the second NO.sub.x sensor.

[0015] The time delay can also be implemented inherently in that a time offset is considered in the course of determining the correlation coefficient.

[0016] The method provides preferably that this time delay is calculated as a function of the respective exhaust gas volume flow rate. For example, it can be provided that the method calculates the time delay of the exhaust gas flow (that is, the NO.sub.x emissions) between the position of the first NO.sub.x sensor and the position of the second NO.sub.x sensor from the exhaust gas volume flow rate and the volume of the exhaust gas system between the two NO.sub.x sensors. The sensor signal of the first NO.sub.x sensor can be time delayed as a function of the calculated time delay in each case.

[0017] Another aspect of the invention relates to a regulating method for regulating the introduction of urea into an SCR exhaust gas aftertreatment system. In this case the urea introduction is regulated with the simultaneous use of the above described linearity indication. For example, the linearity indication can be used as the regulating variable. Preferably, however, the regulating procedure occurs in response to a sensor signal, which is output by the second sensor and evaluated or checked for plausibility with the linearity indication (for example, the correlation factor).

[0018] Another aspect of the invention relates to a device for determining a linearity indication for an above described SCR exhaust gas aftertreatment system having two NO.sub.x sensors in a vehicle. In this case the device includes first means for determining a linearity indication (for example, a correlation coefficient) based on a first signal and a second signal, which is a measure for the linear correlation between both signals. In this respect the first signal has been determined by the first NO.sub.x sensor; and the second signal has been determined by the second NO.sub.x sensor.

[0019] The above embodiments of the measurement method according to the invention can also be transferred in an identical way to the device for determining the linearity indication.

[0020] Preferably the device includes a delay element for the time delay of a sensor signal of the first NO.sub.x sensor, where the time delayed signal is used as the first signal by the means for determining the linearity indication. In this respect the time delay corresponds preferably to approximately the time delay of the exhaust gas flow between the position of the first NO.sub.x sensor and the position of the second NO.sub.x sensor.

[0021] Furthermore, the device includes preferably means for determining the time delay, for example, as a function of the respective exhaust gas flow, as discussed above with reference to the method according to the invention.

[0022] An additional aspect of the invention relates to an SCR exhaust gas aftertreatment system for a vehicle. This exhaust gas aftertreatment system includes, besides the aforementioned components of a conventional SCR exhaust gas aftertreatment system having two NO.sub.x sensors, also the above described device for determining a linearity indication and, in particular, also the above described regulating device.

[0023] Other objects, advantages and novel features of the present invention will become apparent from the following detailed description of one or more preferred embodiments when considered in conjunction with the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

[0024] FIG. 1 depicts a conventional SCR exhaust gas aftertreatment system 5;

[0025] FIG. 2 is a schematic block diagram depicting an embodiment for determining the correlation coefficient;

[0026] FIG. 3 is a graph depicting a first exemplary curve of the time shifted sensor signal 18 (in ppm) of the first NO.sub.x sensor 1 and of the sensor signal 12 (in ppm) of the second NO.sub.x sensor 2, as well as the curve of the resulting correlation coefficient signal 24;

[0027] FIG. 4 is a graph depicting a second exemplary curve of the time shifted sensor signal 18 (in ppm) of the first NO.sub.x sensor 1 and of the sensor signal 12 (in ppm) of the second NO.sub.x sensor 2, as well as the curve of the resulting correlation coefficient signal 24; and

[0028] FIG. 5 is a schematic block diagram depicting an embodiment for regulating the NH3 fill level.

DETAILED DESCRIPTION OF THE DRAWINGS

[0029] FIG. 1 shows a conventional SCR exhaust gas aftertreatment system 5 that has already been described above in the background section of the specification. Such an SCR exhaust gas aftertreatment system 5 has the drawback that owing to cross sensitivity of the second NO.sub.x sensor 2 to NH3, it is not easy to differentiate between the presence of an NH3 slip and an NO.sub.x slip. As explained above, in the case of an NO.sub.x slip there exists a linear correlation between the measured NO.sub.x emissions at the second NO.sub.x sensor 2 and the time synchronized signal of the first NO.sub.x sensor 1. If, in contrast, in the case of an NH3 slip, the NH3 is measured at the second NO.sub.x sensor 2, then the correlation between the two signals is low.

[0030] The invention utilizes this linear correlation between the signals. Preferably to this end the correlation coefficient of the signals of the two NO.sub.x sensors 1 and 2 is calculated. The result is the degree of linear correlation between the two signals. In the case of a value of +1, there exists a totally positive linear correlation between the signals. If the correlation coefficient has a value of 0, then the two signals are not at all linearly dependent on each other. Therefore, an NO.sub.x slip can be inferred from a correlation coefficient exhibiting a value close to 1; and an NH3 slip can be inferred from a correlation coefficient exhibiting a value close to 0. In this way the drawback of the cross sensitivity of the second NO.sub.x sensor 2 to NH3 is eliminated.

[0031] FIG. 2 depicts an exemplary embodiment in block diagram form for determining the correlation coefficient. In order to determine the correlation coefficient, the sensor signal 11 of the first sensor 3 and the sensor signal 12 of the second sensor 2 are used. In block 13 the signal of the exhaust gas volume flow rate (in m.sup.3/h) and the volume of the exhaust gas system between the two NO.sub.x sensors 1 and 2 are used to calculate the time delay 15 of the NO.sub.x emissions from the position of the NO.sub.x sensor 3 to the position of the NO.sub.x sensor 2. As a function of this calculated time delay 15, the sensor signal 11 in the block 16 is time delayed. In block 17 it is determined by use of the status signals 21 and 22 of the first NO.sub.x sensor 1 and/or of the second NO.sub.x sensor 2 as well as by way of the time delayed sensor signal 18 of the first NO.sub.x sensor 1 and by use of the sensor signal 12 of the second NO.sub.x sensor 2 whether the NO.sub.x sensors 1 and 2 are delivering valid measurement values. The determined status is output as the validity signal 19.

[0032] In block 20 the actual calculation of a correlation coefficient signal 23 between the time delayed sensor signal 18 of the first NO.sub.x sensor 1 and the sensor signal 12 of the second NO.sub.x sensor 2 takes place. Furthermore, an averaged correlation coefficient signal 24 and a status signal 25 are determined, the latter indicating whether the correlation coefficient signal 23 is valid.

[0033] The following MatLab program code describes in detail the algorithm running in block 20.

TABLE-US-00001 function [fac_correlation, fac_correlation_avr, fac_correlation_valid] = AlgoCalcCorrelation (x, y, valid, num_correlation, num_avrg, num_reset) % On-line calculation of the variance, the covariance and the correlation coefficient % input variables: % x, y signal values whose correlation is determined % valid status, if the signal values x, y are valid (valid == 1) % num_correlation number of signal points for the calculation % num_avrg number of values for averaging % num_reset number of invalid values for resetting the calculation % output variables: % fac_correlation correlation coefficient between x and y % fac_correlation_avrg averaged correlation coefficient % fac_correlation_valid correlation coefficient is valid % Initialization of the Calculation if init nrcor = 1; nravrg = 0; xold = 0; yold = 0; calcavrg = 0; avrg = 0; fac_correlation = 0; fac_correlation_avrg = 0; fac_correlation_valid =0; k_notvalid = 0; end % Re-start of the Calculation if valid, k_notvalid = 0; else k_notvalid = k_notvalid + 1; end if (k_notvalid > num_reset) nrcor = 1; nravrg = 0; xold = 0; yold = 0; fac_correlation = 0; fac_correlation_valid =0; an = xold; bn = yold; sn = 0; rn = 0; cn = 0; end % Calculation if valid nrcor = nrcor + 1; deltax = x - an; ann = an + deltax / nrcor; sn = sn + (x - ann) * deltax; deltay = y - bn; bnn = bn + deltay / nrcor; rn = rn + (y - bnn) * deltay; cn = cn + (y - bnn) * (x - an) ; an = ann; bn = bnn; xold = x; yold = y: if nrcor == num_correlation fac_correlation = cn / sqrt (sn) / sqrt (rn); an = xold; bn = yold; sn = 0; rn = 0; cn = 0; nrcor = 1; calcavrg = 1; end end if calcavrg nravrg = nravrg +1; avrg = avrg + (fac_correlation - avrg) / nravrg; fac_correlation_avrg = avrg; fac_correlation_valid = 1; if nravrg = = num_avrg; nravrg = 0; avrg = 0; end end

[0034] The calculation algorithm, executed in block 20, calculates "on-line"--that is, without buffering the input signal--the correlation coefficient 23 (called fac_correlation in the source code) for a freely selectable number (called num_correlation in the source code) of measurement values of the input signals 18 (called x in the source code) and 12 (called y in the source code).

[0035] In the first sub-block labeled initialization of the calculation, all of the calculation variables are assigned initial values at the start of the calculation.

[0036] In the sub-block labeled restart of the calculation, the calculation variables are reset in a manner analogous to the initialization routine--that is, the calculation starts again if a defined number (num_reset) of consecutive measurement values are invalid.

[0037] In the sub-block calculation, the running averages an and bn, the variances sn and rn and the covariance cn are calculated from the valid measurement values of the input signals 18 and 12 (x or y respectively in the source code). The results in turn can be used to determine the correlation coefficient fac_correlation according to a freely selectable number of measurement values num_correlation. Thereafter the calculation starts all over again. In addition, a sliding average fac_correlation_avrg (corresponding to the average signal 24 in FIG. 2) is formed from the individual values of the correlation coefficient by means of the last num_avrg calculation results.

[0038] The measurement values of the input signals 18 (x in the source code) and 12 (y in the source code) with an assigned status signal 19 (valid in the source code) having the value "invalid" (that is, valid=0) are not considered for the calculation in the sub-block calculation.

[0039] If the signals 23 (fac_correlation in the source code) and 24 (fac_correlation_avrg in the source code) exhibit valid values, then the status signal 25 (fac_correlation_valid in the source code) is set to 1.

[0040] The above described operation makes it possible to differentiate continuously between an NH3 slip and an NO.sub.x slip without any additional measuring technique.

[0041] FIG. 3 depicts an exemplary curve 30 of the time shifted sensor signal 18 (in ppm--parts per million) of the first NO.sub.x sensor 1 upstream of the SCR catalytic converter and the curve 31 of the sensor signal 12 (in ppm) of the second NO.sub.x sensor 2 downstream of the SCR catalytic converter as well as the curve 33 of the resulting correlation coefficient signal (depicted here is the signal curve of the sliding average 24 of the correlation coefficient multiplied by a factor of 1,000). In the example in FIG. 3, the averaged correlation coefficient exhibits a small amount close to 0 (up to a maximum value of 0.35). That is, the correlation between the two sensor signals 18 and 12 is low. This low value of the correlation coefficient indicates the presence of an NH3 slip.

[0042] FIG. 4 depicts an additional exemplary curve 30' of the time shifted sensor signal 18 (in ppm) of the first NO.sub.x sensor 1 upstream of the SCR catalytic converter and the curve 31' of the sensor signal 12 (in ppm) of the second NO.sub.x sensor 2 downstream of the SCR catalytic converter as well as the curve 33' of the resulting correlation coefficient (depicted here is the signal curve of the sliding average 24 of the correlation coefficient multiplied by a factor of 1,000). It is clear from FIG. 4 that the curve 30' corresponds approximately to the curve 31' multiplied by a linearity factor. Therefore, the correlation coefficient exhibits here a high amount close to 1. That is, the correlation between the two sensor signals 18 and 12 is high. This high value of the correlation coefficient indicates the presence of an NO.sub.x slip.

[0043] The above described method can be used to improve the regulating functions based on the sensor signal of the second NO.sub.x sensor 2, because the method according to the invention makes it possible to differentiate with a degree of certainty between an NO.sub.x slip and an NH3 slip. Therefore, it is possible to prevent, on the one hand, an inadequate metering of NH3 and, thus, higher NO.sub.x emissions and, on the other hand, over-metering of NH3 and, thus, it is possible to minimize the urea consumption and the ammonia slip.

[0044] The differentiation between NH3 and NO.sub.x via the correlation makes it possible to improve a plethora of applications in the exhaust gas aftertreatment. These applications includes, for example: [0045] the calculation of the efficiency of the SCR catalytic converter, [0046] the diagnosis of the NO.sub.x slip, [0047] the control and regulation of the urea metering, [0048] the modeling of the SCR catalytic converter, and [0049] the adaptation of the urea metering, for example, as a consequence of scattering, ageing and errors in the exhaust gas aftertreatment system.

[0050] FIG. 5 depicts an embodiment for regulating the NH3 fill level (or in other words: regulating the urea metering) with an NO.sub.x slip diagnosis and SCR efficiency determination. In this case the configuration shown in FIG. 5 uses the inventive correlation coefficient for a variety of applications. In block 10 a correlation coefficient signal is determined--as already explained in conjunction with FIG. 2--based on the sensor signal 11 of the first NO.sub.x sensor 1 and on the basis of the sensor signal 12 of the second NO.sub.x sensor 2, as a measure for the linear correlation between the time synchronized sensor signals: for example, the correlation coefficient signal 23 or the correlation coefficient signal 24 from FIG. 2. It must be pointed out once again that the number of values that are used for the sliding average 24 can also be selected to be 1.

[0051] The correlation coefficient signal 23 or 24 is used in a block 40, which models the SCR catalytic converter 4, in order to determine the actual NH3 fill level 42 of the SCR catalytic converter 4. Furthermore, the block 40 accepts an actual metering signal 41, which indicates the actual metering with urea. In addition, the block 40 accepts the sensor signals 11 and 12. In order to determine the NH3 actual fill level 42, the sensor signal 12 is evaluated by way of the correlation coefficient signal 23 or 24. In the course of determining the NH3 actual fill level 42, the correlation coefficient signal 23 or 24 is used to differentiate whether the sensor signal 12 indicates NH3 or NO.sub.x. If the correlation coefficient signal 23 or 24 indicates that the sensor signal 12 indicates a certain amount of NH3 (in the case of an NH3 slip), then the NH3 amount corresponding to the sensor signal 12 is subtracted from the current NH3 fill level.

[0052] The determined NH3 actual fill level 42 is compared with an NH3 desired fill level 43; and the difference 44 between the NH3 actual fill level 42 and the NH3 desired fill level 43 is evaluated in block 45, in order to determine a desired metering 46 of urea. As a function of the desired metering 46, the urea is introduced into the SCR catalytic converter 4 by way of the urea metering device 3 (see FIG. 1). The actual metering 41 of urea can deviate from the desired metering 46; for example, because the metering valve cannot inject a currently demanded high actual metering 41. In this case the block 47 represents the regulating path between the desired metering 46 and the determined or estimated actual metering 41, where the metering device 3 (see FIG. 1) is a part of block 47.

[0053] The closed loop control circuit comprising the blocks 45, 47 and 40 serves to quickly regulate the degree of NH3 filling. The slow adaptation of the degree of NH3 filling results from the adaptation of the NH3 desired fill level 43. The inventive correlation coefficient signal 23 or 24 is used in block 48 to determine the NH3 desired fill level 43. In principle, the block 48 determines the NH3 desired fill level 43 by means of one or more state variables of the exhaust gas system 49 (for example, the temperature of the exhaust gas and the size of the exhaust gas mass flow) as well as by means of the NO.sub.x signal 11 of the first NO.sub.x sensor 1. Working on this basis, when the temperature is low (for example, 150.degree. C.), the NH3 desired fill level 43 is high, because at a low temperature the NH3 storage capacity of the SCR catalytic converter 4 is high. In contrast, when the temperature is high (for example, 400.degree. C.), the NH3 desired fill level 43 is low, because in this case the storage capacity is low. In addition, in block 48 the NH3 desired fill level is adapted to the currently present engine emission--that is, the NO.sub.x signal 11 of the first NO.sub.x sensor 1. When the emission is low, the desired fill level 43 is lower than in the case of a high emission. Furthermore, the NH3 desired fill level 43 is adapted as a function of the correlation coefficient signal 23 or 24. If the correlation coefficient signal 23 or 24 indicates that an NH3 slip prevails, then the NH3 desired fill level 43 should be lowered; and if the correlation coefficient signal 23 or 24 indicates that an NO.sub.x slip prevails, then the NH3 desired fill level 43 should be raised.

[0054] In addition to the above described procedure for regulating the metering process, the correlation coefficient signal 23 or 24 is still used in block 50 for diagnosing the NO.sub.x slip. The NO.sub.x slip diagnosis 50 evaluates with the correlation coefficient signal 23 or 24 the sensor signal 12 of the second sensor as to whether this signal involves the indication of NO.sub.x or NH3 and outputs an NO.sub.x slip signal 51. For example, the NO.sub.x slip signal 51 of the NO.sub.x slip diagnosis 51 is marked then as valid (by an additional bit (not illustrated)), when the correlation coefficient signal 23 or 24 indicates that an NO.sub.x slip is, in fact, present. On exceeding a defined threshold value, a warning can be output to the driver by way of the NO.sub.x slip signal 51.

[0055] Furthermore, the correlation coefficient signal 23 or 24 is used in block 52 for determining the efficiency 53 of the SCR catalytic converter 4. The efficiency 53 can be determined, for example, as a ratio between the NO.sub.x signal 11 of the first NO.sub.x sensor 1 and the NO.sub.x signal 12 of the second NO.sub.x sensor 2 (in the case of an NO.sub.x slip). The efficiency that is determined in this way is marked as valid as a function of the correlation coefficient signal 23 or 24 (an additional bit (not illustrated)), when the correlation coefficient signal 23 or 24 indicates that an NO.sub.x slip is, in fact, present.

[0056] The foregoing disclosure has been set forth merely to illustrate the invention and is not intended to be limiting. Since modifications of the disclosed embodiments incorporating the spirit and substance of the invention may occur to persons skilled in the art, the invention should be construed to include everything within the scope of the appended claims and equivalents thereof.

* * * * *