Method and apparatus for controlling an exhaust gas after-treatment system

State of the art

0001The invention relates to a method and a device for controlling an exhaust gas aftertreatment system.

0002A method and a device for controlling an exhaust gas aftertreatment system, in particular for an internal combustion engine, are known from EP-A 0 621 400. There, a method and a device for controlling an internal combustion engine are described, in which fuel is metered to an internal combustion engine. Furthermore, a catalyst is arranged in the exhaust pipe. Fuel is metered to the internal combustion engine in such a way that a part of the fuel reaches the catalytic converter unburned via the exhaust gas lines and reacts there with the nitrogen oxides.

0003Furthermore, catalytic converter systems are known in which a reducing agent is supplied between the internal combustion engine and the catalytic converter. Ammonia is often used as a reducing agent. A urea water solution is metered in, which is converted into ammonia in a hydrolysis catalytic converter. In the actual catalyst, the nitrogen oxides NOX are then reduced to nitrogen with the help of ammonia.

0004The dosage of the reducing agent is very complex because a large number of temperature sensors is required. Furthermore, dosing in dynamic engine operation, ie when accelerating and decelerating, is very complex. It is also problematic that a so-called ammonia breakthrough can occur if the urea water solution is overdosed. This means that if too much reducing agent is metered in, reducing agent and / or incompletely converted reducing agent can get into the environment together with the exhaust gas. A closed control loop using an ammonia sensor is very complex.

Object of the invention

0005The object of the invention is to reduce NOX emissions in a method and a device for controlling an exhaust gas aftertreatment system. This reduction in NOX emissions is achieved through an exact dosage of a reducing agent.

Advantages of the invention

0006By means of the method according to the invention, an exact dosage of a reducing agent for a catalyst is possible. As a result, NOX emissions from the engine can be minimized in both stationary and unsteady operation with a low reductant breakthrough. Furthermore, the application of the control device can be simplified considerably. It is particularly advantageous that only a small number of sensors are required.

0007Advantageous and expedient refinements and developments of the invention are characterized in the subclaims.

drawing

0008The invention is explained below with reference to the embodiments shown in the drawing. It show the Figure 1 is an overview block diagram of an apparatus for controlling an internal combustion engine, 2 shows a block diagram of an apparatus for controlling the supply of reducing agent, Figure 3 is a block diagram of a portion of Figure 2 and Figures 4, 5 and 6 are a flowchart of the procedure of the invention.

0009An internal combustion engine 100 is shown in FIG. This is connected to a catalytic converter 110 via an exhaust pipe 115. From the catalytic converter 110, the exhaust gas reaches the surroundings via an exhaust pipe 120. A metering unit 130 is arranged in the exhaust line 115. Furthermore, at least sensors 170 are arranged on the internal combustion engine and sensors 180 in the exhaust line 120.

0010Furthermore, actuators 160 can be provided which control the internal combustion engine. The output signals of the sensors go to a control unit 155, which applies control signals to both the dosing unit 130 and the actuator 160.

0011In the embodiment shown in FIG. 1, only one control unit 155 is provided, which controls both the internal combustion engine and the exhaust gas aftertreatment system, in particular the metering unit 130. However, it can also be provided that at least two separate control units are provided, one control unit controlling the exhaust gas aftertreatment system. The control unit for controlling the internal combustion engine transmits various sizes via data lines to the control unit for controlling the exhaust gas aftertreatment system.

0012The control unit 155 calculates the control signals with which the actuators 130, 160 and / or other actuators (not shown) are applied, on the basis of various signals which are detected by the sensors 170, 180 and other sensors (not shown). The internal combustion engine is controlled by means of the actuator 160. It can be provided, for example, that the actuator 160 influences the amount of fuel to be injected into the internal combustion engine and / or the start of the injection.

0013The control unit 155 also calculates, depending on various operating parameters of the overall system, the amount of reducing agent that is necessary to reduce the nitrogen oxide (NOX) that occurs in the combustion engine in the catalytic converter 110. This can significantly reduce the amount of NOX that is released to the environment via the exhaust pipe 120.

0014The catalytic converter 110 is located in the exhaust gas of an internal combustion engine. A reducing agent is metered in before the catalytic converter by means of the metering unit 130, depending on the operating point. The metered-in amount of reducing agent is predetermined by the metering unit 130 in such a way that minimized NOx emissions result for each operating state, at the same time preventing an ammonia breakthrough.

0015According to the invention, a reducing agent is fed upstream of the catalyst. On the basis of a load variable QK and at least one first further operating parameter, a variable is specified which determines the quantity of reducing agent supplied. The first further operating parameter is preferably the speed. The load variable QK is preferably a fuel quantity signal. An internal quantity in the control unit 155 can be used as the fuel quantity signal, for example, which is used to determine the control signal for the actuator 160. Furthermore, a size which corresponds to the torque of the internal combustion engine can be used as the load size QK. The load size is referred to below as the fuel quantity signal QK.

0016The size that determines the amount of reducing agent supplied can be corrected as a function of at least one further second operating parameter. Based on the fuel quantity signal and the first further quantity, which is preferably the speed of the internal combustion engine, a basic quantity of reducing agent supplied is specified. This basic quantity is then corrected depending on different sizes.

0017FIG. 2 shows the part of the control unit 155 that controls the exhaust gas aftertreatment system in greater detail. This part in particular provides the control signal for loading the dosing unit 130. An engine block 200 is connected to a sensor 170, further sensors (not shown) and other components 171 of the control unit 155. These sensors deliver various signals relating to the operating state of the internal combustion engine, such as the engine speed N, the cooling water temperature, the boost pressure, the air temperature and / or the air humidity. The other components 171 supply the load quantity QK, which in particular characterizes the amount of fuel injected into the internal combustion engine. All signals that characterize the fuel quantity to be injected can be used as the load variable. In this case, a variable that is present internally in the control unit 155 is preferably used. For example, the control signal for the actuator 160 is used as the fuel quantity QK and or a torque signal.

0018The engine block 200 supplies various signals MNOX, and / or V to a cat block 210. Furthermore, the cat block 210 is supplied with output signals from various sensors 180 which, for example, contain one or more signals T relating to the exhaust gas temperature.

0019The catblock 210 sends a signal MHWL to a correction block 220. The correction block 220 applies a corrected signal to a dosing block 230.

0020The metering block 230 is also supplied with the output signal or outputs from a sensor 172. The dosing block 230 converts these signals into a control signal for the dosing unit 130. This control signal is essentially defined by its control duration ti and its frequency f. These two variables are checked by a monitor 240 and forwarded to the dosing unit 130.

0021The size MNOX and the size V are calculated in the engine block 200 on the basis of various operating parameters of the internal combustion engine. The size MNOX characterizes the current mass flow of nitrogen oxides and the size V characterizes a size that corresponds to the exhaust gas volume flow. This variable V corresponds to the exhaust gas volume flowing through the catalytic converter in one time unit. The size MNOX characterizes the mass flow of substances to be reduced.

0022These variables MNOX and V are preferably calculated using a map. In addition to the speed N and the load size QK, other sizes that characterize the behavior of the internal combustion engine can also be taken into account. These are in particular temperature and pressure values ​​which charac terize the ambient air and / or the air supplied to the internal combustion engine.

0023Starting from the fuel quantity and the speed of the internal combustion engine, the signal MNOX, which characterizes the mass flow of substances to be reduced, in particular of nitrogen oxides NOX, is determined and / or adopted by the engine control.

0024The cat block 210 calculates the quantity MHWL, which corresponds to the required amount of reducing agent, based on the current mass flow MNOX of nitrogen oxides NOX, the exhaust gas volume flow V and / or at least one temperature signal T detected by the temperature sensor 180. The cat block 210 determines the efficiency W of the Catalyst 110 depending on the current operating state of the internal combustion engine and / or the catalyst. The required mass flow MHWL of reducing agents is determined on the basis of the efficiency W of the catalytic converter and the mass flow of nitrogen oxide, which corresponds to the signal MNO X. The catblock is shown in more detail in FIG.

0025The correction block 220 takes into account various influences on the overall system, consisting of the internal combustion engine, injection system and catalytic converter. In particular, the aging of the overall system is taken into account.

0026For this purpose it is provided, for example, that a variable is determined which characterizes the running time of the system and / or the exhaust gas aftertreatment system. For example, the consumption of reducing agent can be used as a corresponding quantity. On the basis of this variable, a correction value is determined which takes into account the aging of the system. The correction value is chosen such that it compensates for the change, in particular the decrease, in the efficiency of the system. If the efficiency of the system decreases, a correction value results, which results in a reduction in the amount of reducing agent. According to the invention, a correction is made depending on a variable that takes into account the running time and / or the aging of the system and / or the internal combustion engine.

0027Starting from the required mass flow to the reducing agent MHWL, the metering block 230 calculates signals for controlling the metering unit 130 using further operating parameters, which are detected by the sensor 172. In the exemplary embodiment shown, the metering unit 130 is acted upon by a clocked signal. The frequency f and / or the period ti of the signal determines the metered amount of reducing agent.

0028The monitoring 240 following the dosing block 230 checks the two variables ti and f to determine whether all criteria for an injection of reducing agents have been met. It is preferably provided that a so-called signal range check is carried out. For this purpose, it is checked whether the two signals ti and / or f are greater than a lower threshold and / or less than an upper threshold. It can also be provided that a check is carried out to determine whether there are operating states in which the reducing agent can be metered in. For example, the temperature of the catalyst can be detected by means of a temperature sensor. Dosing only takes place if the temperature is higher than a threshold value. It is particularly advantageous if a sensor is provided that emits a signal that indicates the amount of reducing agent that is still in the storage container for the reducing agent.

0029The catblock 210 is shown in more detail in FIG. Elements already described in FIGS. 1 and 2 are designated by corresponding reference symbols. An essential component of the catblock 210 are various characteristic diagrams 300 which are subjected to a temperature calculation 310 with at least one temperature signal T, a space velocity calculation 320 with a signal RG and the engine block 200 with the signal MNOX.

0030A signal W, which characterizes the efficiency of the catalytic converter, is present at the output of the characteristic diagrams 300. A signal V, which characterizes the volume flow in the exhaust gas, is fed from the engine block 200 to the space velocity calculation 320.

0031The temperature calculation 310 processes an output signal TV of a temperature sensor 180a, which is arranged in front of the catalytic converter 110 in the exhaust gas pipe 115, and an output signal TN of a temperature sensor 180b, which is arranged in the exhaust pipe 120 after the catalytic converter 110.

0032The signal MNOX also arrives at a link point 335, in which it is linked in the output signal of a specification block 330. The output signal MHWLW of the node 335 arrives at a second node 340, where it is linked to the output signal W of the characteristic diagrams 300. The output signal MHWLS of the node 340 arrives at the input of a node 355 at whose second input the output signal of a transient correction 350 is present. On the one hand, the transient correction 350 is supplied with the output signal T of the temperature calculation 310 and the difference between the two input signals TV and TN of the temperature calculation 310 via the node 352 as a signal TD. The output signal MHWLI of the node 355 arrives at a node 365 where it is linked to the output signal SF of a distance correction 360. The signal MHWL is present at the output of node 365, which is fed to correction block 220.

0033At node 335, the nitrogen oxide mass flow MNOX is converted into a desired mass flow MHWLW of reducing agent by means of a constant factor. The relationship applies:<maths id="math0001" num=""><math display="block"><mrow><mtext>MHWLW = MNOX * K</mtext></mrow></math><img file="EP1024254A2_D0001.tif" /></maths>

0034The mass flows are given in grams per hour. Size K is a constant that, for example, assumes a value of almost 2 for a 32.5% urea water solution. In the case of differently concentrated urea water solutions or other reducing agents, a corresponding value must be selected for the constant K. The value for the constant K is stored in the specification 330.

0035The desired mass flow MHWLW for the reducing agent would be metered if the catalyst 110 were able to reduce the total nitrogen oxide by converting the entire reducing agent to nitrogen. Since the degree of implementation is dependent on various operating parameters, the entire efficiency W of the system consisting of catalyst 110 and injection system 130 is stored in map 300. The map 300 is preferably a four-dimensional map or a plurality of three-dimensional maps. The efficiency in the characteristic diagram 300 is stored as a function of the nitrogen oxide mass flow MNOX, the characteristic catalyst temperature T and the space velocity RG of the exhaust gas. Starting from at least the signal MNOX, which characterizes the amount of substances to be reduced, and / or the characteristic catalyst temperature T, the signal w, which characterizes the system efficiency, is determined.

0036The space velocity RG is predetermined on the basis of the exhaust gas volume flow V, which is determined by the engine block 210, and a constant, which is essentially determined by the volume of the catalytic converter, by the space velocity calculation 320. Instead of the space velocity, the exhaust gas volume flow V can also be used directly as an input variable.

0037The characteristic catalytic converter temperature is determined on the basis of the first exhaust gas temperature TV, which represents the temperature in the exhaust pipe before the catalytic converter, and the second exhaust gas temperature TN, which represents the temperature after the catalytic converter in the exhaust pipe.

0038The characteristic catalyst temperature T is a fictitious temperature, which is determined on the basis of the temperature TV before the catalyst and the temperature TN after the catalyst by the temperature calculation 310. This temperature value reflects the real behavior of the catalytic converter as a function of the temperature in both stationary and unsteady engine operation.

0039It is preferably provided that arithmetic averaging and / or subsequent weighting is carried out with a constant value or a value dependent on the operating parameters. It is preferably provided that the weighting takes place as a function of the volume flow V. It is particularly advantageous if the time course of various operating parameters is included in the weighting. Thus, it can be provided that the weighting takes place depending on the change in the load size over time.

0040Furthermore, it can be provided that the value T is stored in corresponding characteristic diagrams as a function of the variables TV and TN and, if appropriate, the further operating parameters.

0041As an alternative to the size MNOX, which characterizes the nitrogen oxide mass flow, the nitrogen oxide concentration can also be used as an input variable.

0042At node 340, the basic quantity MHWL0 of reducing agent is determined on the basis of the desired mass flow MHWLW and the efficiency W. The two variables are preferably multiplied at node 340. Based on the desired amount of reducing agent and the signal W that characterizes the system efficiency, this means that the basic value MHWL0 of reducing agent is determined.

0043The map 300 indicates the efficiency W for both the stationary and the transient engine operation. If there are large temperature differences between the inlet of the catalytic converter and the outlet of the catalytic converter during transient operation, an additional correction factor is used. This correction factor is provided by the transient correction 350. This means that, based on the first exhaust gas temperature TV and the second exhaust gas temperature TN, a correction variable SF is specified. This factor is dimensioned in such a way that it avoids overdosing in transient engine operation. With the size SF, which is predefined by the distance calculation 360, the amount of reducing agents is reduced once more by a further attempt to ensure that no ammonia breakthrough occurs. This value SF takes into account system tolerances. It is specified in such a way that no ammonia breakthrough occurs even at maximum tolerance.

0044It is advantageous in this embodiment that the effective catalyst temperature T can be calculated very precisely by using two temperature sensors which are arranged before and after the catalyst 110. Since the efficiency W conversion of nitrogen oxide is essentially determined by the behavior of the catalyst and in particular by the effective catalyst temperature, a very precise metering of reducing agents is possible. The more precisely the effective catalyst temperature T which characterizes the reaction behavior is determined, the smaller the safety distance SF can be chosen in favor of a high degree of efficiency for the conversion of nitrogen oxide without ammonia breakthrough.

0045It is also particularly advantageous that the control unit 155 is divided into functional blocks, which essentially comprise an engine block, a cat block and the metering block. As a result, the effort involved in the application, for example in the event of changes to one of the components engine, catalytic converter or metering unit 130, can be significantly reduced. As a result, costs for development and application can be significantly reduced, especially with small numbers of control units.

0046Another embodiment for determining the control signals for the metering unit 130 is shown in FIG. 4 as a flow chart. In a step 400, a basic quantity MHWLO of the reducing agent is read out from a characteristic map on the basis of operating parameters. The basic quantity MHWLO of reducing agents is preferably specified as a function of the load size QK and the speed N of the internal combustion engine.

0047In the subsequent step 410, the basic quantity MHWL0 of reducing agent is corrected depending on the current operating state of the internal combustion engine. In this correction, further operating parameters are taken into account that do not go into the map of step 400. These are in particular the air humidity, various temperature values ​​that determine the temperature and / or pressure value of the ambient air or of the air supplied to the internal combustion engine and / or a temperature value that identifies the temperature of the internal combustion engine. For this purpose, a first correction value K1 is determined depending on an operating parameter. As a result, the amount of reducing agent can be corrected in particular at the start.

0048In the event of a load jump, the nitrogen oxide emission changes with a time delay and only reaches the new stationary value late. If the amount of fuel injected suddenly increases by a very large value to a new value, the nitrogen oxide emission only increases after a certain delay to a new value which corresponds to the increased amount of fuel. Furthermore, the efficiency of the conversion of nitrogen oxide also changes only with a time delay. The amount of reducing agent MHWL read from the map in step 400, which applies to the new steady-state operating state, therefore assumes a wrong value for a short time, that is to say if the amount of fuel increases, the amount of reducing agent takes on a too high value and if the amount of fuel decreases it too low .

0049According to the invention, the amount of reducing agent is adapted to this dynamic behavior of the nitrogen oxide emission and / or the efficiency of the reaction. This takes into account the effect that the exhaust gas temperature and the efficiency of the catalyst and / or the nitrogen oxide emission change with a time delay.

0050This correction takes place in step 420. In step 420, a second correction value K2 is determined, which takes into account the dynamics of the internal combustion engine and the catalyst system. The dynamic behavior is taken into account by means of a temperature difference. This temperature difference is the deviation between a measured temperature and a target value TS of the exhaust gas temperature. The temperature is preferably measured after the catalyst. This setpoint TS for the exhaust gas temperature corresponds to the exhaust gas temperature in dynamic operation. This value is specified in step 425 as a function of various operating parameters, such as the quantity of fuel QK to be injected and the speed N of the internal combustion engine. The actual exhaust gas temperature is measured by means of a temperature sensor in, after and / or before the catalytic converter 110.

0051Based on this temperature difference ΔT, the injected fuel quantity QK and / or the speed N of the internal combustion engine, the second correction value K2 is specified in step 420. One or more maps are preferably used for this purpose.

0052The efficiency W of the catalytic converter and the emission of nitrogen oxides from the engine depend to a large extent on the exhaust gas temperature. According to the invention, based on the difference between the current exhaust gas temperature, which is detected by means of a sensor, and the temperature value TS applicable to the respective operating point, the amount of reducing agent is adjusted, ie a correction value K2 is determined.

0053In the subsequent step 430, a third correction factor K3 is determined, which takes into account the aging of the catalyst system over the total operating time of the system. This takes into account the change in efficiency, in particular the reduction in efficiency, of the catalyst and / or of the metering unit 130. This step corresponds essentially to block 220 in FIG. 2

0054It is particularly advantageous if a diagnostic function is provided in a further step 440. This diagnostic function requires a nitrogen oxide and / or an ammonia sensor. The nitrogen oxide sensor monitors the operability of the system and determines whether nitrogen oxide conversion takes place. If, for example, an insufficient concentration of reducing agents is mixed in, an increased emission of nitrogen oxides is obtained, which is detected by means of the nitrogen oxide sensor. This can be the case, for example, when the storage container for the reducing agent is empty. Accordingly, an ammonia breakthrough, that is, an excessively high concentration of reducing agent, can be reliably detected with an ammonia sensor. On the basis of this diagnosis, a fourth correction factor K4 is determined. In a simplified embodiment, this step can also be omitted. This can reduce the number of sensors required.

0055As in block 360 of FIG. 3, a fifth correction factor K5 is determined in step 450, which corresponds to the value SF in FIG. 3.

0056Then in step 460, the amount MHWL of reducing agent is calculated on the basis of the basic amount MHWL0 stored in the characteristic diagram 400 and the correction factors K1, K2, K3, K4 and K5. The sizes are preferably added. Alternatively, a multiplicative linking of individual sizes or a mixed linking of the sizes can also be provided. This means that individual correction variables have an additive effect and others have a multiplicative effect.

0057In step 470, as in step 230 of FIG. 2, the control time ti and the control frequency f for the metering unit 130 are calculated. In step 480, the diagnosis of the system is carried out in step 240.

0058In this embodiment, the necessary amount of reducing agent MHWL is stored directly in a map. Correction values ​​for correcting the amount of reducing agent are stored in several maps for different positive and negative temperature differences. It is preferably provided that maps are provided in which the correction value is stored as a function of the speed and the injected fuel quantity QK. Different maps are provided for different temperature differences between the target temperature TS and the measured temperature. Interpolation is then carried out between these maps.

0059Another embodiment is shown in FIG. In a first step 500, the target temperature TS is specified as a function of the operating state of the internal combustion engine. For this purpose, it is provided in particular that the target temperature TS is stored as a function of the fuel quantity QK to be injected and the speed N of the internal combustion engine. According to the invention, further operating parameters can also be taken into account.

0060These are, in particular, the air humidity, various temperature values ​​which characterize the temperature and / or pressure value of the ambient air or the air supplied to the internal combustion engine and / or a temperature value which characterizes the temperature of the internal combustion engine.

0061For this purpose, the fuel quantity QK to be injected and the rotational speed N are processed in step 500. A temperature difference ΔT is then determined on the basis of the target temperature TS and a measured temperature value T temperature of the catalyst. This value ΔT indicates the deviation of the current temperature from the temperature applicable for this stationary operating state.

0062In a next step 510, the basic value MHWL0 for the quantity of reducing agent is read out from various characteristic maps depending on the quantity of fuel QK to be injected and the speed N. It is provided that different maps are used for different temperature differences AT. For intermediate values ​​of the temperature difference for which no maps are stored, the values ​​for the amount of reducing agent MHWL0 are determined in step 520 by means of an interpolation.

0063Then in step 530, a second correction factor K2 is determined, which takes into account the influence of further ambient conditions, such as the cooling water temperature TW. Subsequently, further correction variables K3, K4 and K5 are determined as described in FIG. 4. According to the invention, a correction is made depending on a variable which indicates that a cold start is present.

0064In the embodiment according to FIGS. 4, 5 and 6, the basic quantity MHWL0 of supplied reducing agent is specified on the basis of at least the load quantity QK, the speed N of the internal combustion engine and / or a temperature signal which is determined by the exhaust gas temperature. In the embodiments according to FIGS. 5 and 6, the temperature signal is predetermined on the basis of the amount of fuel injected, the speed of the internal combustion engine and the measured exhaust gas temperature. The temperature signal indicates the deviation of the measured temperature variable from the target temperature TS. The setpoint temperature is also referred to as the expected temperature variable.

0065Starting from the temperature signal, which characterizes the deviation between the measured and the expected temperature, and at least the injected fuel quantity, the basic quantity MHWL0 of the reducing agent supplied is specified.

0066Another embodiment is shown in FIG. In this embodiment, the efficiency W of the catalytic converter is specified on the basis of various operating parameters and the required amount of reducing agent MHWLO is calculated using the efficiency W and the amount MNOX of emitted nitrogen oxides.

0067The efficiency W of the catalyst essentially depends on the exhaust gas temperature, the operating point of the internal combustion engine, which is defined by the fuel quantity QK and the speed N, and the nitrogen oxide emission of the engine. With the difference ΔT between the current temperature value T for the respective operating point and the stationary value TS of the respective operating point stored in the characteristic diagram, the efficiency W of the system is determined from characteristic diagrams. The basic amount MHWLO for the reducing agent is calculated on the basis of the efficiency W and the amount MNOX of nitrogen oxides emitted. This value MHWLO is then corrected as described in FIGS. 4 and 5.

0068This embodiment is shown in more detail in FIG. 6. In a first step 600, the target temperature TS is determined. The target temperature TS, which is stationary in certain operating states, is read from a map based on the injected fuel quantity QK and the speed of the internal combustion engine. The temperature difference ΔT is determined on the basis of the temperature T measured with a sensor.

0069In the subsequent step 610, the efficiency of the overall system is determined. The efficiency is preferably specified as a function of the temperature difference ΔT, the fuel quantity QK to be injected and / or the speed N. Several maps are provided in which the efficiency W for a fixed temperature difference ΔT as a function of the injected fuel quantity QK and the speed N are stored. In the subsequent interpolation 620, the respective efficiency is then calculated using an interpolation.

0070In step 630, the quantity MNOX of nitrogen oxides is determined on the basis of various operating parameters, in particular the fuel quantity QK to be injected and the speed N of the internal combustion engine. Based on the amount MNOX of nitrogen oxides and the efficiency W, the amount MHWL0 of reducing agent is then determined in step 640. In the subsequent steps 530, 430, 440, 450, 460, 470 and 480, a corresponding correction is made as in the figures described above.

0071In this embodiment, starting from the temperature signal and at least the fuel quantity QK, a signal W is specified which characterizes a system efficiency. Starting from the signal W, which characterizes the system efficiency, and the signal MNOX, which characterizes the amount of substances to be reduced, the basic amount MHWL0 of the reducing agent supplied is specified. The signal that characterizes the amount of substances to be reduced is specified on the basis of the injected fuel amount and the speed of the internal combustion engine.