DE10019245A1 - Controlling rich regeneration of NOx storage unit of lean burn engine, determines oxygen storage during first regeneration, regenerates at rate related to NOx stored, and desulfurizes when regeneration interval is slight - Google Patents

Abstract

The quantity of oxygen in the unit is determined from the oxygen sensor peak signal amplitude, during a first regeneration process. NOx is removed from the unit at a frequency related inversely to the quantity of NOx stored in it. Desulfurization is carried out to restore the capacity of the unit, once regeneration time falls below a minimal value. An Independent claim is included for the corresponding regeneration control system. It includes an oxygen sensor for the exhaust gases flowing through the unit and a control/regulation module programmed to carry out the method. Preferred features: Deterioration of the unit is indicated, when a predetermined number of desulfurization processes have been carried out, without increase in regeneration interval. A regeneration adjustment factor is related to the capacity of the unit, and filling time is also adjusted in accordance with the factor. NOx storage quantity is adjusted, to saturate a predetermined fraction of the capacity of the unit. The initial value of filling time of the unit, is determined from a reference table, as a function of engine speed and loading, and/or from the mass air flow rate. It is determined as an inverse power of the product of engine load and engine speed. A related process for filling and regenerating a lean burn NOx storage unit (34) is described, based on the foregoing principles.

Description

This invention relates to a system and a method for regulating the regeneration cycle of a NO _x collecting vessel located downstream of a lean-burn engine in the exhaust system thereof.

Known control systems of lean-burn engines include an air-fuel mixture ratio controller that supplies the engine with an amount of fuel that is proportional to a measured air mass flow so as to periodically achieve a lean air-fuel mixture ratio and thereby lower fuel consumption. Since a typical one in the exhaust gas path of the engine is the three-way catalytic converter is less effective in converting the NO _x portion generated by the engine if there is an excess of oxygen in the lean engine operation, a lean NO _{x is} proposed in the prior art - Install the collecting vessel in the exhaust gas path downstream of the three-way catalytic converter in order to chemically bind the residual portion of NO _x generated during lean operation.

Since the capacity of the receptacle for the storage of NO _{x is} clearly limited, the stored NO _{x is} "flushed out" from the receptacle periodically by supplying the engine with an enriched air-fuel mixture, the mixing ratio of which is 14, relative to the stoichiometric range. 65 is enriched and z. B. has a value below 13. During this flushing process, excess HC and CO components break through the three-way catalytic converter, pass through the collecting vessel, react there with the stored NO _x and reduce it to harmless N ₂ and O ₂ . The amount of fuel excess required for purging to release the stored NO _x is often expressed in terms of a "purging time", that is, a period of time during which the purging fuel is supplied in the enriched air-fuel mixture with a substantially constant air mass flow rate.

The timing of the purging event must be controlled or regulated in such a way that the collecting vessel does not otherwise exceed its absorption capacity for NO _x , since NO _{x would} then pass through the collecting vessel and cause an increase in the NO _x component in the exhaust gas. In order to avoid flushing out only part of the filling of the collecting vessel, the frequency of the flushing processes is preferably regulated depending on the exhaust gas standards applicable to the fuel enrichment in the exhaust gas occurring during the flushing events.

It has been recognized in the prior art that the absorption capacity of the NO _x collecting vessel is in turn a function of many variable variables, which include the temperature, the history, the level of sulfurization and thermal damage, i.e. the extent of damage to that in the collecting vessel Contain NOx absorbent material due to excessive heating.

Here reference is made to the US-Patent 5,437,153 referred to, which also teaches that begins with the collection vessel of the absorbed NO _x in kreme adjusting rate to fall when the collecting vessel comes close to its maximum capacity. Accordingly, US Pat. No. 5,437,153 teaches the use of a NO _x target capacity which is considerably less than the actual storage capacity of the NO _x collecting vessel, thereby giving the collecting vessel a perfect instantaneous NO _x absorption effect, that is to say that the collection vessel the entire generated by the motor can absorb _NOx amount as long as the amount of _NOx remains a mounted below the target capacity. To rejuvenate the collecting vessel, a rinsing event is always initiated when accumulated estimates of the motor-generated NO _{x reach} the target capacity of the collecting vessel.

Actually embedded in the receiver _NOx amount depends on the concen tration of NO _x in the engine supplied gas, the exhaust gas flow rate, in order gebungsfeuchtigkeit, collecting vessel temperature and other variables from. Thus both the capacity of the receiver and the actual collecting vessel embedded in on NO _x amount complex functions form many variables. It is desirable to monitor and regulate the rinsing and filling processes of the collecting vessel so as to ensure that the collecting vessel works under optimal operating conditions at all times.

If the engine is operated with a sulfur-containing fuel, the collecting vessel absorbs sulfur and thereby reduces on the one hand the absolute capacity for NO _x and on the other hand the instantaneous NO _x output of the collecting vessel. If such a sulfurization of the receptacle exceeds a critical value, the absorbed SO _x must be desorbed or "burned out" by desulfurization. The temperature of the collecting vessel is raised above about 650 ° C in the presence of an excess of HC and CO. By way of example only, U.S. Patent 5,746,049 teaches a desulfurization process for the receiver that increases the temperature of the receiver above at least 650 ° C by supplying a secondary air stream to the exhaust gas upstream of the NO _x receiver when the engine is enriched with air. Cold mixture is operated, and this is based on the resulting exothermic reaction, which raises the temperature of the collecting vessel to the desired level for burning out the SO _x in the collecting vessel.

The lean NO _x control method according to the invention uses information that depends on the amount of NO _x stored in the collecting vessel and the maximum NO _x storage capacity of the collecting vessel, so that the condition or the "health" of the collecting vessel is diagnosed and the purging parameters of the collecting vessel vessel, such as the decrease rate of the NO _x capacity, the flushing time and the flushing strength can be compared online while the engine is working in the vehicle. The decrease rate of the NO _x capacity is the one with which the NO _x capacity of the collecting vessel decreases during the NO _x filling process. In addition, a decision is made to desulfurize the collecting vessel on the basis of the observed reduction in the absorption capacity of the collecting vessel for nitrate and the dependent increase in the rate of decrease of the NO _x capacity. In this way, the collecting vessel works optimally with regard to the NO _x conversion performance and minimizes the CO and NO _x content in the exhaust and the fuel consumption of the vehicle. Intelligent desulfurization of the collecting vessel ensures that the NO _x conversion capacity of the collecting vessel always remains above a given minimum value.

More precisely, in accordance with the invention, the rate of decrease of the NO _x capacity of the collecting vessel is monitored and a regulation of the frequency and the rinsing depth of the collecting vessel as well as a regulation of the desulfurization of the same are carried out in a closed loop. The purging frequency of the receptacle inversely depends on the rate at which NO _{x is} stored in the receptacle, while the purge depth depends on the amount of NO _x which is released from the receptacle and converted into N ₂ and O ₂ . A programmed computer regulates the filling and rinsing times of the collecting vessel on the basis of the voltage amplitude of an oxygen sensor of the switched type and the time-dependent reaction of the same. The flush frequency, which in the ideal case directly from the rate of decrease in NO _x - is dependent on storage capacity controlled so that the collecting vessel is not ne NO _x -Speicherkapazitätsgrenze addition is filled.

In addition, according to the invention, the collecting vessel is filled up to a predetermined fraction of its actual capacity on the basis of the NO _x capacity decrease rate and then completely emptied during a flushing process. Since the capacity of the receptacle decreases over time due to its deterioration, a purge optimization routine in a closed loop generates a balancing factor which is used to adjust the NO _x capacity decrease rate in order to achieve a storage of NO _x which is sufficient for the receptacle to to fill the desired fraction of its capacity. Since the capacity of the receptacle substantially decreases as the actual NO _x capacity decrease rate will indicate when it becomes equal to or greater than a predetermined maximum NO _x capacity decrease rate, desulfurization of the receptacle is performed to try the original capacity of the collecting vessel again. When a predetermined number of Ent desulphurization processes of the collecting vessel without any reduction of the NO _x - capacity decrease rate has been performed, the collection vessel should be replaced, and an operator will be informed accordingly by a display.

The initial value of the NO _x capacity decrease rate of the collecting vessel is determined by mapping the engine system and the collecting vessel. The map recording generates values for the filling rate and the optimal NO _x capacity decrease rate of the collecting vessel as a function of the engine load or the air mass flow rate.

The air-fuel mixture ratio does not change very much during the operation in which the collecting vessel is used, and changes in the engine speed hardly affect the NO _x capacity decrease rates. In this way, the primary change in the NO _x capacity decrease rate of the receptacle remains within a small inverse power (e.g., inverse square) of the engine load or mass flow rate.

The above and other objects and other advantageous features of the invention will become apparent from the following detailed description of a preferred embodiment management example even clearer if reference to the accompanying drawings is taken, which show:

Fig. 1 is a diagram of an engine control system that embodies the principles of the invention;

FIG. 2 shows graphically the voltage response of an oxygen sensor depending on the air / fuel ratio;

Fig. 3 different curves the (a) the air-fuel mixture ratio of the engine, (b) the reaction of the oxygen sensor lying in the exhaust, (c) the EGO data collection and (d) the CO content in the exhaust depending on a short purge time ( 1 ), an average rinsing time ( 2 ) and a long rinsing time ( 3 ) Chen;

FIG. 4 shows in more detail the response of the oxygen sensor depends on a short rinsing time (1), a medium length rinsing (2) and a long rinse time (3);

Fig. 5 is a normalized oxygen sensor saturation time t _sat, as a function of purge time t _p;

Fig. 6 is a normalized saturation time t _sat in response to a Sau erstofffühlerspitzenspannung V _p for the case that this peak voltage is smaller than a reference voltage V _ref;

Fig. 7 shows the relationship between a flush time t _p, and a filling time t _F of the collecting vessel, the t and the optimal rinsing time _PT and two under optimum flushing dots represents 1 and 2;

Fig. 7a shows the relationship between the purge time and of the filling time when the rinsing time for all fill times is optimal. The optimal rinsing time t _PT and the filling time t _FT represent the preferred operating point T of the system. Two suboptimal points A and B, which lie on the reaction curve, are also shown;

Fig. 8 shows the relationship between the purge time t _p and the filling time t _F for four different operating states of a collecting vessel in which the reduction of NO _x capacitance thereof progressively increases, and also the extrapolated flushing times for the oxygen storage section t _Posc the entire rinsing time t _p ;

Fig. 9 shows the relationship between the NO _x capacitance and the rinse time for four different operating states of the receiver, which is due to the sulphurisation and / or thermal damage pogressiv increasingly ver deteriorated;

FIG. 10 is a flowchart for the optimization of the flushing time t _p of the collecting vessel;

Fig. 11 is a flow diagram for a system optimization;

FIG. 12 is a flow chart for determining whether a desulphurisation of the collecting vessel is necessary;

. Graphically the relationship is subjected between the on storage in the collecting vessel th oxygen carrier portion and the relative amount of time during which the collecting vessel initially a stream of NO _x Figure 13;

FIG. 14 is a graph showing a Spülkraftstoffanteil depending on the relative filling time;

FIG. 15 is a table of Grundfüllrate R _ij (NO _x -Kapazitätsabnahme) of the collecting vessel for various speeds and loads at given points ta Bella driven recorded values of temperature, air-fuel mixing ratio, exhaust gas recirculation amount and advancing the ignition timing;

. Figs. 16a-16d, a list of core field conditions for the air-fuel mixing ratio, the exhaust gas recirculation EGR, advancing the ignition timing and the temperature of the receiver for which the fill rate R shown in Figure 15 _ij of the collection vessel were determined;

FIG. 17 shows how the capacity decrease modification factor of the collecting vessel changes with the temperature; FIG.

FIG. 18 shows how the modification factors for the air-fuel mixture ratio, the exhaust gas recirculation quantity or rate, the advance of the ignition timing, depend on a change in the values of the air-fuel mixture ratio, the recirculated exhaust gas quantity and the value of the values in FIG. 16 Change before changing the ignition timing and

Fig. 19 is a flowchart for determining when a rinsing operation of the catch on the vessel is to be initiated.

Referring now to the drawings and initially to FIG. 1, a powertrain control module (PCM), which is an electronic engine control unit and includes a ROM, RAM and a CPU, is generally designated 10 . The PCM controls a set of injectors 12 , 14 , 16 and 18 that inject fuel into the cylinders of a four-cylinder internal combustion engine 20 . The fuel injectors have a conventional function and design and are positioned in such a way that they inject fuel into the respectively assigned cylinders in precise quantities measured by the control unit 10 . The control unit 10 transmits a fuel injector signal to the injectors to maintain an air / fuel mixture ratio ("AFR") determined by the control unit 10 . An air or air mass flow sensor 22 lies in the intake manifold 24 of the engine and generates a signal which indicates the air mass flow resulting from the position of a throttle valve 26 . The air mass flow signal is used in the control unit 10 to calculate an air mass value which indicates the mass of the air flowing into the intake system per unit of time. A heated exhaust gas oxygen sensor (HEGO) 28 detects the oxygen content in the exhaust gas flow generated by the engine and transmits a signal to the control unit 10 . The HEGO sensor 28 is used to control the air-fuel mixture ratio of the engine, particularly while the engine is operating at the stoichiometric mixture ratio.

An exhaust system that includes one or more exhaust pipes leads the exhaust gas generated by the combustion of an air-fuel mixture in the engine to a conventional, controlled three-way catalytic converter 30 . The three-way catalytic converter 30 contains a catalytic material that chemically changes the composition of the exhaust gas generated by the engine and generates a catalytically treated exhaust gas. The catalytically treated exhaust gas is passed through an exhaust pipe 32 to a downstream NO _x collecting vessel 34 , which consists of materials previously described and finally discharged into the atmosphere through an exhaust 36 .

A second HEGO sensor 38 is located downstream of the collecting vessel 34 and generates a signal which is fed to the control unit 10 for the diagnosis and regulation according to the invention. The second HEGO sensor 38 is used to monitor the HC effect of the three-way catalytic converter 30 by means of known methods which compares the signal amplitude of the second HEGO sensor 38 with that of the first HEGO sensor 28 during a conventional stoichiometrically controlled limited operating cycle. A lying in the middle of the receptacle tempera ture sensor 42 generates an output signal indicating the instantaneous temperature T of the receptacle 34 . Still other sensors (not shown) provide the control unit 10 with additional information about the work or behavior of the engine such as. B. the camshaft position, the crankshaft position, the win kel speed, the throttle valve position and air temperature. Information from these sensors is used by the control unit 10 to control engine operation.

Fig. 2 shows a typical voltage response of a switched oxygen sensor, for. B. the second HEGO sensor 38 , depending on the air-fuel mixture ratio. The voltage output signal of the second HEGO sensor 38 switches between low and high level when the exhaust gas mixture changes from lean to rich relative to the stoichiometric air-fuel mixture ratio of approximately 14.65. Since the air-fuel mixture ratio is low during the filling time, NO _x generated by the engine flows through the three-way catalytic converter 30 and an exhaust pipe 32 into the collecting vessel 34 , where this NO _{x is} stored or stored.

Fig. 3 shows a typical operation of the receiver with the rinse cycle. The upper waveform ( FIG. 3a) shows the relationship between the lean filling time t _F and the fat rinsing time t _p for three different rinsing times 1 , 2 and 3 . The second waveform ( Fig. 3b) shows the response of the second HEGO sensor 38 to the three flushing times. The CO and HC quantities flowing through the collecting vessel, which influence the downstream sensor 38 , serve as an indicator of the effectiveness the rinsing process of the collecting vessel. The peak voltage level of the oxygen sensor in the exhaust pipe is an indicator of the amounts of NO _x and O ₂ that are still stored in the collecting vessel. With a short purge time 1 , a very weak reaction of the oxygen sensor occurs because the collecting vessel has not yet been completely rinsed out, which results in a narrow needle-shaped tip of the CO content in the exhaust and in the voltage response of the second HEGO sensor, which is strongly dependent on it. In this case, the peak voltage V _{p of} the sensor does not reach the reference voltage V _ref . For a moderate or optimal purge time 2 , the voltage response of the second HEGO sensor becomes equal to the reference voltage V _ref and indicates that the collecting vessel is almost flushed out because a still acceptable very small amount of CO is present in the exhaust gas. For a long purging time 3 , the peak voltage of the second HEGO sensor exceeds the reference voltage V _ref and thus indicates that the collecting vessel has either been completely or excessively purged, and this results in an undesirably high CO and HC emission in the exhaust gas, as illustrated by the waveform in Fig. 3d.

The data collection window for the voltage of the second HEGO sensor is shown in the signal form in FIG. 3c. During this window, the PCM collects data based on the output voltage of the second HEGO sensor 38 . FIG. 4 shows an enlarged view of the response of the sensor 38 to the three different rinsing times shown in FIG. 3. The time interval Δt ₂₁ is equal to the time period during which the sensor voltage exceeds the reference value V _ref . For a sensor peak voltage V _P below the reference voltage V _ref , the PCM 10 achieves a smooth transition to the metric of FIG. 5 by linearly extrapolating the sensor saturation time t _sat from t _sat = t _satrefr to t _sat = 0. The PCM 10 uses the relationship shown in FIG. 6 and makes the sensor saturation time t _sat proportional to the peak voltage V _{p of} the sensor, as shown in FIG. 6.

In FIG. 5, the relationship between the normalized time t _sat Sauerstoffsensorsätti is supply and the purge time t _p shown. The saturation time t _{sat of} the sensor is the normalized time period in which the output signal of the second HEGO sensor is above V _ref and is equal to Δt ₂₁ / Δt _21norm , where Δt _{21norm is} the normalization factor. The saturation time t _{sat of} the sensor is normalized by the desired value t _{sat_desired} . For a given filling time t _F and a given state of the collecting vessel there is an optimal flushing time t _{Psat_desired} , which results in an optimal normalized saturation time t _sat = 1, in which the contents of HC and CO in the exhaust gas are not excessively large and at which is still an acceptable absorption effect of the NO _{x container} . The rinse time is too long for a saturation time t _sat <1 and should be reduced. The rinse time is too short for a saturation time t _sat <1 and should be increased. In this way, a re regulation of the purging of the collecting vessel in a closed loop based on the output signal of the second HEGO sensor 38 can be achieved.

FIG. 7 shows the target relationship between the flushing time t _P and the filling time t _F for a given operating state of the engine and a given operating state of the collecting vessel. The two sub-optimal rinsing times t _{P_subopt1} and t _{P_subopt2} each _correspond to an under or over-rinsing of the collecting _vessel 34 for a defined filling time t _FT . The rinsing time t _P in which the NO _x collected in the collecting vessel is rinsed out in an optimal manner is referred to as t _PT . This point corresponds to a desired or desired rinse time t _sat = t _{sat_desired} . This purge time minimizes CO emissions in the exhaust gas during the specified fill time t _FT . This procedure also results in a determination of the purge time t _{P_osc} for stored oxygen, which depends on the amount of oxygen stored directly in the collecting vessel. Oxygen can, for. B. directly absorbed in the form of cerium oxide. The purging time t _{P_osc} for stored oxygen can either be carried out by extrapolating two or more purging times to the point t _F = 0 or by performing the t _P optimization in the vicinity of the point t _F = 0. The working point T2 is achieved by deliberately setting t _FT2 <t _FT and finding t _PT2 through the optimization.

Fig. 7a illustrates optimizing the filling time t _F. For a given filling time t _FT , the optimal rinsing time t _PT according to FIG. 7 is determined. Then, by gradually approaching a value t _FB that is a little smaller than the initial value t _FT and gradually approaching a value t _FA that is slightly larger than the initial value t _FT , the filling time is fluctuating (trembling) approximated. The purging time optimization is used at all three points T, A and B to determine the variation of t _P depending on t _F. The change in t _P from A to T and also from B to T is evaluated. In FIG. 7a the change from B to T is greater than the change from A to T. The absolute value of these differences is controlled or regulated in such a way that it remains within a certain tolerance DELTA_MIN, as will be explained in more detail below with reference to FIG. 11 . The absolute value of the differences is proportional to the slope of the curve, which represents t _{P as a} function of t _F.

This optimization process defines the working point T as the "shoulder" of the t _P as a function of t _F indicating curve. T _psat represents the saturation value of the rinsing time for infinitely long filling times.

The results of the optimization program for the flushing time t _P and the filling time t _F are shown in FIG. 8 for four different states of the collecting vessel, which have four different collecting levels for NO _x and oxygen. Both the flushing time t _P and the filling time t _F were optimized using the procedures illustrated in FIGS. 7 and 7a. The point determined by FIG. 8 is referred to as the optimal working point T1, for which the flushing time t is _PT1 and the filling time t is _FT1 . The index "1" indicates a condition A and that the collecting vessel has not deteriorated. As the receptacle deteriorates due to sulfur contamination, thermal damage and other factors, states B, C and D of the receptacle are reached. The rinse and fill time optimization programs are run continuously when the motor is in quasi-steady state. Optimal working points T2, T3 and T4 are achieved, which correspond to the receptacle states B, C and D. Both the NO _x saturation level, which is reflected in t _PT1 , t _PT2 , t _PT3 and t _PT4 , as well as the purging times t _PoscT1 , t _PoscT2 , t _PoscT3 and t _PoscT4 , which are dependent on oxygen _storage , change with the collecting _{vessel state} and become typical with the deterioration of the receptacle smaller. The amount of purge fuel for the NO _x portion of the purge is equal to t _PNox = t _PT - t _Posc . It can be seen that the amount of flushing fuel is equivalent to the flushing time for a given operating state. The control unit 10 controls the actual amount of purge fuel by modifying the period of time during which the engine 20 is operated with a predetermined rich air-fuel mixture ratio. To simplify the description, it is assumed that the flushing time is equivalent to the proportion of flushing fuel in the operating state assumed in the description. In this way, the purge time required for the stored NO _x mass and the amount of oxygen can be determined directly and used for diagnosis and control.

Fig. 9 illustrates the relationship between the NO _x purge time t _PNOx and the NO _x capacity of the collecting vessel. States A, B and C are decided so that there is an acceptable NO _x storage effect and fuel consumption, while state D is unacceptable. Therefore, when approaching state D, a desulfurization operation is set for the receiver to restore the NO _x storage capacity of the receiver and to reduce the fuel consumption associated with a high NO _x purge frequency. The change in t _Posc can provide additional information about the aging of the receptacle due to the change in oxygen storage.

Fig. 10 illustrates a flowchart for the optimization of the flushing time t _P. The task of this program is to optimize the purging air-fuel mixture ratio peak for a given value of the filling time t _F. This program is contained within the software for system optimization, as will be explained below with reference to FIG. 11. At decision block 46 , the status of a flag for a purge event is checked, and if the flag is set, a lean NO _x purge is performed, which is indicated in block 48 . The indicator for a rinsing process is set when the filling of the collecting vessel is complete. For example, the flag would be set in block 136 of FIG. 19 if this purging event setting method is used. In block 50 , the output voltage of the oxygen sensor (EGO) is sampled during a predetermined data collection window and the peak voltage V _P and the transition times t ₁ and t _{2 are} determined if these occur. During the data collection window, the waveform change of the output signal from the EGO sensor is sampled, as shown in FIG. 3. If it has been determined by decision block 52 that V _{P is} greater than V _ref , as indicated in blocks 54 and 56 , the sensor saturation time t _{sat is} proportional to Δt ₂₁ , which is the time period during which the output voltage of the EGO sensor is above V _ref . In the case V _P <V _ref , as indicated in block 58 , t _sat is determined from a linearly extrapolated function. For this function, which is shown in FIG. 6, it applies that t _{sat is} determined by making it proportional to the peak amplitude V _P. This achieves a smooth transition from the case V _P <V _ref to the case V _P <V _ref and thus achieves a continuous positive and negative error function t _{sat_error} (k), as indicated by block 60 , which is suitable for feedback control , wherein this error function t _{sat_error} (k) is equal to a desired value or t _{sat_desired} for the saturation time of the sensor minus the current saturation time t _sat . The error function t _{sat_error} (k) is then normalized in block 62 by dividing it by the desired sensor saturation time t _{sat_desired} .

The resulting normalized error t _{sat_error_norm} (k) is used as an input variable for a feedback controller, e.g. B. uses a PID controller (proportional differential integral controller). The output variable of the PID controller is a multiplicative correction factor for the rinsing time of the collecting vessel or PURGE_MUL, as shown in block 64 . There is a direct monotonous relationship between t _{sat_error_norm} (k) and PURGE_MUL. If t _{sat_error_norm} (k) <0, the collecting vessel is flushed below optimally and PURGE_MUL must be increased from its basic value in order to supply more CO for flushing NO _x . If t _{sat_error_norm} (k) <0, the collecting vessel has been excessively purged and PURGE_MUL must be reduced from its basic value in order to deliver less CO for the NO _x purge. As indicated in block 66, this results in a new value for the flushing time t _P (k + 1) = t _P (k) × PURGE_MUL. The purging time optimization continues until the absolute value of the difference between the old and new purging time values becomes less than an allowable tolerance, as indicated in blocks 68 and 70 . If | t _P (k + 1) - t _P (k) | ≧ ε, this means that the optimal rinsing time t _P of the PID control is not within the permitted tolerance ε. Accordingly, as block 70 indicates, the new purge time calculated in block 66 is used in the subsequent purge cycles until block 68 is satisfied. The filling time t _F is adjusted using the equation (2) given below (during the t _P optimization) until the optimal rinsing time t _{P is} reached. If | t _P (k + 1) - t _P (k) | <ε, the purge time optimization converges, and the instantaneous value of the purge time is stored, as indicated by block 72 , and the optimization procedure can return to the program for optimizing the time t _F shown in FIG. 11. Instead of only changing the purge time t _P , the relative enrichment of the air / fuel mixture used for the purge event (see FIG. 3) can also be changed in the same way.

Fig. 11 shows a flow chart for system optimization, the Spülzeit- also includes both the Füllzeitoptimierung. According to block 74, the filling time optimization is only carried out when the engine is operating in quasi-stationary mode. In this context, it should be mentioned that quasi-steady-state operation occurs when the rates of change of certain engine operating parameters, such as engine speed, engine load, air flow, ignition timing, exhaust gas recirculation EGR, are below certain levels. In block 76 , the step size FILL_STEP is selected for the fill time equal to STEP_SIZE, which, if FILL_STEP <0, results in an extended fill time. As explained below and shown in FIG. 14, STEP_SIZE is matched for the capacity utilization rate R _ij .

In block 78 , the rinse time optimization described above with reference to FIG. 10 is carried out. It optimizes the flushing time t _P for a given filling time. The size PURGE_MUL at the end of the purge time optimization carried out in block 78 is stored as size CTRL_START and, as block 80 indicates, the fill time multiplication factor FILL_MUL is incremented by FILL_STEP. The filling step is multiplied by FILL_MUL in block 82 and thus the stepwise approximation of t _{F is} promoted. In block 84 , the purging optimization of FIG. 10 is carried out for the new value of the filling time t _F (k + 1). The size PURGE_MUL reached at the end of the flushing time optimization carried out in FIG. 10 is stored as the value CTRL_END in block 86 . The change in the flushing multiplication factor CTRL_DIFF = ABS (CTRL_END-CTRL_START) is also stored in block 86 and compared in block 88 with a reference value DELTA_MIN. DELTA_MIN corresponds to the tolerance explained with reference to FIG. 7a, and CTRL_END and CTRL_START correspond to the two values of t _P that were found in FIG. 7a for A and T or for B and T. If the change in the flushing multiplication factor is greater than DELTA_MIN, the sign of FILL_STEP is changed in order to find an optimal filling time in the opposite direction, as indicated in block 90 . If the change in the rinsing multiplication factor is less than DELTA_MIN, the search for an optimal filling time t _{F continues} in the same direction as indicated in block 92 . In block 94 , FILL_MUL is incremented with the selected step size FILL_STEP. In block 96 the filling time t _F (k + 1) is changed by multiplication by FILL_MUL. The result is the selection of the optimum value point t _PT as the working point and the continuous trembling or fluctuating movement around this working point. If the engine is not operating in a quasi-steady state during this procedure, the fill time optimization routine is exited, as indicated by block 74 , and the fill time specified in equation (2) below is used.

Figure 12 illustrates a flow diagram for a program for desulfurization of the receiver according to the invention. In block 100 , a reference value t _{PNOxref is} read from a reference _table , which represents the rinsing time for a collecting vessel that has not deteriorated under given operating conditions. The value t _PNOxref can be a function of the air mass flow rate of the air / fuel mixture ratio and other parameters. In block 102 , the current rinsing time t _P (k) is called again and compared with t _PNOxref minus a predetermined tolerance _value TOL and if it is found that t _P (k) <t _PNOxref - TOL, a desulfurization of the collecting _vessel is determined . Desulphurization involves heating the receiver to approximately 650 ° C for approximately ten minutes with an air / fuel mixture that is richer than the stoichiometric mixture ratio and has a value of 0.98 λ. A desulfurization counter D is reset in block 104 and continues to be counted whenever a desulfurization process is carried out, as indicated in block 106 . After the desulfurization process has ended, the optimal rinsing and filling time is determined in block 108 , as previously described in connection with FIG. 11. The new flush time t _P (k + 1) is compared to the reference time t _PNOxref minus the tolerance value TOL in block 110 and if it is found in decision block 112 that t _P (k + 1) <t _PNOxref - TOL performed at least two additional desulphurizations. If the receiver still fails the test, a malfunction indicator lamp (MIL) is illuminated, as indicated by block 114 , and the receiver should then be replaced with a new one. If the condition is met and t _P (k) ≧ t _PNOxref - TOL. Has the receptacle 34 not deteriorated to the extent that immediate maintenance is required and normal operation is resumed.

A NO _x purge event is determined when a given capacity, which is less than the actual capacity of the collecting vessel 34, has been filled, or has been used up by the storage of NO _x . Oxygen is stored in the receptacle either in the form of cerium oxide or as NO _x , and the sum of both represents the amount of oxygen carrier stored. FIG. 13 illustrates the relationship between the oxygen carrier stored in the receptacle 34 and the period of time during which the receptacle was received 34 is subjected to a NO _x flow. NO _x storage takes place at a slower rate than that of oxygen storage. The optimal operating point based on the NO _x generation time corresponds to the "shoulder" of the curve or approximately 60-70% of the relative NO _x generation time in this figure. A value of 100% on the abscissa corresponds to the saturated NO _x capacity of the collecting vessel 34 . The values for stored NO _x and stored oxygen are also shown. The capacity utilization rate R _ij is the initial slope of this curve, the percentage of the stored oxygen carrier divided by the percentage of the NO _x generation time.

FIG. 14 is similar to FIG. 13 with the exception that the relative proportion of the flushing fuel is plotted against the relative filling time t _F. The capacity utilization rate R _ij (% of the flushing fuel /% of the filling time) can be seen as the initial slope of this curve. For a given calibration of the air-fuel mixture ratio, the recirculating exhaust gas EGR and the ignition timing SPK for a given speed and load, the relationship of the relative amount of NO _x generated is linearly dependent on the relative fill rate t _F. Fig. 14 illustrates the relationship between the amount of HC and CO-containing scavenging fuel used to purge the receiver and the time during which the receiver is subjected to NO _x flow. The amount of fuel required for purging is divided between the amount required to purge the stored oxygen and the amount required to purge the NO _x stored as nitrate.

The decrease in the NO _x storage capacity in the collecting vessel can be expressed by the following equations

The basic or unmodified collection vessel capacity use RS (%) is given by equation (1), which represents a time-weighted summation of the cell filling rate R _ij (% / s) over all work cells controlled by the filling process of the collection vessel as a function of speed and load. The relative fill rate of the cell R _ij (% flushing fuel /% fill time) is obtained by dividing the change in the flushing time by the fill time t _F , which corresponds to a 100% fill for this cell. It should be noted that equation (1) is provided only as a reference, while equation (2) with its modifying factors is the real working equation. The modification factors in equation (2) are M ₁ (T) for the temperature T of the collecting vessel, M ₂ for the air-fuel mixture ratio (AFR), M ₃ for the recirculated exhaust gas quantity (EGR) and M ₄ for the advance SPK of the ignition point . The individual filling rates R _{ij of} the cells are summed up to an amount below 100%, at which the capacity of the collecting vessel is essentially reached but is not fully used. For this capacity, the sum t _{F of} the times spent in all cells gives the filling time of the collecting vessel. The result of this calculation, which results from equation (2), is the actual capacity utilization RSM (%) of the collecting vessel. The basic filling rate for a given area is multiplied by the time t _k spent in this area, then multiplied by M ₂ , M ₃ and M ₄ and added up continuously. The sum is changed by the temperature modification factor M ₁ (T). When the sum of RSM modified in this way comes close to 100%, the collecting vessel is almost completely filled with NO _x and a rinsing process is defined.

Fig. 15 shows a map table with stored data of the Grundfüllrate R _ij of the collection vessel. The overall system, which consists of the engine and exhaust gas purification system, which contains the three-way catalytic converter and the collecting vessel, is represented by a speed-load matrix map. A representative calibration of the air-fuel mixture ratio ("AFR"), EGR and EGR advance is used. The temperature T _{ij of} the collecting vessel is recorded for each speed load range. Was FIGS. 16a-16d show representative tabular representations of the mapped conditions for the air-fuel mixture ratio, exhaust gas recirculation, EGR, the advancing of the ignition timing and for the temperature T _ij of the collecting vessel, for guessing the filler R _ij of the collection vessel in Figure 15 determines. .

If the current operating conditions of the vehicle deviate from the map states shown in FIG. 16, corrections are made to the modifying factors M ₁ (T), M ₂ (AFR), M ₃ (EGR) and M ₄ (advance ignition timing). The correction for M ₁ (T) is shown in FIG. 17. As shown, since the NO _x absorption capacity of the collection vessel reaches a maximum value at an optimal temperature T ₀ , which is 350 ° C. in one embodiment, a correction is made which reduces the NO _x capacity of the collection vessel when its temperature T rises above the optimum temperature T ₀ or falls below it.

FIGS. 18a-18c show the corrections modifiers M _2, M ₃ and M _4. These corrections are applied when the actual values of the air-fuel mixture ratio, the amount of exhaust gas returned and the advance of the ignition timing differ from the values stored in the map of FIG. 15.

Fig. 19 shows the flowchart for determining the basic filling time of the collecting vessel 34 , that is, when the time for rinsing the collecting vessel 34 has come. If, as determined in block 120 , a flushing process has been completed and the engine is running in lean mode, as determined by block 122 , the collecting vessel is filled, as indicated by block 124 . The filling time is based on an estimate of the decrease in the NO _x storage capacity R _ij with a suitable modification for the air / fuel mixture ratio, for the exhaust gas recirculation quantity EGR, the advance of the ignition point and the temperature of the collecting vessel. The engine speed and the load are read out in block 126 and a basic fill rate R _{ij is then} determined in block 128 from a reference table using the speed and the load as input indices ( FIG. 15). The collection vessel temperature, the air / fuel mixture ratio of the engine, the amount of exhaust gas returned, the advance ignition timing and the time t _k are determined in block 130 ( FIGS. 16a-16d) and used in block 132 to calculate a time-weighted sum RSM, which is based on the in a given speed-load range time period based. When RSM approaches 100%, a purge is determined as indicated by blocks 134 and 136 . Otherwise, the filling process of the collecting vessel continues in block 122 . The filling time determined in FIG. 19 is the basic filling time, and this will change if the collecting vessel is sulphured or has undergone thermal damage. However, with the procedures described above ( FIGS. 7a, 8 and 11), where the optimal filling time has been determined by a trembling process, the need for a desulfurization process is determined and also whether thermal damage to the collecting vessel has occurred.

The set value for the purge time t _p must contain a component t _posc for the oxygen purge and t _pNOx for the NO _x purge. Therefore, t _P = t _Posc + t _PNOx . The control unit 10 contains a reference _table which provides the value of t _Posc , which is strongly dependent on the temperature. For a _receptacle containing cerium oxide, t _{Posc obeys} the Arrhenius equation. T _Posc = C _exp (-E / kT), where C is a constant that depends on the type and condition of the collecting vessel, E is an activation energy and T is the absolute temperature.

Although a preferred embodiment in the foregoing description  the invention has been described in detail, the relevant experts readily recognize that features of the invention in different ways, who deviates from the scope of the attached claims, changes that can.

Claims (10)

1. A method for controlling the purging of an amount of NO _x , which has previously been collected in a lean NO _x collecting vessel, which is part of an exhaust gas purification system of an internal combustion engine, which contains a sensor which is used to generate an oxygen concentration in the the collecting vessel from the engine flowing exhaust gas signal is set up, characterized in that the method comprises:
Determining the amount of oxygen previously stored in the collecting vessel on the basis of the peak amplitude of the signal from the oxygen sensor, which was generated during a first purging process of the collecting vessel;
Flushing NO _x from the receptacle at a frequency inverse to the amount of NO _x stored in the receptacle, and performing a desulfurization operation of the receptacle to restore the capacity of the receptacle when the purge time is shorter than a predetermined minimum purge time. 2. The method according to claim 1, characterized in that a deterioration tion of the collecting vessel is displayed when a predetermined number of Desulfurization operations of the collecting vessel can be carried out without that the rinsing time is extended. 3. The method according to claim 2, characterized in that it further contains:
Generation of a rinse adjustment multiplier depending on the capacity of the collecting vessel;
Adjusting the filling time as a function of the multiplier, in order to achieve a storage of an amount of NO _x which is sufficient to fill the collecting vessel up to a predetermined fraction of its capacity. 4. The method according to claim 3, characterized in that an initial value the filling time of the collecting vessel from a reference table as a function of Engine speed and the load is determined. 5. The method according to claim 3, characterized in that an initial value the filling time of the collecting vessel from a reference table as a function of Air mass flow rate is determined. 6. The method according to claim 3, characterized in that an initial value the filling time of the collecting vessel an inverse power of the product from Mo gate load and engine speed. 7. A method for filling and purging a lean NO _x collecting vessel, which lies in the exhaust gas path of an internal combustion engine upstream of an oxygen sensor, so that the collecting vessel is essentially filled to its capacity during a filling time and is essentially emptied during a subsequent purging time , characterized in that the method comprises:
By temporally integrating the filling rate taken from a characteristic value table with which the collecting vessel fills, information is derived as to whether the collecting vessel has been filled with NO _x up to a predetermined fraction of its capacity;
a purging process is carried out in which the strength of the purging process is just sufficient to purge the NO _x stored in the collecting vessel, in which the output signal of the oxygen sensor is monitored using a time and voltage-dependent oxygen sensor metric and the purging time is continuously adjusted to its optimum value that the flushing strength is just sufficient to flush the NO _x out of the collecting vessel;
the rinsing time is continuously compared with a reference rinsing time which corresponds to that of a deteriorated collecting vessel, and if the reference rinsing time is exceeded, one or more desulfurization processes are initiated;
the optimum rinsing time is compared with the reference rinsing time after desulfurization, and
if the rinsing time does not return to a value longer than the reference rinsing time, a deterioration of the collecting vessel is indicated. 8. System for controlling the purging process of a lean NO _x collecting vessel located in the exhaust gas path of an internal combustion engine, characterized in that the system comprises:
an oxygen sensor responsive to the exhaust gas flowing through the receiver;
a control module that is programmed to determine the amount of NO _x stored in the collection vessel based on a peak amplitude of the voltage derived from the oxygen sensor during a purging operation of the collection vessel, the control module being further programmed to remove the collection vessel from To purge NO _x at a frequency inversely related to the amount of NO _x stored in the receptacle and to desulfurize the receptacle to restore its capacity when the purge time is shorter than a predetermined minimum purge time. 9. System according to claim 8, characterized in that the tax he / control module is still programmed to worsen the Display receptacle when a predetermined number of deviations was carried out without the Rinsing time has extended. 10. System according to claim 9, characterized in that the STEU he / control module is further programmed to generate a Spülzeitabgleichmulti Multiplier depending on the capacity of the collecting vessel and to compare the fill time as a function of the multiplier to the incorporation of a sufficient amount of NO _x to fill the receptacle up to a certain fraction of its capacity.